3GPP TR 36.814 V9.0.

0 (2010-03)
Technical Report
3rd Generation Partnership Project;
Technical Specification Group Radio Access Network;
Evolved Universal Terrestrial Radio Access (E-UTRA);
Further advancements for E-UTRA physical layer
aspects
(Release 9)

The present document has been developed within the 3
rd
Generation Partnership Project (3GPP
TM
) and may be further elaborated for the purposes of 3GPP.

The present document has not been subject to any approval process by the 3GPP

Organizational Partners and shall not be implemented.
This Specification is provided for future development work within 3GPP

only. The Organizational Partners accept no liability for any use of this
Specification.
Specifications and reports for implementation of the 3GPP
TM
system should be obtained via the 3GPP Organizational Partners' Publications Offices.
3GPP
Keywords
UMTS, Radio
3GPP
Postal address
3GPP support office address
650 Route des Lucioles - Sophia Antipolis
Valbonne - FRANCE
Tel.: +33 4 92 94 42 00 Fax: +33 4 93 65 47 16
Internet
http://www.3gpp.org
Copyright Notification
No part may be reproduced except as authorized by written permission.
The copyright and the foregoing restriction extend to reproduction in all media.
© 2010, 3GPP Organizational Partners (ARIB, ATIS, CCSA, ETSI, TTA, TTC).
All rights reserved.
UMTS™ is a Trade Mark of ETSI registered for the benefit of its members
3GPP™ is a Trade Mark of ETSI registered for the benefit of its Members and of the 3GPP Organizational Partners
LTE™ is a Trade Mark of ETSI currently being registered for the benefit of its Members and of the 3GPP
Organizational Partners
GSM® and the GSM logo are registered and owned by the GSM Association
3GPP TR 36.814 V9.0.0 (2010-03) 2 Release 9
Contents
Foreword................................................................................................................................................6
1 Scope...................................................................................................................................................7
2 References...........................................................................................................................................7
3 Definitions, symbols and abbreviations...............................................................................................7
3.1 Definitions.............................................................................................................................................................7
3.2 Symbols.................................................................................................................................................................7
3.3 Abbreviations........................................................................................................................................................8
4 Introduction.........................................................................................................................................8
5 Support of wider bandwidth................................................................................................................8
5.1 MAC-PHY interface.............................................................................................................................................8
5.2 DL control signalling............................................................................................................................................8
5.3 UL control signalling............................................................................................................................................9
6 Uplink transmission scheme..............................................................................................................10
6.1 Uplink spatial multiplexing.................................................................................................................................10
6.2 Uplink transmit diversity.....................................................................................................................................11
6.2.1 Transmit Diversity for Uplink Control Channel..............................................................................................11
6.3 Uplink multiple access........................................................................................................................................12
6.4 Uplink reference signals......................................................................................................................................12
6.5 Uplink power control..........................................................................................................................................12
7 Downlink transmission scheme.........................................................................................................12
7.1 Physical channel mapping...................................................................................................................................12
7.2 Downlink spatial multiplexing............................................................................................................................12
7.2.1 Feedback in support of downlink spatial multiplexing....................................................................................14
7.3 Enhanced multiuser MIMO transmission...........................................................................................................15
7.4 Downlink reference signals.................................................................................................................................15
7.4.1 Demodulation reference signal........................................................................................................................15
7.4.2 CSI reference signal.........................................................................................................................................16
7.5 Downlink transmit diversity................................................................................................................................16
8 Coordinated multiple point transmission and reception.....................................................................16
8.1 Downlink coordinated multi-point transmission.................................................................................................16
8.1.1 Terminology and definitions ...........................................................................................................................16
8.1.2 Radio-interface specification areas..................................................................................................................17
8.1.3 Feedback in support of DL CoMP...................................................................................................................18
8.1.3.1 Explicit Feedback in support of DL CoMP....................................................................................................18
8.1.3.2 Implicit Feedback in support of DL CoMP....................................................................................................19
8.1.4 Overhead in support of DL CoMP operation...................................................................................................20
8.2 Uplink coordinated multi-point reception...........................................................................................................20
9 Relaying functionality.......................................................................................................................20
9.1 Relay-eNodeB link for inband relay...................................................................................................................21
9.1.1 Resource partitioning for relay-eNodeB link...................................................................................................21
9.1.2 Backward compatible backhaul partitioning ...................................................................................................21
9.1.3 Backhaul resource assignment.........................................................................................................................22
9.2 Relay-eNodeB link for outband relay.................................................................................................................23
9.3 Relay-eNodeB link for inband relay Type 1b.....................................................................................................23
9A Heterogeneous deployments...........................................................................................................23
9A.1 Rel-8/9 schemes...............................................................................................................................................23
9A.2 Non-Rel-8/9 schemes.......................................................................................................................................24
9A.2.1 CA-based scheme..........................................................................................................................................24
9A.2.2 Non-CA based schemes................................................................................................................................24
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 3 Release 9
10 Evaluation of techniques for Advanced E-UTRA...........................................................................24
10.1 Cell spectral efficiency and cell-edge spectral efficiency against 3GPP target................................................25
10.1.1 3GPP Case1 (3GPP spatial channel model)...................................................................................................26
10.1.1.1 FDD, Downlink............................................................................................................................................26
10.1.1.2 TDD, Downlink............................................................................................................................................29
10.1.1.3 FDD, Uplink.................................................................................................................................................31
10.1.1.4 TDD, Uplink................................................................................................................................................33
10.2 Cell spectral efficiency and cell-edge spectral efficiency for ITU-R requirements..........................................34
10.2.1 Indoor (InH channel model)...........................................................................................................................36
10.2.1.1 FDD, Downlink (InH) .................................................................................................................................36
10.2.1.2 TDD, Downlink (InH) .................................................................................................................................37
10.2.1.3 FDD, Uplink (InH) ......................................................................................................................................38
10.2.1.4 TDD, Uplink (InH) ......................................................................................................................................39
10.2.2 Microcellular (UMi channel model)..............................................................................................................40
10.2.2.1 FDD, Downlink (UMi) ................................................................................................................................40
10.2.2.2 TDD, Downlink (UMi) ...............................................................................................................................42
10.2.2.3 FDD, Uplink (UMi) ....................................................................................................................................44
10.2.2.4 TDD, Uplink (UMi) ....................................................................................................................................45
10.2.3 Base coverage urban (UMa channel model)..................................................................................................47
10.2.3.1 FDD, Downlink (UMa) ..............................................................................................................................47
10.2.3.2 TDD, Downlink (UMa) ...............................................................................................................................49
10.2.3.3 FDD, Uplink (UMa) ....................................................................................................................................51
10.2.3.4 TDD, Uplink (UMa) ....................................................................................................................................53
10.2.4 High speed (RMa channel model)..................................................................................................................54
10.2.4.1 FDD, Downlink (RMa) ...............................................................................................................................54
10.2.4.2 TDD, Downlink (RMa) ...............................................................................................................................55
10.2.4.3 FDD, Uplink (RMa) ....................................................................................................................................56
10.2.4.4 TDD, Uplink (RMa) ....................................................................................................................................57
Annex A: Simulation model..................................................................................................................58
A.1 Link simulation Scenarios..................................................................................................................................58
A.2 System simulation Scenarios..............................................................................................................................58
A.2.1 System simulation assumptions......................................................................................................................58
A.2.1.1 Reference system deployments.....................................................................................................................58
A.2.1.1.1 Homogeneous deployments.......................................................................................................................58
A.2.1.1.2 Heterogeneous deployments......................................................................................................................60
A.2.1.1.3 Assumptions for Coordinated Multi point Transmission and Reception Evaluations ..............................77
A.2.1.1.4 Assumptions for Relay Evaluations ..........................................................................................................78
A.2.1.1.5 Assumptions for indoor RRH/Hotzone Evaluations..................................................................................81
A.2.1.2 Channel models.............................................................................................................................................82
A.2.1.3 Traffic models...............................................................................................................................................82
A.2.1.3.1 FTP traffic models......................................................................................................................................83
A.2.1.3.2 Performance metrics...................................................................................................................................84
A.2.1.3.2.1 Reference points......................................................................................................................................84
A.2.1.3.3 File dropping criteria..................................................................................................................................85
A.2.1.3.4 Estimation of range ofλ and K...................................................................................................................85
A.2.1.4 System performance metrics.........................................................................................................................86
A.2.1.5 Scheduling and resource allocation...............................................................................................................87
A.2.1.6 Antenna gain for a given bearing and downtilt angle...................................................................................87
A.2.1.6.1 Polarized antenna modelling......................................................................................................................87
A.2.1.6.2 Antenna gain for a given direction and mechanical tilt angle....................................................................88
A.2.1.7 Advanced receivers modeling ......................................................................................................................90
A.2.1.7.1 Iterative soft interference cancellation receivers .......................................................................................90
A.2.1.8 Effective IoT.................................................................................................................................................90
A.2.2 System level simulator calibration..................................................................................................................91
A.2.3 Downlink CoMP evaluation assumptions for intra-NodeB CoMP.................................................................94
A.3 Evaluation assumptions for IMT-A....................................................................................................................95
A.4 Detailed simulation results.................................................................................................................................98
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 4 Release 9
Annex B:
Channel models (Informative)....................................................................................98
B.1 Channel models ...............................................................................................................................................100
B.1.1 Void 100
B.1.2 Primary module.............................................................................................................................................100
B.1.2.1 Path loss models..........................................................................................................................................101
B.1.2.1.1 Autocorrelation of shadow fading............................................................................................................104
B.1.2.2 Primary module channel model parameters................................................................................................104
B.1.2.2.1 Generic model..........................................................................................................................................104
B.2 References........................................................................................................................................................111
Annex C:
Change history...........................................................................................................112
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 5 Release 9
Foreword
This Technical Report has been produced by the 3
rd
Generation Partnership Project (3GPP).
The contents of the present document are subject to continuing work within the TSG and may change following formal
TSG approval. Should the TSG modify the contents of the present document, it will be re-released by the TSG with an
identifying change of release date and an increase in version number as follows:
Version x.y.z
where:
x the first digit:
1 presented to TSG for information;
2 presented to TSG for approval;
3 or greater indicates TSG approved document under change control.
y the second digit is incremented for all changes of substance, i.e. technical enhancements, corrections,
updates, etc.
z the third digit is incremented when editorial only changes have been incorporated in the document.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 6 Release 9
1 Scope
This document is related to the technical report for physical layer aspect of the study item “Further advancements for E-
UTRA” [1]. The purpose of this TR is to help TSG RAN WG1 to define and describe the potential physical layer
evolution under consideration and compare the benefits of each evolution techniques, along with the complexity
evaluation of each technique.
This activity involves the Radio Access work area of the 3GPP studies and has impacts both on the Mobile Equipment
and Access Network of the 3GPP systems.
This document is intended to gather all information in order to compare the solutions and gains vs. complexity, and
draw a conclusion on way forward.
This document is a ‘living’ document, i.e. it is permanently updated and presented to TSG-RAN meetings.
2 References
The following documents contain provisions which, through reference in this text, constitute provisions of the present
document.
• References are either specific (identified by date of publication, edition number, version number, etc.) or
non-specific.
• For a specific reference, subsequent revisions do not apply.
• For a non-specific reference, the latest version applies. In the case of a reference to a 3GPP document
(including a GSM document), a non-specific reference implicitly refers to the latest version of that document
in the same Release as the present document.
[1] 3GPP TD RP-080137: "Proposed SID on LTE-Advanced".
[2] 3GPP TR 21.905: "Vocabulary for 3GPP Specifications".
[3] 3GPP TR 36.913: “Requirements for Evolved UTRA (E-UTRA) and Evolved UTRAN (E-
UTRAN)
[4] ITU-R Report M.2135, Guidelines for evaluation of radio interface technologies for IMT-
Advanced, 2008-11, http://www.itu.int/publ/R-REP-M.2135-2008/en
[5] 3GPP TS 36.213: “Evolved Universal Terrestrial Radio Access (E-UTRA); Physical layer
procedures (Release 8)”
[6] 3GPP TSG RAN1 R1-094954: "Annex A4 Self-evaluation results for TR36.814".
3 Definitions, symbols and abbreviations
3.1 Definitions
Void
3.2 Symbols
Void
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 7 Release 9
3.3 Abbreviations
For the purposes of the present document, the abbreviations defined in 3GPP TS 21.905 [2] and the following apply:
4 Introduction
At the 3GPP TSG RAN #39 meeting, the Study Item description on "Further Advancements for E-UTRA (LTE-
Advanced)" was approved [1]. The study item covers technology components to be considered for the evolution of E-
UTRA, e.g. to fulfil the requirements on IMT-Advanced. This technical report covers the physical-layer aspects of these
technology components.
5 Support of wider bandwidth
Carrier aggregation, where two or more component carriers are aggregated, is supported by LTE-Advanced in order to
support wider transmission bandwidths e.g. up to 100MHz and for spectrum aggregation..
A terminal may simultaneously receive or transmit one or multiple component carriers depending on its capabilities:
- An LTE-Advanced terminal with reception and/or transmission capabilities for carrier aggregation can
simultaneously receive and/or transmit on multiple component carriers.
- An LTE Rel-8 terminal can receive and transmit on a single component carrier only, provided that the structure
of the component carrier follows the Rel-8 specifications.
- It shall be possible to configure all component carriers LTE Release 8 compatible, at least when the
aggregated numbers of component carriers in the UL and the DL are same. Consideration of non-backward-
compatible configurations of LTE-A component carriers is not precluded
The L1 specification shall support carrier aggregation for both contiguous and non-contiguous component carriers with
each component carrier limited to a maximum of 110 Resource Blocks using the Release 8 numerology
- For contiguous carrier aggregation, the needed frequency spacing between the contiguous component carriers
will be studied by RAN WG4. This study should include the supported number of RBs per component carrier
and the needed guard bands between and at the edges for a certain aggregation case.
- If possible, the same solution should be used in the L1 specifications for contiguous and non-contiguous
aggregation.
It will be possible to configure a UE to aggregate a different number of component carriers of possibly different
bandwidths in the UL and the DL. In typical TDD deployments, the number of component carriers and the bandwidth of
each component carrier in UL and DL will be the same. RAN WG4 will study the supported combinations of
aggregated component carrier and bandwidths.
5.1 MAC-PHY interface
From a UE perspective, there is one transport block (in absence of spatial multiplexing) and one hybrid-ARQ entity per
scheduled component carrier. Each transport block is mapped to a single component carrier only. A UE may be
scheduled over multiple component carriers simultaneously.
5.2 DL control signalling
The design principles for downlink control signalling of control region size, uplink and downlink resource assignments,
and downlink HARQ ACK/NACK indication are described below.
- Independent control region size is applied for each component carrier. On any carrier with a control region, Rel-8
design (modulation, coding, mapping to resource elements) for PCFICH is reused.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 8 Release 9
- For signalling of resource assignments for downlink (PDSCH) and uplink (PUSCH) transmission, following
mechanisms are supported,
- PDCCH on a component carrier assigns PDSCH resources on the same component carrier and PUSCH
resources on a single linked UL component carrier. Rel-8 PDCCH structure (same coding, same CCE-based
resource mapping) and DCI formats are used on each component carrier.
- PDCCH on a component carrier can assign PDSCH or PUSCH resources in one of multiple component
carriers using the carrier indicator field(CIF), where
- Configuration for the presence of CIF is semi-static and UE specific (i.e. not system-specific or cell-
specific)
- CIF is a fixed 3-bit field, and Rel-8 PDCCH structure (same coding, same CCE-based resource mapping)
is reused.
- CIF location is fixed irrespective of DCI format size
- Cross-carrier assignments can be configured both when the DCI formats have the same or different sizes
- Explicit CIF at least for the case of same DCI format size
- There will be an upper limit on the total number of blind decodes
- For signalling of downlink HARQ ACK/NACK indication, following principles are applied.
- PHICH physical transmission aspects from Rel-8 (orthogonal code design, modulation, scrambling sequence,
mapping to resource elements) are reused.
- PHICH is transmitted only on the downlink component carrier that was used to transmit the UL grant
- At least in case that the number of downlink component carriers are more than or equal to that of uplink
component carriers and no carrier indicator field is used, the Rel-8 PHICH resource mapping rule is reused.
5.3 UL control signalling
The design principles for uplink control signalling of HARQ ACK/NACK, scheduling request and channel state
information (CSI) on PUCCH are described below.
- The Rel-10 PUCCH design supports up to five DL component carriers.
- For signalling of HARQ ACK/NACK on PUCCH for downlink (PDSCH) transmission, following mechanisms
are supported:
- All HARQ ACK/NACK for a UE can be transmitted on PUCCH in absence of PUSCH transmission.
- In general, transmission of one ACK/NACK for each DL component carrier transport block is supported.
- In case of power limitation, limited transmission of ACK/NACK for the DL component carrier transport
blocks is supported.
- The design of the ACK/NACK resource allocation should consider performance and power control
aspects, while not aiming to optimise for the case of large number of UEs being simultaneously scheduled
on multiple DL component carriers.
-The scheduling request is transmitted on PUCCH and is semi-statically mapped onto one UE specific UL
component carrier.
-Periodic CSI reporting on PUCCH is supported for up to five DL component carriers. The CSI is semi-statically
mapped onto one UE specific UL component carrier and the design follows the Rel-8 principles for
CQI/PMI/RI, considering ways to reduce reporting overhead or to extend CSI payload.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 9 Release 9
6 Uplink transmission scheme
6.1 Uplink spatial multiplexing
Uplink spatial multiplexing of up to four layers is supported by LTE-Advanced.
In the uplink single user spatial multiplexing, up to two transport blocks can be transmitted from a scheduled UE in a
subframe per uplink component carrier. Each transport block has its own MCS level. Depending on the number of
transmission layers, the modulation symbols associated with each of the transport blocks are mapped onto one or two
layers according to the same principle as in Rel-8 E-UTRA downlink spatial multiplexing. The transmission rank can be
adapted dynamically. It is possible to configure the uplink single user spatial multiplexing transmission with or without
the layer shifting. In case of the layer shifting, shifting in time domain is supported.
If layer shifting is configured, the HARQ-ACKs for all transport blocks are bundled into a single HARQ-ACK. One-bit
ACK is transmitted to the UE if all transport blocks are successfully decoded by the eNodeB. Otherwise, one-bit NACK
is transmitted to the UE.
If layer shifting is not configured, each transport block has its own HARQ-ACK feedback signalling.
For FDD and TDD, precoding is performed according to a predefined codebook. If layer shifting is not configured,
precoding is applied after the layer mapping. If layer shifting is configured, precoding is applied after the layer shifting
operation. Application of a single precoding matrix per uplink component carrier is supported. In case of full-rank
transmission, only identity precoding matrix is supported. For uplink spatial multiplexing with two transmit antennas, 3-
bit precoding codebook as defined in Table 6.1-1 is used.
Table 6.1-1: 3-bit precoding codebook for uplink spatial multiplexing with two transmit antennas
Codebook
index
Number of layers υ
1 2
0
1
]
1

¸

1
1
2
1
1
]
1

¸

1 0
0 1
2
1
1
1
]
1

¸

−1
1
2
1
2
1
]
1

¸

j
1
2
1
3
1
]
1

¸

− j
1
2
1
-
4
1
]
1

¸

0
1
2
1
5
1
]
1

¸

1
0
2
1
For uplink spatial multiplexing with four transmit antennas, a 6-bit precoding codebook is used. The subset of the
precoding codebook used for 1-layer transmission is defined in Table 6.1-2. The baseline for the subset of the precoding
codebook used for 2-layer transmission is defined in Table 6.1-3. For 3-layer transmission, the number of precoding
matrices is 20, and only BPSK or QPSK alphabets are used for non-zero elements in precoding matrices.
Table 6.1-2: 6-bit precoding codebook for uplink spatial multiplexing with four transmit antennas:
precoding matrices for 1-layer transmission.
Codebook
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 10 Release 9
Index
0 to 7
1
1
1
1
]
1

¸

−1
1
1
1
2
1
1
1
1
1
]
1

¸

j
j
1
1
2
1
1
1
1
1
]
1

¸


1
1
1
1
2
1
1
1
1
1
]
1

¸



j
j
1
1
2
1
1
1
1
1
]
1

¸

j
j
1
1
2
1
1
1
1
1
]
1

¸

1
1
2
1
j
j
1
1
1
1
]
1

¸



j
j
1
1
2
1
1
1
1
1
]
1

¸



1
1
2
1
j
j
Index
8 to 15
1
1
1
1
]
1

¸


1
1
1
1
2
1
1
1
1
1
]
1

¸



j
j
1
1
2
1
1
1
1
1
]
1

¸




1
1
1
1
2
1
1
1
1
1
]
1

¸



j
j
1
1
2
1
1
1
1
1
]
1

¸



j
j
1
1
2
1
1
1
1
1
]
1

¸



1
1
2
1
j
j
1
1
1
1
]
1

¸



j
j
1
1
2
1
1
1
1
1
]
1

¸



1
1
2
1
j
j
Index
16 to 23
1
1
1
1
]
1

¸

0
1
0
1
2
1
1
1
1
1
]
1

¸


0
1
0
1
2
1
1
1
1
1
]
1

¸

0
0
1
2
1
j
1
1
1
1
]
1

¸


0
0
1
2
1
j
1
1
1
1
]
1

¸

1
0
1
0
2
1
1
1
1
1
]
1

¸

−1
0
1
0
2
1
1
1
1
1
]
1

¸

j
0
1
0
2
1
1
1
1
1
]
1

¸

− j
0
1
0
2
1
Table 6.1-3: 6-bit precoding codebook for uplink spatial multiplexing with four transmit antennas:
precoding matrices for 2-layer transmission.
Codebook
Index
0 to 7
1
1
1
1
]
1

¸

− j 0
1 0
0 1
0 1
2
1
1
1
1
1
]
1

¸

j 0
1 0
0 1
0 1
2
1
1
1
1
1
]
1

¸


1 0
1 0
0
0 1
2
1
j
1
1
1
1
]
1

¸



1 0
1 0
0
0 1
2
1
j
1
1
1
1
]
1

¸



j 0
1 0
0 1
0 1
2
1
1
1
1
1
]
1

¸


j 0
1 0
0 1
0 1
2
1
1
1
1
1
]
1

¸

1 0
1 0
0
0 1
2
1
j
1
1
1
1
]
1

¸

−1 0
1 0
0
0 1
2
1
j
Index
8 to
15
1
1
1
1
]
1

¸

1 0
0 1
1 0
0 1
2
1
1
1
1
1
]
1

¸

−1 0
0 1
1 0
0 1
2
1
1
1
1
1
]
1

¸


1 0
0 1
1 0
0 1
2
1
1
1
1
1
]
1

¸



1 0
0 1
1 0
0 1
2
1
1
1
1
1
]
1

¸

0 1
1 0
1 0
0 1
2
1
1
1
1
1
]
1

¸


0 1
1 0
1 0
0 1
2
1
1
1
1
1
]
1

¸

− 0 1
1 0
1 0
0 1
2
1
1
1
1
1
]
1

¸



0 1
1 0
1 0
0 1
2
1
6.2 Uplink transmit diversity
For UEs with multiple transmit antennas, an uplink Single Antenna Port Mode is defined, where the UE behaviour is
same as the one with single antenna from eNodeB’s perspective. For a given UE, the uplink Single Antenna Port Mode
can be independently configured for its PUCCH, PUSCH and SRS transmissions.
The uplink Single Antenna Port Mode is the default mode before eNodeB is aware of the UE transmit antenna
configuration.
6.2.1 Transmit Diversity for Uplink Control Channel
For uplink control channels with Rel-8 PUCCH format 1/1a/1b, the spatial orthogonal-resource transmit diversity
(SORTD) scheme is supported for transmissions with two antenna ports. In this transmit diversity scheme, the same
modulation symbol from the uplink channel is transmitted from two antenna ports, on two separate orthogonal
resources.
For the UE with four transmit antennas, the 2-tx transmit diversity scheme is applied.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 11 Release 9
6.3 Uplink multiple access
DFT-precoded OFDM is the transmission scheme used for PUSCH both in absence and presence of spatial
multiplexing. In case of multiple component carriers, there is one DFT per component carrier. Both frequency-
contiguous and frequency-non-contiguous resource allocation is supported on each component carrier.
Simultaneous transmission of uplink L1/L2 control signalling and data is supported through two mechanisms
- Control signalling is multiplexed with data on PUSCH according to the same principle as in Rel-8
- Control signalling is transmitted on PUCCH simultaneously with data on PUSCH
6.4 Uplink reference signals
Similar to LTE, two types of uplink reference signals are supported by LTE-Advanced:
- Demodulation reference signal
- Sounding reference signal
The precoding applied for the demodulation reference signal is the same as the one applied for the PUSCH. Cyclic shift
separation is the primary multiplexing scheme of the demodulation reference signals.
The baseline for sounding reference signal in LTE-Advanced operation is non-precoded and antenna-specific. For
multiplexing of the sounding reference signals, Rel-8 principles are reused.
6.5 Uplink power control
Scope of uplink power control in LTE-Advanced is similar to Rel’8:
- UL power control mainly compensates for slow-varying channel conditions while reducing the interference
generated towards neighboring cells
- Fractional path-loss compensation or full path-loss compensation is used on PUSCH and full path-loss
compensation on PUCCH
LTE-Advanced supports component carrier specific UL power control for both contiguous and non-contiguous carrier
aggregation for closed-loop case, and for open loop at least for the cases that the number of downlink component
carriers is more than or equal to that of uplink component carriers.
7 Downlink transmission scheme
7.1 Physical channel mapping
LTE-Advanced supports the PDSCH to be mapped also to MBSFN (non-control) region of MBSFN subframes that are
not used for MBMS
- In case of PDSCH mapping to MBSFN subframes, both normal and extended cyclic prefix can be used for
control and data region, same CP length is used for control and data
- Relation between CP length of normal and MBSFN subframes in the control region is the same as in Rel-8
7.2 Downlink spatial multiplexing
Downlink spatial multiplexing of up to eight layers is supported byLTE-Advanced.
In the downlink 8-by-X single user spatial multiplexing, up to two transport blocks can be transmitted to a scheduled
UE in a subframe per downlink component carrier. Each transport block is assigned its own modulation and coding
scheme. For HARQ ACK/NAK feedback on uplink, one bit is used for each transport block.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 12 Release 9
A transport block is associated with a codeword. For up to four layers, the codeword-to-layer mapping is according to
section 6.3.3.2 of TS 36.211. For above four layers as well as the case of mapping one codeword to three or four layers,
which is for retransmission of one out of two codewords that were initially transmitted with more than four layers, the
layer mapping shall be done according to Table 7.2-1.. Complex-valued modulation symbols
) 1 ( ),..., 0 (
(q)
symb
) ( ) (
− M d d
q q
for code word
q
shall be mapped onto the layers
[ ]
T
i x i x i x ) ( ... ) ( ) (
) 1 ( ) 0 ( −
·
υ
,
1 ,..., 1 , 0
layer
symb
− · M i
where υ is the number of layers and
layer
symb
M
is
the number of modulation symbols per layer.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 13 Release 9
Table 7.2-1: Codeword-to-layer mapping for above four layers and the case of mapping one
codeword to three or four layers
Number of layers Number of code
words
Codeword-to-layer mapping
1 ,..., 1 , 0
layer
symb
− · M i
3 1
) 2 3 ( ) (
) 1 3 ( ) (
) 3 ( ) (
) 0 ( ) 2 (
) 0 ( ) 1 (
) 0 ( ) 0 (
+ ·
+ ·
·
i d i x
i d i x
i d i x
3
) 0 (
symb
layer
symb
M M ·
4 1
) 3 4 ( ) (
) 2 4 ( ) (
) 1 4 ( ) (
) 4 ( ) (
) 0 ( ) 3 (
) 0 ( ) 2 (
) 0 ( ) 1 (
) 0 ( ) 0 (
+ ·
+ ·
+ ·
·
i d i x
i d i x
i d i x
i d i x
4
) 0 (
symb
layer
symb
M M ·
5 2
) 1 2 ( ) (
) 2 ( ) (
) 0 ( ) 1 (
) 0 ( ) 0 (
+ ·
·
i d i x
i d i x
3 2
) 1 (
symb
) 0 (
symb
layer
symb
M M M · ·
) 2 3 ( ) (
) 1 3 ( ) (
) 3 ( ) (
) 1 ( ) 4 (
) 1 ( ) 3 (
) 1 ( ) 2 (
+ ·
+ ·
·
i d i x
i d i x
i d i x
6 2
) 2 3 ( ) (
) 1 3 ( ) (
) 3 ( ) (
) 0 ( ) 2 (
) 0 ( ) 1 (
) 0 ( ) 0 (
+ ·
+ ·
·
i d i x
i d i x
i d i x
3 3
) 1 (
symb
) 0 (
symb
layer
symb
M M M · ·
) 2 3 ( ) (
) 1 3 ( ) (
) 3 ( ) (
) 1 ( ) 5 (
) 1 ( ) 4 (
) 1 ( ) 3 (
+ ·
+ ·
·
i d i x
i d i x
i d i x
7 2
) 2 3 ( ) (
) 1 3 ( ) (
) 3 ( ) (
) 0 ( ) 2 (
) 0 ( ) 1 (
) 0 ( ) 0 (
+ ·
+ ·
·
i d i x
i d i x
i d i x
4 3
) 1 (
symb
) 0 (
symb
layer
symb
M M M · ·
) 3 4 ( ) (
) 2 4 ( ) (
) 1 4 ( ) (
) 4 ( ) (
) 1 ( ) 6 (
) 1 ( ) 5 (
) 1 ( ) 4 (
) 1 ( ) 3 (
+ ·
+ ·
+ ·
·
i d i x
i d i x
i d i x
i d i x
8 2
) 3 4 ( ) (
) 2 4 ( ) (
) 1 4 ( ) (
) 4 ( ) (
) 0 ( ) 3 (
) 0 ( ) 2 (
) 0 ( ) 1 (
) 0 ( ) 0 (
+ ·
+ ·
+ ·
·
i d i x
i d i x
i d i x
i d i x
4 4
) 1 (
symb
) 0 (
symb
layer
symb
M M M · ·
) 3 4 ( ) (
) 2 4 ( ) (
) 1 4 ( ) (
) 4 ( ) (
) 1 ( ) 7 (
) 1 ( ) 6 (
) 1 ( ) 5 (
) 1 ( ) 4 (
+ ·
+ ·
+ ·
·
i d i x
i d i x
i d i x
i d i x
7.2.1 Feedback in support of downlink spatial multiplexing
The baseline for feedback in support of downlink single-cell single-user spatial multiplexing is codebook-based
precoding feedback.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 14 Release 9
7.3 Enhanced multiuser MIMO transmission
Enhanced multi-user MIMO transmission is supported in LTE-Advanced. Basic principle of MU-MIMO in LTE-
Advanced is as follows:
- Switching between SU- and MU-MIMO transmission is possible without RRC reconfiguration
7.4 Downlink reference signals
Two types of downlink reference signals are considered for LTE-Advanced:
- Reference signals targeting PDSCH demodulation
- Reference signals targeting CSI estimation (for CQI/PMI/RI/etc reporting when needed)
The reference signal structure can be used to support multiple LTE-Advanced features, e.g. CoMP and spatial
multiplexing.
7.4.1 Demodulation reference signal
Reference signals targeting PDSCH demodulation are:
- UE-specific, i.e, the PDSCH and the demodulation reference signals intended for a specific UE are subject to the
same precoding operation.
- Present only in resource blocks and layers scheduled by the eNodeB for transmission.
- Reference signals transmitted on different layers are mutually orthogonal at the eNodeB.
The design principle for the demodulation reference signals is an extension to multiple layers of the concept of Rel-8
UE-specific reference signals used for beamforming (details on pattern, location, etc are FFS). Complementary use of
Rel-8 cell-specific reference signals by the UE is not precluded.
Normal subframes with normal CP:
For rank 1and 2, the same DM-RS structure (including patterns, spreading and scrambling) as in LTE Rel-9 is used, as
illustrated in Figure 7.3.1-1. For rank 2, DM-RS for 1st layer and that for 2nd layer are multiplexed by means of code
division multiplexing (CDM) by using orthogonal cover code (OCC) over two consecutive resource elements (blue) in
time domain.
DM-RS CRS
time
f
r
e
q
u
e
n
c
y
Figure 7.4.1-1: DM-RS pattern for rank 1 and 2
For rank 3 and 4, the DM-RS pattern is illustrated in Figure 7.3.1-2. DM-RS for 1st layer and that for 2nd layer are
multiplexed by means of CDM by using OCC over two consecutive resource elements (blue) in time domain. DM-RS
for 3rd layer and that for 4th layer are multiplexed by means of CDM by using OCC over two consecutive resource
elements (green) in time domain. DM-RS for 1st and 2nd layers and that for 3rd and 4th layers are multiplexed by
means of frequency division multiplexing (FDM).
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 15 Release 9
DM-RS
CRS
time
f
r
e
q
u
e
n
c
y
Figure 7.4.1-2: DM-RS pattern for rank 3 and 4
7.4.2 CSI reference signal
Reference signals targeting CSI estimation are
- cell specific
- sparse in frequency and time.
- punctured into the data region of normal/MBSFN subframe
7.5 Downlink transmit diversity
For the downlink transmit diversity with more than four transmit antennas applied to PDCCH, and PDSCH in non-
MBSFN subframes, the Rel-8 transmit diversity scheme is used.
8 Coordinated multiple point transmission and
reception
Coordinated multi-point (CoMP) transmission/reception is considered for LTE-Advanced as a tool to improve the
coverage of high data rates, the cell-edge throughput and/or to increase system throughput in both high load and low
load scenarios.
8.1 Downlink coordinated multi-point transmission
8.1.1 Terminology and definitions
Downlink coordinated multi-point transmission implies dynamic coordination among multiple geographically separated
transmission points.
General terminology:
- Serving cell: Cell transmitting PDCCH assignments (a single cell). This is the serving cell of Rel-8 (concept that
already exists)
CoMP categories:
- Joint Processing (JP): data is available at each point in CoMP cooperating set (definition below)
- Joint Transmission: PDSCH transmission from multiple points (part of or entire CoMP cooperating set) at a
time
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 16 Release 9
- data to a single UE is simultaneously transmitted from multiple transmission points, e.g. to (coherently or
non-coherently) improve the received signal quality and/or cancel actively interference for other UEs
- Dynamic cell selection: PDSCH transmission from one point at a time (within CoMP cooperating set)
- Coordinated Scheduling/Beamforming (CS/CB): data is only available at serving cell (data transmission from
that point) but user scheduling/beamforming decisions are made with coordination among cells corresponding to
the CoMP cooperating set.
CoMP sets:
- CoMP cooperating set
- Set of (geographically separated) points directly or indirectly participating in PDSCH transmission to UE.
Note that this set may or need not be transparent to the UE.
- CoMP transmission point(s): point or set of points actively transmitting PDSCH to UE
- CoMP transmission point(s) is a subset of the CoMP cooperating set
- For Joint transmission, the CoMP transmission points are the points in the CoMP cooperating set
- For Dynamic cell selection, a single point is the transmission point at every subframe. This transmission
point can change dynamically within the CoMP cooperating set.
- For Coordinated scheduling/beamforming, the CoMP transmission point corresponds to the “serving
cell”
- CoMP measurement set: set of cells about which channel state/statistical information related to their link to the
UE is reported as discussed in section 8.1.3
- The CoMP measurement set may be the same as the CoMP cooperating set
- The actual UE reports may down-select cells for which actual feedback information is transmitted (reported
cells)
In addition, we have:
RRM measurement set: in support of RRM measuremets (already in Rel-8) and therefore not CoMP specific
8.1.2 Radio-interface specification areas
Downlink coordinated multi-point transmission should include the possibility of coordination between different cells.
From a radio-interface perspective, there is no difference from the UE perspective if the cells belong to the same
eNodeB or different eNodeBs. If inter-eNodeB coordination is supported, information needs to be signalled between
eNodeBs.
Potential impact on the radio-interface specifications is foreseen in mainly three areas:
- Feedback and measurement mechanisms from the UE
- Reporting of dynamic channel conditions between the multiple transmission points and the UE
- For TDD, channel reciprocity may be exploited
- Reporting to facilitate the decision on the set of participating transmission points
- For TDD, channel reciprocity may be exploited
- Preprocessing schemes
- Joint processing prior to transmission of the signal over the multiple transmission points
- Downlink control signaling to support the transmission scheme
- Reference signal design
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 17 Release 9
- Depending on the transmission scheme, specification of additional reference signals may be required.
- CSI RS design should take potential needs of DL CoMP into account
New forms of feedback and signaling may be needed to support CoMP that are, for example, configured by RRC for a
given UE.
- As baseline, the network need not explicitly signal to the UE the CoMP transmission point(s) and the UE
reception/demodulation of CoMP transmissions (CS/CB, or JP with MBSFN subframes) is the same as that for
non CoMP (SU/MU-MIMO).”
- Any additional feedback designed for CoMP shall be consistent with the feedback framework for SU/MU-MIMO.
8.1.3 Feedback in support of DL CoMP
The three main categories of CoMP feedback mechanisms have been identified to be:
- Explicit channel state/statistical information feedback
- Channel as observed by the receiver, without assuming any transmission or receiver processing
- Implicit channel state/statistical information feedback
- feedback mechanisms that use hypotheses of different transmission and/or reception processing, e.g.,
CQI/PMI/RI
- UE transmission of SRS can be used for CSI estimation at eNB exploiting channel reciprocity.
Combinations of full or subset of above three are possible.
Look at these types of feedback mechanisms for the evaluations. UL overhead (number of bits) associated with each
specific feedback mechanism needs to be identified. The feedback overhead (UL) vs, DL performance tradeoff should
be assessed with the goal to target minimum overhead for a given performance.
For the CoMP schemes that require feedback, individual per-cell feedback is considered as baseline. Complementary
inter-cell feedback might be needed. UE CoMP feedback reports target the serving cell (on UL resources from serving
cell) as baseline when X2 interface is available and is adequate for CoMP operation in terms of latency and capacity. In
this case, the reception of UE reports at cells other than the serving cell is a network implementation choice.
The feedback reporting for cases with X2 interface not available or not adequate (latency and capacity), and for cases
where feedback reports to the serving cell causes large interference (e.g., in heteronegenous deployment scenarios) for
CoMP operation needs to be discussed and, if found needed, a solution needs to be identified.
Do not have to confine the CoMP studies to payload sizes currently supported by PUCCH operation..
Two possibilities should be studied the “container” of the DL CoMP feedback:
- Expand the supported PUCCH payload sizes
- Use periodic/a-periodic reports on PUSCH
8.1.3.1 Explicit Feedback in support of DL CoMP
This section lists different forms of explicit feedback in support of DL CoMP. They are all characterized by having a
channel part and a noise-and-interference part.
Channel part:
- For each cell in the UE’s measurement set that is reported in a given subframe, one or several channel properties
are reported
- Channel properties include (but are not limited to) the following (‘i‘ is the cell index):
- Channel matrix (Hi) – short term (instantaneous)
- The full matrix Hi, or
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 18 Release 9
- main eigen component(s) of Hi
- Transmit channel covariance (Ri), where Ri = (sum{Hij†Hij})/J, j=0,1,2,…,J-1, (‘j’ is span over time or
frequency)
- The full matrix Ri, or
- main eigen component(s) of Ri
- Inter-cell channel properties may also be reported
Noise-and interference part, e.g.,
- Interference outside the
- cells reported by the UE
- CoMP transmission points
- Total receive power (Io) or total received signal covariance matrix
- Covariance matrix of the noise-and-interference
- the full matrix, or
- main eigen component(s)
8.1.3.2 Implicit Feedback in support of DL CoMP
This section lists different forms of implicit feedback in support of DL CoMP.
- There are hypotheses at the UE and the feedback is based on one or a combination of two or more of the
following, e.g.:
- Single vs. Multi user MIMO
- Single cell vs. Coordinated transmission
- Within coordinated transmission: Single point (CB/CS) vs. multi-point (JP) transmission
Within Joint processing CoMP:
- Subsets of transmission points or subsets of reported cells (Joint Transmission)
- CoMP transmission point(s) (Dynamic Cell Selection)
- Transmit precoder (i.e. tx weights)
- JP: multiple single-cell or multi-cell PMI capturing coherent or non-coherent channel across reported
cells
- CB/CS: Single-cell or multiple single-cell PMIs capturing channel from the reported cell(s) to the UE
- Transmit precoder based on or derived from the PMI weight
- Other types of feedbacks, e.g. main Multi-cell eigen-component, instead of PMI are being considered
- Receive processing (i.e. rx weights)
- Interference based on particular tx/rx processing
There may be a need for the UE to convey to the network the hypothesis or hypotheses used (explicit signalling of
hypothesis to eNB). And/or, there may be a semi-static hypothesis configuration e.g. grouping of hypotheses (explicit
signalling of hypothesis to the UE). And/or, precoded RS may be used to allow UE to generate refined CQI/RI feedback
8.1.3.3 UE transmission of SRS in support of DL CoMP
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 19 Release 9
This section lists UE transmission of SRS related feedback in support of DL CoMP. UE transmission of SRS can be
used for CSI estimation at multiple cells exploiting channel reciprocity. Enhanced SRS schemes may be considered.
8.1.4 Overhead in support of DL CoMP operation
DL CoMP operation overhead is very much related to the DL-RS structure (see section 7.2).
- Studies on CSI RS impact on PDSCH transmissions to Rel-8 UEs for various RS densities needed
- There should be no impact from CSI RS transmission on transmission of PBCH/PSS/SSS
8.2 Uplink coordinated multi-point reception
Coordinated multi-point reception implies coordination among multiple, geographically separated points. Uplink
coordinated multi-point reception is expected to have very limited impact on the RAN1 specifications. Uplink CoMP
reception can involve joint reception (JR) of the transmitted signal at multiple reception points and/or coordinated
scheduling (CS) decisions among cells to control interference and may have some RAN1 specification impact.
The need for extended CP operation in certain UL subframes should be further investigated.
9 Relaying functionality
Relaying is considered for LTE-Advanced as a tool to improve e.g. the coverage of high data rates, group mobility,
temporary network deployment, the cell-edge throughput and/or to provide coverage in new areas.
The relay node is wirelessly connected to the radio-access network via a donor cell. With respect to the relay node’s
usage of spectrum, its operation can be classified into:
- inband, in which case the eNB -relay link shares the same carrier frequency with relay-UE links. Rel-8 UEs
should be able to connect to the donor cell in this case.
- outband, in which case the eNB-relay link does not operate in the same carrier frequency as relay-UE links. Rel-
8 UEs should be able to connect to the donor cell in this case.
For both inband and outband relaying, it shall be possible to operate the eNB-to-relay link on the same carrier frequency
as eNB-to-UE links
With respect to the knowledge in the UE, relays can be classified into
- transparent, in which case the UE is not aware of whether or not it communicates with the network via the relay.
- non-transparent, in which case the UE is aware of whether or not it is communicating with the network via the
relay.
Depending on the relaying strategy, a relay may
- be part of the donor cell
- control cells of its own
In the case the relay is part of the donor cell, the relay does not have a cell identity of its own (but may still have a relay
ID). At least part of the RRM is controlled by the eNodeB to which the donor cell belongs, while parts of the RRM may
be located in the relay. In this case, a relay should preferably support also LTE Rel-8 UEs. Smart repeaters, decode-and-
forward relays, different types of L2 relays, and Type 2 relay are examples of this type of relaying.
In the case the relay is in control of cells of its own, the relay controls one or several cells and a unique physical-layer
cell identity is provided in each of the cells controlled by the relay. The same RRM mechanisms are available and from
a UE perspective there is no difference in accessing cells controlled by a relay and cells controlled by a “normal”
eNodeB. The cells controlled by the relay should support also LTE Rel-8 UEs. Self-backhauling (L3 relay), “Type 1
relay nodes”, “Type 1a relay nodes”, and “Type 1b relay nodes” use this type of relaying.
At least “Type 1” and “Type 1a” relay nodes are part of LTE-Advanced.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 20 Release 9
A “Type 1” relay node is an inband relaying node characterized by the following:
- It control cells, each of which appears to a UE as a separate cell distinct from the donor cell
- The cells shall have their own Physical Cell ID (defined in LTE Rel-8) and the relay node shall transmit its own
synchronization channels, reference symbols, …
- In the context of single-cell operation, the UE shall receive scheduling information and HARQ feedback directly
from the relay node and send its control channels (SR/CQI/ACK) to the relay node
- It shall appear as a Rel-8 eNodeB to Rel-8 UEs (i.e. be backwards compatible)
- To LTE-Advanced UEs, it should be possible for a relay node to appear differently than Rel-8 eNodeB to allow
for further performance enhancement.
“Type 1a” and “Type 1b” relay nodes are characterised by the same set of features as the “Type 1” relay node above,
except “Type 1a” operates outband and “Type 1b” operates inband with adequate antenna isolation. A “Type 1a” relay
node is expected to have little or no impact on RAN1 specifications.
A “Type 2” relay node is an inband relaying node characterized by the following:
- It does not have a separate Physical Cell ID and thus would not create any new cells.
- It is transparent to Rel-8 UEs; a Rel-8 UE is not aware of the presence of a Type 2 relay node.
- It can transmit PDSCH.
- At least, it does not transmit CRS and PDCCH.
9.1 Relay-eNodeB link for inband relay
9.1.1 Resource partitioning for relay-eNodeB link
In order to allow inband relaying, some resources in the time-frequency space are set aside for the backhaul link (Un)
and cannot be used for the access link (Uu). At least the following scheme will be supported for this resource
partitioning:
General principle for resource partitioning at the relay:
- eNB → RN and RN → UE links are time division multiplexed in a single carrier frequency (only one is active at
any time)
- RN → eNB and UE → RN links are time division multiplexed in a single carrier frequency (only one is active at
any time)
Multiplexing of backhaul links in FDD:
- eNB → RN transmissions are done in the DL frequency band
- RN → eNB transmissions are done in the UL frequency band
Multiplexing of backhaul links in TDD:
- eNB → RN transmissions are done in the DL subframes of the eNB and RN
- RN → eNB transmissions are done in the UL subframes of the eNB and RN
9.1.2 Backward compatible backhaul partitioning
Due to the relay transmitter causing interference to its own receiver, simultaneous eNodeB-to-relay and relay-to-UE
transmissions on the same frequency resource may not be feasible unless sufficient isolation of the outgoing and
incoming signals is provided e.g. by means of specific, well separated and well isolated antenna structures. Similarly, at
the relay it may not be possible to receive UE transmissions simultaneously with the relay transmitting to the eNodeB.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 21 Release 9
One possibility to handle the interference problem is to operate the relay such that the relay is not transmitting to
terminals when it is supposed to receive data from the donor eNodeB, i.e. to create “gaps” in the relay-to-UE
transmission. These “gaps” during which terminals (including Rel-8 terminals) are not supposed to expect any relay
transmission can be created by configuring MBSFN subframes as exemplified in Figure 9.1.2-1. Relay-to-eNodeB
transmissions can be facilitated by not allowing any terminal-to-relay transmissions in some subframes.

Data Ctrl
transmission gap
(“MBSFN subframe”)
Ctrl
One subframe
No relay-to-UE transmission
eNB-to-relay transmission
Figure 9.1.2-1: Example of relay-to-UE communication using normal subframes (left) and eNodeB-to-
relay communication using MBSFN subframes (right).
9.1.3 Backhaul resource assignment
In case of downlink backhaul in downlink resources, the following is valid
- At the RN, the access link downlink subframe boundary is aligned with the backhaul link downlink subframe
boundary, except for possible adjustment to allow for RN transmit/receive switching
- The set of downlink backhaul subframes, during which downlink backhaul transmission may occur, is semi-
statically assigned
- The set of uplink backhaul subframes, during which uplink backhaul transmission may occur, can be semi-
statically assigned, or implicitly derived from the downlink backhaul subframes using the HARQ timing
relationship
- A new physical control channel (here referred to as the “R-PDCCH”) is used to dynamically or “semi-
persistently” assign resources, within the semi-statically assigned sub-frames, for the downlink backhaul data
(corresponding to the “R-PDSCH” physical channel). The R-PDCCH may assign downlink resources in the
same and/or in one or more later subframes.
- The “R-PDCCH” is also used to dynamically or “semi-persistently” assign resources for the uplink backhaul
data (the “R-PUSCH” physical channel). The R-PDCCH may assign uplink resources in one or more later
subframes.
- Within the PRBs semi-statically assigned for R-PDCCH transmission, a subset of the resources is used for each
R-PDCCH. The actual overall set of resources used for R-PDCCH transmission within the above mentioned
semi-statically assigned PRBs may vary dynamically between subframes. These resources may correspond to the
full set of OFDM symbols available for the backhaul link or be constrained to a subset of these OFDM symbols.
The resources that are not used for R-PDCCH within the above mentioned semi-statically assigned PRBs may be
used to carry R-PDSCH or PDSCH.
- The detailed R-PDCCH transmitter processing (channel coding, interleaving, multiplexing, etc.) should reuse
Rel-8 functionality to the extent possible, but allow removing some unnecessary procedure or bandwidth-
wasting procedure by considering the relay property.
- If the “search space” approach of R8 is used for the backhaul link, use of common search space, which can be
semi-statically configured (and potentially includes entire system bandwidth), is the baseline. If RN-specific
search space is configured, it could be implicitly or explicitly known by RN.
- The R-PDCCH is transmitted starting from an OFDM symbol within the subframe that is late enough so that the
relay can receive it.
- “R-PDSCH” and “R-PDCCH” can be transmitted within the same PRBs or within separated PRBs.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 22 Release 9
- No “R-PCFICH” exists for indication of backhaul control region size in time and frequency domains
9.2 Relay-eNodeB link for outband relay
If relay-eNB and relay-UE links are isolated enough in frequency (possibly with help of additional means such as
antenna separation), then there is no interference issue in activating both links simultaneously. Therefore, it becomes
possible for relay-eNodeB link to reuse the channels designed for UE-eNodeB link.
9.3 Relay-eNodeB link for inband relay Type 1b
If the outgoing and incoming signals at the relay are adequately isolated in the spatial domain, e.g., by appropriate
arrangement of the respective antennas for the Un and Uu links, the eNB→ RN and RN→ UE (RN→ eNB and UE→
RN) links can be activated simultaneously without the need for the time division multiplexing. The operation of Type
1b relay nodes may not be supported in all deployment scenarios.
9A Heterogeneous deployments
Heterogeneous deployments consist of deployments where low power nodes are placed throughout a macro-cell layout.
Different scenarios under consideration that could occur in heterogeneous deployments are discussed in Appendix
A.2.1.1.2
The interference characteristics in a heterogeneous deployment can be significantly different than in a homogeneous
deployment. Examples hereof are given in Figure 9A-1. In case (a), a macro user with no access to the CSG cell will be
interfered by the HeNB, in case (b) a macro user causes severe interference towards the HeNB and in case (c), a CSG
user is interfered by another CSG HeNB. On the right hand side, case (d), path-loss based cell association (e.g. by using
biased RSRP reports) may improve the uplink but at the cost of increasing the downlink interference of non-macro
users at the cell edge.

CSG
Rx power (solid, dotted->macro), 1/pathloss (dashed)
Inter-cell interference (dotted)

(b)
(d)
CSG
(a)
(c)
CSG
CSG
Rx power (solid, dotted->macro), 1/pathloss (dashed)
Inter-cell interference (dotted)

(b)
(d)
CSG
(a)
(c)
CSG
Figure 9A-1: Examples of interference scenarios in heterogeneous deployments.
In these scenarios, preliminary results indicate that methods for handling the uplink and downlink interference towards
data as well as L1/L2 control signaling, synchronization signals and reference signals are important. Such methods may
operate in time, frequency and/or spatial domains.
9A.1 Rel-8/9 schemes
Several methods for handling the above-mentioned interference scenarios using the functionality present in Rel-8/9 can
be envisioned. Examples hereof include using different carrier frequencies for different cell layers, or, if the same
carrier is used across different cell layers, by restricting the transmission power during part of the time for at least one
cell layer to reduce interference to control signalling on other cell layers or through different power control schemes as
discussed in [TR36.921]. Enhancements to these types of schemes can be considered in releases beyond Rel-9.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 23 Release 9
9A.2 Non-Rel-8/9 schemes
A common property of the non-Rel-8/9 schemes below is that they provide means for coordination of the control
channel interference between cell layers. Examples of some mechanism that might be used for interference handling can
be found in Section 8.
9A.2.1 CA-based scheme
Carrier aggregation (CA) with cross-carrier scheduling using CIF , described in Section 5.2 and agreed to be part of
Rel-10, can be used for heterogeneous deployments. Downlink interference for control signaling can be handled by
partitioning component carriers in each cell layer into two sets, one set used for data and control and one set used
mainly for data and possibly control signaling with reduced transmission power. One example is illustrated in
Figure 9A.2.1-1. For the data part, downlink interference coordination techniques can be used. Rel-8/9 terminals can be
scheduled on one component carrier while Rel-10 terminal capable of carrier aggregation can be scheduled on multiple
component carriers. Time synchronization between the cell layers is assumed in this example.

Macro
Pico
f
1
f
2
f
1
f
2
f
1
f
2
f
1
f
2
Macro UE
• Control signaling on f
1
• Data on f
1
and/or f
2
Pico UE
• Control signaling on f
2
• Data on f
1
and/or f
2
Macro UE
• Control signaling on f
1
and/or f
2
• Data on f
1
and/or f
2
Figure 9A.2.1-1: One example of carrier aggregation applies to heterogeneous deployments.
9A.2.2 Non-CA based schemes
Several schemes for control-channel interference handling not relying on carrier aggregation can be envisioned,
including, but not limited to, time-domain and frequency-domain schemes.
In a time-domain scheme, interference handling for downlink control signalling can be achieved by restricting the
transmission during part of the time, possibly in combination with time shifting, for at least one cell layer to reduce
interference to control signalling on other cell layers. Time synchronization between the cell layers is assumed in this
approach.
In a frequency-domain scheme, the downlink control signalling is transmitted over part of the carrier bandwidth.
Interference management can be achieved by using different parts of the carrier bandwidth for control signalling in
different cell layers. Time synchronization between the cell layers is assumed.
10 Evaluation of techniques for Advanced E-UTRA
Section 10.1 presents the evaluation results of the techniques for advanced E-UTRA against the 3GPP targets [3]. The
evaluation results in Section 10.2 are in support of the submission of the 3GPP "LTE Release 10 & beyond (LTE-
Advanced)" to the ITU-R as a candidate technology for the IMT-Advanced.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 24 Release 9
10.1 Cell spectral efficiency and cell-edge spectral efficiency
against 3GPP target
Cell spectral efficiency and cell-edge spectral efficiency are evaluated through extensive simulations conducted by a
number of companies. The tables in the following subsections show the simulation results focusing on LTE-Advanced
configurations, i.e., downlink and uplink MU-MIMO, downlink and uplink CoMP, and uplink SU-MIMO.
Tables 10.1-1, 2, 3, and 4 show the source specific simulation parameters that characterize the performances of
downlink FDD, downlink TDD, uplink FDD, and uplink TDD, respectively. Note, however, that the performance
differences among sources could also be explained by other factors such as detailed signal processing algorithms at the
transmitter and the receiver.
In the tables for downlink, control channel (CCH) duration (L OFDM symbols) and the number of the MBSFN
subframes are the factors that affect the overhead. In the MBSFN subframes, there are no CRS in data regions, which
effectively reduce the overhead. Channel estimation and receiver types are the factors that affect the demodulation
performance. CSI-assumption at eNB is the factor that characterizes the transmit signal processing at eNB for MU-
MIMO and CoMP schemes. In the tables for uplink, the PUCCH bandwidth is the factor that affects the overhead.
Table 10.1-1 Simulation parameters for 3GPP targets fulfillment (DL, FDD)
 
CCH duration
(L symbols)
Num of MBSFN
subframe
Channel
estimation
Receiver
type
CSI assumption
at eNB
Source 1 3 6 Real MMSE
Short-term
for CS/CB-CoMP
Long-term +
Short-term
for JP-CoMP
Source 3 3 6 Real MMSE with IRC Long-term
Source 4 3 6 Real MMSE with IRC Short-term
Source 5 3 6 Real MMSE
Long-term +
Short-term
Source 7 3 6 Ideal MMSE
Long-term +
Short-term
Source 11 3 6 Real MMSE with IRC Short-term
Source 15 3 6 Real MMSE with IRC Short-term
Source 18 3 6 Real MMSE with IRC Long-term
Table 10.1-2 Simulation parameters for 3GPP targets fulfillment (DL, TDD)
  CCH duration
(L symbols)
Num of MBSFN
subframe
Channel
estimation
Receiver
type
CSI assumption
at eNB
Source 1
3 2 Real MMSE Short-term
Source 2
3 0 Real MMSE with IRC Short-term
Source 3
3 2 Real MMSE with IRC Long-term
Source 4
3 2 Real MMSE with IRC Short-term
Source 5
3 0 Real MMSE Short-term
Source 10
3 2 Real MMSE with IRC Short-term
Source 19
3 2 Real MMSE with IRC Short-term
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 25 Release 9
Table 10.1-3 Simulation parameters for 3GPP targets fulfillment (UL, FDD)
 
PUCCH
bandwidth
(RB/10 MHz)
Channel
estimation
Receiver type
Source 1
6 Real MMSE
Source 3
4 Real MMSE with IRC
Source 4
4 Real MMSE with IRC
Source 5
4 Real MMSE
Source 15
4 Real MMSE with IRC
Source 18
6 Real MMSE with IRC
Table 10.1-4 Simulation parameters for 3GPP targets fulfillment (UL, TDD)
 
PUCCH
bandwidth
(RB/10 MHz)
Channel
estimation
Receiver type
Source 1 4 Real MMSE
Source 2
4 Real MMSE with IRC
Source 3
4 Real MMSE with IRC
Source 4
4 Real MMSE with IRC
Source 5
4 Real MMSE with IRC
Source 10
4 Real MMSE with IRC
Tables in the following subsections capture the spectrum efficiency results from individual sources. As performance
references, the 3GPP targets and the averaged results of Rel-8 schemes (2-by-2/4-by-2/4-by-4 SU-MIMO with L=3 for
downlink, and 1-by-2/1-by-4 SIMO, and 1-by-4 MU-MIMO for uplink) assuming real channel estimation.
10.1.1 3GPP Case1 (3GPP spatial channel model)
10.1.1.1 FDD, Downlink
Tables 10.1.1.1-1, 10.1.1.1-2, and 10.1.1.1-3 show the spectrum efficiency results of 2-by-2, 4-by-2, and 4-by-4 MU-
MIMO schemes with eNB antenna configuration (C). Results show that the MU-MIMO schemes satisfy the 3GPP
target and provide significant gain compared with the Rel-8 SU-MIMO scheme.
Table 10.1.1.1-1 Performance of DL MU-MIMO 2 x 2 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.77 0.110
Source 5 2.74 0.091
Source 15 2.56 0.070
Rel-8 SU-MIMO
(2 x 2, L=3)
2.23 0.079
3GPP target 2.40 0.070
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 26 Release 9
Table 10.1.1.1-2 Performance of DL MU-MIMO 4 x 2 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.55 0.120
Source 4
3.23 0.118
Source 5
3.41 0.127
Source 7
3.83 0.123
Source 11
3.15 0.104
Source 15
3.42 0.114
Rel-8 SU-MIMO
(4 x 2, L=3)
2.53 0.100
3GPP target 2.60 0.090
Table 10.1.1.1-3 Performance of DL MU-MIMO 4 x 4 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 5.45 0.210
Source 4
4.49 0.205
Source 5
4.42 0.223
Source 11
4.63 0.196
Source 15
4.45 0.182
Rel-8 SU-MIMO
(4 x 4, L=3)
3.41 0.143
3GPP target 3.70 0.120
Tables 10.1.1.1-4, 10.1.1.1-5, and 10.1.1.1-6 show the spectrum efficiency results of 2-by-2, 4-by-2, and 4-by-4 CS/CB-
CoMP schemes with eNB antenna configuration (C). Results show that the CS/CB schemes satisfy the 3GPP target and
achieve significant gain compared with the Rel-8 SU-MIMO scheme.
Table 10.1.1.1-4 Performance of DL CS/CB-CoMP 2 x 2 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 2.54 0.088
Source 15 2.58 0.070
Rel-8 SU-MIMO
(2 x 2, L=3)
2.23 0.079
3GPP target 2.40 0.070
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 27 Release 9
Table 10.1.1.1-5 Performance of DL CS/CB-CoMP 4 x 2 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.28 0.150
Source 3 3.34 0.144
Source 15 3.51 0.121
Source 18 3.24 0.100
Rel-8 SU-MIMO
(4 x 2, L=3)
2.53 0.100
3GPP target 2.60 0.090
Table 10.1.1.1-6 Performance of DL CS/CB-CoMP 4 x 4 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 4.97 0.239
Source 15 4.58 0.195
Source 18 4.44 0.180
Rel-8 SU-MIMO
(4 x 4, L=3)
3.41 0.143
3GPP target 3.70 0.120
Tables 10.1.1.1-7, 10.1.1.1-8, and 10.1.1.1-9 show the spectrum efficiency results of 2-by-2, 4-by-2, and 4-by-4 JP-
CoMP schemes with eNB antenna configuration (C). Results show that the JP-CoMP schemes provide additional
performance enhancement over MU-MIMO in Table 10.1.1-1-1, 10.1.1-1-2, and Table 10.1.1-1-3 (compared with the
same source).
Table 10.1.1.1-7 Performance of DL JP-CoMP 2 x 2 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.92 0.120
Source 15 2.47 0.088
Rel-8 SU-MIMO
(2 x 2, L=3)
2.23 0.079
3GPP target 2.60 0.090
Table 10.1.1.1-8 Performance of DL JP-CoMP 4 x 2 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 4.40 0.180
Source 15 3.34 0.143
Rel-8 SU-MIMO
(4 x 2, L=3)
2.53 0.100
3GPP target 2.60 0.090
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 28 Release 9
Table 10.1.1.1-9 Performance of DL JP-CoMP 4 x 4 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 5.89 0.31
Source 15 4.48 0.227
Rel-8 SU-MIMO
(4 x 4, L=3)
3.41 0.143
3GPP target 3.70 0.120
10.1.1.2 TDD, Downlink
Tables 10.1.1.2-1, 10.1.1.2-2, and 10.1.1.2-3 show the spectrum efficiency results of 2-by-2, 4-by-2, and 4-by-4 MU-
MIMO schemes with eNB antenna configuration (C). Table 10.1.1.2-4 shows the spectrum efficiency results of 4-by-2
MU-MIMO schemes with eNB antenna configuration (E). Results show that the MU-MIMO schemes satisfy the 3GPP
target and provide significant gain compared with the Rel-8 SU-MIMO scheme.
Table 10.1.1.2-1 Performance of DL MU-MIMO 2 x 2 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.96 0.120
Source 5 2.80 0.106
Rel-8 SU-MIMO
(2 x 2, L=3)
2.17 0.083
3GPP target 2.40 0.070
Table 10.1.1.2-2 Performance of DL MU-MIMO 4 x 2 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 4.23 0.160
Source 2
3.45 0.131
Source 4
4.16 0.173
Source 5
3.68 0.150
Source 10
3.52 0.172
Source 19
3.5 0.121
Rel-8 SU-MIMO
(4 x 2, L=3)
2.52 0.096
3GPP target 2.60 0.090
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 29 Release 9
Table 10.1.1.2-3 Performance of DL MU-MIMO 4 x 4 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 6.16 0.250
Source 2 4.67 0.194
Source 5 4.07 0.182
Source 19 4.70 0.160
Rel-8 SU-MIMO
(4 x 4, L=3)
3.28 0.154
3GPP target 3.70 0.120
Table 10.1.1.2-4 Performance of DL MU-MIMO 4 x 2 (E) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 10
3.35 0.150
Rel-8 SU-MIMO
(4 x 2, L=3)
2.42 0.073
3GPP target 2.60 0.090
Tables 10.1.1.2-5, 10.1.1.2-6, and 10.1.1.2-7 show the spectrum efficiency results of 2-by-2, 4-by-2, and 4-by-4 CS/CB-
CoMP schemes with eNB antenna configuration (C). Results show that the CS/CB schemes satisfy the 3GPP target and
achieve significant gain compared with the Rel-8 SU-MIMO scheme.
Table 10.1.1.2-5 Performance of DL CS/CB-CoMP 2 x 2 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 2.54 0.088
Rel-8 SU-MIMO
(2 x 2, L=3)
2.17 0.083
3GPP target 2.40 0.070
Table 10.1.1.2-6 Performance of DL CS/CB-CoMP 4 x 2 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 3.38 0.146
Rel-8 SU-MIMO
(4 x 2, L=3)
2.52 0.096
3GPP target 2.60 0.090
Table 10.1.1.2-7 Performance of DL CS/CB-CoMP 4 x 4 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 5.06 0.244
Rel-8 SU-MIMO
(4 x 4, L=3)
3.28 0.154
3GPP target 3.70 0.120
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 30 Release 9
Tables 10.1.1.2-8, 10.1.1.2-9, and 10.1.1.2-10 show the spectrum efficiency results of 2-by-2, 4-by-2, and 4-by-4 JP-
CoMP schemes with eNB antenna configuration (C). Table 10.1.1.2-11 shows the spectrum efficiency results of 4-by-2
JP-CoMP schemes with eNB antenna configuration (E). Results show that the JP-CoMP schemes provide additional
performance enhancement over MU-MIMO in Table 10.1.1.2-1, 10.1.1.2-2, 10.1.1.2-3, and 10.1.1.2-4 (compared with
the same source).
Table 10.1.1.2-8 Performance of DL JP-CoMP 2 x 2 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.15 0.130
Rel-8 SU-MIMO
(2 x 2, L=3)
2.17 0.083
3GPP target 2.40 0.070
Table 10.1.1.2-9 Performance of DL JP-CoMP 4 x 2 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 4.89 0.210
Source 10 4.38 0.188
Rel-8 SU-MIMO
(4 x 2, L=3)
2.52 0.096
3GPP target 2.60 0.090
Table 10.1.1.2-10 Performance of DL JP-CoMP 4 x 4 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 6.61 0.330
Rel-8 SU-MIMO
(4 x 4, L=3)
3.28 0.154
3GPP target 3.70 0.120
Table 10.1.1.2-11 Performance of DL JP-CoMP 4 x 2 (E) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 10 4.05 0.175
Rel-8 SU-MIMO
(4 x 2, L=3)
2.42 0.073
3GPP target 2.60 0.090
10.1.1.3 FDD, Uplink
Tables 10.1.1.3-1 and 10.1.1.3-2 show the spectrum efficiency results of 2-by-4 SU-MIMO with eNB antenna
configuration (A) and (C), respectively. Results show that the 2-by-4 SU-MIMO schemes satisfy the 3GPP target and
provide significant gain compared with the Rel-8 SIMO scheme.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 31 Release 9
Table 10.1.1.3-1 Performance of UL SU-MIMO 2 x 4 (A) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 2.46 0.112
Source 15 2.26 0.093
Source 18 2.36 0.086
Rel-8 SIMO
(1 x 4)
1.95 0.075
3GPP target 2.00 0.070
Table 10.1.1.3-2 Performance of UL SU-MIMO 2 x 4 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.33 0.070
Source 3 2.37 0.118
Source 4
2.43 0.096
Source 5
2.34 0.080
Source 15
2.01 0.096
Source 18 2.12 0.087
Rel-8 SIMO
(1 x 4)
2.00 0.075
3GPP target 2.00 0.070
Tables 10.1.1.3-3 and 10.1.1.3-4 show the spectrum efficiency results of 1-by-2 CoMP with eNB antenna configuration
(A) and (C), respectively. Results show that the Rel-8 SIMO scheme satisfies the 3GPP target. Tables 10.1.1.3-5 and
10.1.1.3-6 show spectrum efficiency results of 2-by-4 CoMP with eNB antenna configuration (A) and (C), respectively.
Results show that multi point reception provides additional performance enhancement compared with the Rel-8 SIMO
scheme.
Table 10.1.1.3-3 Performance of UL CoMP 1 x 2 (A) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 18 1.44 0.052
Rel-8 SIMO
(1 x 2)
1.42 0.050
3GPP target 1.20 0.040
Table 10.1.1.3-4 Performance of UL CoMP 1 x 2 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 18 1.40 0.051
Rel-8 SIMO
(1 x 2)
1.33 0.047
3GPP target 1.20 0.040
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 32 Release 9
Table 10.1.1.3-5 Performance of UL CoMP 2 x 4 (A) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 18 2.43 0.092
Rel-8 SIMO
(1 x 4)
1.95 0.075
3GPP target 2.00 0.070
Table 10.1.1.3-6 Performance of UL CoMP 2 x 4 (C) (3GPP Case1, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 18 2.18 0.092
Rel-8 SIMO
(1 x 4)
2.00 0.075
3GPP target 2.00 0.070
10.1.1.4 TDD, Uplink
Tables 10.1.1.4-1 and 10.1.1.4-2 show the spectrum efficiency results of 2-by-4 SU-MIMO with eNB antenna
configuration (A) and (C), respectively. Results show that the 2-by-4 SU-MIMO schemes satisfy the 3GPP target and
provide gain compared with the Rel-8 SIMO and Rel-8 MU-MIMO schemes.
Table 10.1.1.4-1 Performance of UL SU-MIMO 2 x 4 (A) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 2.23 0.101
Source 5
2.33 0.080
Rel-8 MU-MIMO
(1 x 4)
2.09 0.070
3GPP target 2.00 0.070
Table 10.1.1.4-2 Performance of UL SU-MIMO 2 x 4 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.16 0.080
Source 3 2.15 0.107
Source 4
2.13 0.082
Rel-8 SIMO
(1 x 4)
1.83 0.064
3GPP target 2.00 0.070
Table 10.1.1.4-3 shows the spectrum efficiency results of 1-by-2 CoMP with eNB antenna configuration (C). Results
show that the Rel-8 SIMO scheme satisfies the 3GPP target and multi point reception provides performance
enhancement compared with the Rel-8 SIMO scheme.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 33 Release 9
Table 10.1.1.4-3 Performance of UL CoMP 1 x 2 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 10 1.51 0.051
Rel-8 SIMO
(1 x 2)
1.24 0.045
3GPP target 1.20 0.040
Table 10.1.1.4-4 shows the spectrum efficiency results of 2-by-4 MU-MIMO with eNB antenna configuration (C).
Result shows that MU-MIMO provides performance enhancement compared with the Rel-8 SIMO scheme.
Table 10.1.1.4-4 Performance of UL MU-MIMO 2 x 4 (C) (3GPP Case1, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 2.59 0.079
Rel-8 SIMO
(1 x 4)
1.83 0.064
3GPP target 2.00 0.070
10.2 Cell spectral efficiency and cell-edge spectral efficiency for
ITU-R requirements
Cell spectral efficiency and cell-edge spectral efficiency are evaluated through extensive simulations conducted by a
number of companies. The simulation assumptions applied in the following evaluations are shown in Annex A.3.
Detailed information covering a range of possible configurations is provided in Annex A.4 and [5]. The tables in the
following subsections show the simulation results focusing on LTE-Advanced configurations, i.e., downlink and uplink
MU-MIMO, downlink and uplink CoMP, and uplink SU-MIMO.
Tables 10.2-1, 2, 3, and 4 show the source specific simulation parameters that characterize the performances of
downlink FDD, downlink TDD, uplink FDD, and uplink TDD, respectively. Note, however, that the performance
differences among sources could also be explained by other factors such as detailed signal processing algorithms at the
transmitter and the receiver. More detailed assumption on each component is shown in [5].
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 34 Release 9
Table 10.2-1 Simulation parameters (DL, FDD)
CCH duration
(L symbols)
Num of MBSFN
subframe
Channel
estimation
Receiver
type
CSI assumption
at eNB
Source 1 3 0 Real MMSE Short-term
Source 2 2 0 Real
(Ideal for JP-
CoMP)
MMSE Short-term
Source 3 3 6 Real MMSE with IRC Long-term
Source 4 3 0 Real MMSE Short-term
Source 5 3 6 Real MMSE Short-term
Source 7 3 0 Ideal MMSE No
Source 8 3 (Normal
subframe)
2 (MBSFN
subframe)
6 Real MMSE Short-term
Source 9 3 0 Real MMSE with IRC Short-term
Source 10 2 0 Real MMSE with IRC Short-term
Source 11 3 (Normal
subframe)
2 (MBSFN
subframe)
6 Real MMSE with IRC Short-term and
Long-term
Source 12 3 6 Real MMSE with IRC Short-term
Source 13 3 0 Real MMSE with IRC Long-term
Source 15 3 6 Real MMSE with IRC Long-term
Source 16 2 0 Ideal MMSE No
Source 18 2 6 Real MMSE with IRC Long-term
Table 10.2-2 Simulation parameters (DL, TDD)
CCH duration
(L symbols)
Num of MBSFN
subframe
Channel
estimation
Receiver
type
CSI assumption
at eNB
Source 1 2 0 Real MMSE Short-term
Source 2 2 0 Real
(Ideal for JP-
CoMP)
MMSE Short-term
Source 3 3 2 Real MMSE with IRC Long-term
Source 4 2 0 Real MMSE Short-term
Source 7 3 0 Real MMSE No
Source 8 3 (Normal
subframe)
2 (MBSFN
subframe)
2 Real MMSE Short-term
Source 9 2 0 Real MMSE with IRC Short-term
Source 10 2 0 Real MMSE with IRC Short-term
Source 12 3 2 Real MMSE with IRC Short-term
Source 13 3 0 Real MMSE with IRC Long-term
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 35 Release 9
Table 10.2-3 Simulation parameters (UL, FDD)
PUCCH
bandwidth
(RB/10 MHz)
Channel
estimation
Receiver type
Source 1
6 Real MMSE
Source 2
4 Real MMSE
Source 3
4 Real MMSE with IRC
Source 4
4 Real MMSE
Source 5
4 Real MMSE
Source 7
4 Ideal MMSE
Source 8
4 Real MMSE
Source 9
4 Real MMSE with IRC
Source 10
4 Real MMSE with IRC
Source 11
4 Real MMSE with IRC
Source 12
6 Real MMSE with IRC
Source 14
3 (InH)
4 (Others)
Real MMSE with IRC
and MMSE with
SIC for rank 2
Source 18
4 Real MMSE with IRC
Table 10.2-4 Simulation parameters (UL, TDD)
PUCCH
bandwidth
(RB/10 MHz)
Channel
estimation
Receiver type
Source 1
4 Real MMSE
Source 2
4 Real MMSE
Source 3
4 Real MMSE with IRC
Source 4
4 Real MMSE
Source 8
4 Real MMSE
Source 9
4 Real MMSE with IRC
Source 10
4 Real MMSE with IRC
Source 12
3 Real MMSE with IRC
Tables in the following subsections capture the spectrum efficiency results from individual sources. As performance
references, the ITU-R requirements and the averaged results of Rel-8 schemes (4-by-2 SU-MIMO with L=3 for
downlink, and 1-by-4 SIMO for uplink) assuming real channel estimation.
10.2.1 Indoor (InH channel model)
10.2.1.1 FDD, Downlink (InH)
Tables 10.2.1.1-1 and 10.2.1.1-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna
configuration (C) and (A), respectively. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and
provide significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 36 Release 9
Table 10.2.1.1-1 Performance of DL MU-MIMO 4 x 2 (C) (InH, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 5.68 0.224
Source 8 5.39 0.210
Source 15 5.66 0.232
Rel-8 SU-MIMO
(4 x 2, L=3)
4.00 0.200
ITU-R requirement 3.00 0.100
Table 10.2.1.1-2 Performance of DL MU-MIMO 4 x 2 (A) (InH, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 8 5.39 0.211
Rel-8 SU-MIMO
(4 x 2, L=3)
4.14 0.198
ITU-R requirement 3.00 0.100
Table 10.2.1.1-3 shows the spectrum efficiency results of 8-by-2 MU-MIMO scheme with the eNB antenna
configuration (C/E). Results show that increasing the number of eNB antennas contributes to additional performance
enhancement in cell spectral efficiency.
Table 10.2.1.1-3 Performance of DL MU-MIMO 8 x 2 (C/E) (InH, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 8 5.60 0.203
Rel-8 SU-MIMO
(4 x 2, L=3)
4.00 0.200
ITU-R requirement 3.00 0.100
10.2.1.2 TDD, Downlink (InH)
Table 10.2.1.2-1 shows the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna configuration
(C). Results show that the MU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared
with the 4-by-2 SU-MIMO scheme in Rel-8.
Table 10.2.1.2-1 Performance of DL MU-MIMO 4 x 2 (C) (InH, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 5.63 0.221
Source 8 5.03 0.186
Source 9 5.80 0.220
Source 10 5.91 0.180
Rel-8 SU-MIMO
(4 x 2, L=3)
3.90 0.211
ITU-R requirement 3.00 0.100
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 37 Release 9
Table 10.2.1.2-2 shows the spectrum efficiency results of 8-by-2 MU-MIMO scheme with the eNB antenna
configuration (C/E). Results show that increasing the number of eNB antennas enhances cell spectral efficiency.
Table 10.2.1.2-2 Performance of DL MU-MIMO 8 x 2 (C/E) (InH, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 8 5.18 0.198
Rel-8 SU-MIMO
(4 x 2, L=3)
3.90 0.211
ITU-R requirement 3.00 0.100
10.2.1.3 FDD, Uplink (InH)
Table 10.2.1.3-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration
(A). Results show that the SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared
with the 1-by-4 SU-MIMO scheme in Rel-8. Use of advanced receiver such as IRC and SIC provides additional gains.
Table 10.2.1.3-1 Performance of UL SU-MIMO 2 x 4 (A) (InH, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 4.01 0.276
Source 3 4.31 0.274
Source 5 3.91 0.265
Source 12 4.39 0.139
Source 14 5.84 0.400
Source 18 4.65 0.273
Rel-8 SIMO
(1 x 4)
3.25 0.226
ITU-R requirement 2.25 0.070
Table 10.2.1.3-2 shows the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna configuration (A).
Results show that multi point reception provides performance enhancement compared with the Rel-8 1-by-4 SIMO
scheme.
Table 10.2.1.3-2 Performance of UL CoMP 1 x 4 (A) (InH, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.42 0.271
Source 18 3.32 0.239
Rel-8 SIMO
(1 x 4)
3.25 0.226
ITU-R requirement 2.25 0.070
Table 10.2.1.3-3 shows the spectrum efficiency results of 2-by-4 CoMP schemes with eNB antenna configuration (A).
Results show that additional transmit antenna provides further performance enhancement.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 38 Release 9
Table 10.2.1.3-3 Performance of UL CoMP 2 x 4 (A) (InH, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 4.12 0.278
Source 18 4.74 0.274
Rel-8 SIMO
(1 x 4)
3.25 0.226
ITU-R requirement 2.25 0.070
10.2.1.4 TDD, Uplink (InH)
Table 10.2.1.4-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration
(A). Results show that the 2-by-4 SU-MIMO schemes satisfy the ITU-R requirements. Performance improvement from
the Rel-8 1-by-4 SIMO scheme is also demonstrated. Use of advanced receiver such as IRC and SIC provides additional
gains.
Table 10.2.1.4-1 Performance of UL SU-MIMO 2 x 4 (A) (InH, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.78 0.243
Source 3 3.98 0.260
Source 12 4.43 0.124
Rel-8 SIMO
(1 x 4)
3.04 0.206
ITU-R requirement 2.25 0.070
Table 10.2.1.4-2 shows the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna configuration (A).
Results show that multi point reception provides performance enhancement compared with the Rel-8 1-by-4 SIMO
scheme.
Table 10.2.1.4-2 Performance of UL CoMP 1 x 4 (A) (InH, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.41 0.250
Rel-8 SIMO
(1 x 4)
3.04 0.206
ITU-R requirement 2.25 0.070
Table 10.2.1.4-3 shows the spectrum efficiency results of 2-by-4 CoMP schemes with eNB antenna configuration (A).
Results show that additional transmit antenna provides further performance enhancement in cell spectral efficiency.
Table 10.2.1.4-3 Performance of UL CoMP 2 x 4 (A) (InH, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.84 0.240
Rel-8 SIMO
(1 x 4)
3.04 0.206
ITU-R requirement 2.25 0.070
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 39 Release 9
10.2.2 Microcellular (UMi channel model)
10.2.2.1 FDD, Downlink (UMi)
Tables 10.2.2.1-1 and 10.2.2.1-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna
configuration (C) and (A), respectively. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and
provide significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8 that cannot meet the requirement. Use of
MBSFN subframes (overhead reduction), IRC receiver, and short-term feedback tends to improve the cell- and cell-
edge spectrum efficiency performances.
Table 10.2.2.1-1 Performance of DL MU-MIMO 4 x 2 (C) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.51 0.079
Source 2 2.86 0.083
Source 4 3.14 0.078
Source 5 2.97 0.076
Source 7 2.18 0.063
Source 8 2.88 0.105
Source 11 3.17 0.084
Source 12 2.95 0.103
Source 15 2.74 0.081
Rel-8 SU-MIMO
(4 x 2, L=3)
2.14 0.068
ITU requirement 2.60 0.075
Table 10.2.2.1-2 Performance of DL MU-MIMO 4 x 2 (A) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 5 3.15 0.103
Source 8 2.61 0.100
Source 12 2.65 0.093
Rel-8 SU-MIMO
(4 x 2, L=3)
1.96 0.061
ITU requirement 2.60 0.075
Table 10.2.2.1-3 shows the spectrum efficiency results of 8-by-2 MU-MIMO schemes with the eNB antenna
configuration (C/E). Obviously, increasing the number of eNB antennas contributes to further performance
enhancement.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 40 Release 9
Table 10.2.2.1-3 Performance of DL MU-MIMO 8 x 2 (C/E) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.62 0.138
Source 5 3.72 0.140
Source 8 3.03 0.130
Source 10 3.36 0.089
Rel-8 SU-MIMO
(4 x 2, L=3)
2.14 0.068
ITU requirement 2.60 0.075
Tables 10.2.2.1-4 and 10.2.2.1-5 show the spectrum efficiency results of 4-by-2 CS/CB-CoMP schemes with eNB
antenna configuration (C) and (A), respectively. Table 10.2.2.1-6 shows the spectrum efficiency results of 8-by-2
CS/CB-CoMP schemes with eNB antenna configuration (C/E). Results show that the CS/CB schemes satisfy the ITU
requirements and achieve significant gain compared with the Rel-8 4-by-2 SU-MIMO scheme.
Table 10.2.2.1-4 Performance of DL CS/CB-CoMP 4 x 2 (C) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 2.84 0.092
Source 11 3.11 0.086
Source 12 2.99 0.114
Source 13 3.21 0.084
Source 18 3.15 0.083
Rel-8 SU-MIMO
(4 x 2, L=3)
2.14 0.068
ITU requirement 2.60 0.075
Table 10.2.2.1-5 Performance of DL CS/CB-CoMP 4 x 2 (A) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 12 2.71 0.103
Rel-8 SU-MIMO
(4 x 2, L=3)
1.96 0.061
ITU requirement 2.60 0.075
Table 10.2.2.1-6 Performance of DL CS/CB-CoMP 8 x 2 (C) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 3.99 0.105
Source 4 3.05 0.088
Rel-8 SU-MIMO
(4 x 2, L=3)
2.14 0.068
ITU requirement 2.60 0.075
Table 10.2.2.1-7 shows the spectrum efficiency results of 4-by-2 JP-CoMP schemes with eNB antenna configuration
(C). Results show that the JP-CoMP schemes provide additional performance enhancement over MU-MIMO in Table
10.2.2-1-1 (compared with the same source).
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 41 Release 9
Table 10.2.2.1-7 Performance of DL JP-CoMP 4 x 2 (C) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.45 0.108
Source 2 3.06 0.097
Rel-8 SU-MIMO
(4 x 2, L=3)
2.14 0.068
ITU requirement 2.60 0.075
Tables 10.2.2.1-8 and 10.2.2.1-9 show the spectrum efficiency results of 8-by-2 SU-MIMO scheme with eNB antenna
configuration (A) and (C), respectively. Results show that use of 8 transmit antenna gives additional degrees of
freedom, resulting in enhanced performance.
Table 10.2.2.1-8 Performance of DL SU-MIMO 8 x 2 (A) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 16 3.65 0.140
Rel-8 SU-MIMO
(4 x 2, L=3)
1.96 0.061
ITU requirement 2.60 0.075
Table 10.2.2.1-9 Performance of DL SU-MIMO 8 x 2 (C) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 16 3.74 0.153
Rel-8 SU-MIMO
(4 x 2, L=3)
2.14 0.068
ITU requirement 2.60 0.075
10.2.2.2 TDD, Downlink (UMi)
Tables 10.2.2.2-1 and 10.2.2.2-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna
configuration (C) and (A), respectively. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and
provide significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8 that cannot meet the requirement. Use of
MBSFN subframes (overhead reduction), IRC receiver, and short-term feedback tends to improve the cell- and cell-
edge spectrum efficiency performances.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 42 Release 9
Table 10.2.2.2-1 Performance of DL MU-MIMO 4 x 2 (C) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.85 0.087
Source 2 3.07 0.084
Source 4 3.07 0.076
Source 7 3.33 0.090
Source 8 2.71 0.103
Source 9 3.22 0.095
Source 10 2.94 0.084
Source 12 2.90 0.105
Rel-8 SU-MIMO
(4 x 2, L=3)
2.20 0.076
ITU-R requirement 2.60 0.075
Table 10.2.2.2-2 Performance of DL MU-MIMO 4 x 2 (A) (UMi, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 12 2.65 0.092
Rel-8 SU-MIMO
(4 x 2, L=3)
1.99 0.067
ITU-R requirement 2.60 0.075
Table 10.2.2.2-3 shows the spectrum efficiency results of 8-by-2 MU-MIMO schemes with the eNB antenna
configuration (C/E). Obviously, increasing the number of eNB antennas contributes to further performance
enhancement.
Table 10.2.2.2-3 Performance of DL MU-MIMO 8 x 2 (C/E) (UMi, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.85 0.106
Source 2 3.46 0.089
Source 9 3.33 0.099
Source 10 3.47 0.096
Rel-8 SU-MIMO
(4 x 2, L=3)
2.20 0.076
ITU-R requirement 2.60 0.075
Tables 10.2.2.2-4 and 10.2.2.2-5 show the spectrum efficiency results of 4-by-2 CS/CB-CoMP schemes with eNB
antenna configuration (C), and 8-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C), respectively. Table
10.2.2.1-6 shows the spectrum efficiency results of 4-by-2 JP-CoMP schemes with eNB antenna configuration (C).
Results show that the CoMP schemes satisfy the ITU requirements and achieve significant gain compared with the Rel-
8 4-by-2 SU-MIMO scheme.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 43 Release 9
Table 10.2.2.2-4 Performance of DL CS/CB-CoMP 4 x 2 (C) (UMi, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 3.28 0.087
Source 3 2.78 0.089
Source 13 3.20 0.083
Rel-8 SU-MIMO
(4 x 2, L=3)
2.20 0.076
ITU-R requirement 2.60 0.075
Table 10.2.2.2-5 Performance of DL CS/CB-CoMP 8 x 2 (C) (UMi, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 3.83 0.096
Source 4 3.20 0.096
Rel-8 SU-MIMO
(4 x 2, L=3)
2.20 0.076
ITU-R requirement 2.60 0.075
Table 10.2.2.2-6 Performance of DL JP-CoMP 4 x 2 (C) (UMi, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.76 0.082
Source 2 3.03 0.096
Rel-8 SU-MIMO
(4 x 2, L=3)
2.20 0.076
ITU-R requirement 2.20 0.060
10.2.2.3 FDD, Uplink (UMi)
Table 10.2.2.3-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration
(A). Results show that the SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared
with the 1-by-4 SU-MIMO scheme in Rel-8. Use of advanced receiver such as IRC and SIC provides additional gains.
Table 10.2.2.3-1 Performance of UL SU-MIMO 2 x 4 (A) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.23 0.087
Source 3 2.52 0.112
Source 5 2.24 0.088
Source 11 1.93 0.075
Source 12 1.98 0.051
Source 14 2.45 0.071
Source 18 2.08 0.092
Rel-8 SIMO
(1 x 4)
1.93 0.074
ITU-R requirement 1.80 0.050
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 44 Release 9
Tables 10.2.2.3-2 and 10.2.2.3-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna
configuration (A) and 2-by-4 CoMP schemes with eNB antenna configuration (A), respectively. Results show that multi
point reception provides performance enhancement compared with the Rel-8 1-by-4 SIMO scheme.
Table 10.2.2.3-2 Performance of UL CoMP 1 x 4 (A) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.22 0.093
Source 18 2.03 0.089
Rel-8 SIMO
(1 x 4)
1.93 0.074
ITU-R requirement 1.80 0.050
Table 10.2.2.3-3 Performance of UL CoMP 2 x 4 (A) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.45 0.102
Source 18 2.18 0.102
Rel-8 SIMO
(1 x 4)
1.93 0.074
ITU-R requirement 1.80 0.050
Table 10.2.2.3-4 shows the spectrum efficiency results of 2-by-4 MU-MIMO schemes with eNB antenna configuration
(A). Results show that MU-MIMO provides performance enhancement compared with SU-MIMO shown in Table
10.2.2.3-1 (compared with the same source).
Table 10.2.2.3-4 Performance of UL MU-MIMO 2 x 4 (A) (UMi, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 12 2.51 0.086
Rel-8 SIMO
(1 x 4)
1.93 0.074
ITU-R requirement 1.80 0.050
10.2.2.4 TDD, Uplink (UMi)
Table 10.2.2.4-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration
(A). Results show that the SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared
with the 1-by-4 SU-MIMO scheme in Rel-8.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 45 Release 9
Table 10.2.2.4-1 Performance of UL SU-MIMO 2 x 4 (A) (UMi, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.12 0.081
Source 3 2.29 0.100
Source 12 1.90 0.058
Rel-8 SIMO
(1 x 4)
1.87 0.072
ITU-R requirement 1.80 0.050
Tables 10.2.2.4-2 and 10.2.2.4-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna
configuration (A) and 2-by-4 CoMP schemes with eNB antenna configuration (A), respectively. Results show that multi
point reception provides performance enhancement compared with the Rel-8 1-by-4 SIMO scheme.
Table 10.2.2.4-2 Performance of UL CoMP 1 x 4 (A) (UMi, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.15 0.083
Rel-8 SIMO
(1 x 4)
1.87 0.072
ITU-R requirement 1.80 0.050
Table 10.2.2.4-3 Performance of UL CoMP 2 x 4 (A) (UMi, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.35 0.097
Rel-8 SIMO
(1 x 4)
1.87 0.072
ITU-R requirement 1.80 0.050
Table 10.2.2.4-4 shows the spectrum efficiency results of 2-by-4 MU-MIMO schemes with eNB antenna configuration
(A). Results show that MU-MIMO provides performance enhancement compared with SU-MIMO shown in Table
10.2.2.4-1 (compared with the same source).
Table 10.2.2.4-4 Performance of UL MU-MIMO 2 x 4 (A) (UMi, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 12 2.79 0.068
Rel-8 SIMO
(1 x 4)
1.87 0.072
ITU-R requirement 1.80 0.050
Table 10.2.2.4-5 shows the spectrum efficiency results of 1-by-8 MU-MIMO schemes with eNB antenna configuration
(C/E). Results show that additional eNB receive antennas contributes to additional performance enhancement.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 46 Release 9
Table 10.2.2.4-5 Performance of UL MU-MIMO 1 x 8 (C/E) (UMi, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 2.97 0.079
Rel-8 SIMO
(1 x 4)
1.82 0.067
ITU-R requirement 1.80 0.050
10.2.3 Base coverage urban (UMa channel model)
10.2.3.1 FDD, Downlink (UMa)
Tables 10.2.3.1-1 and 10.2.3.1-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna
configuration (C) and (A), respectively. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and
provide significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8 that cannot meet the requirement. Use of
MBSFN subframes (overhead reduction), IRC receiver, and short-term feedback tends to improve the cell- and cell-
edge spectrum efficiency performances.
Table 10.2.3.1-1 Performance of DL MU-MIMO 4 x 2 (C) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.14 0.056
Source 5 2.22 0.060
Source 7 2.06 0.060
Source 8 2.48 0.063
Source 11 2.66 0.070
Source 12 2.36 0.076
Source 15 2.22 0.062
Rel-8 SU-MIMO
(4 x 2, L=3)
1.61 0.050
ITU-R requirement 2.20 0.060
Table 10.2.3.1-2 Performance of DL MU-MIMO 4 x 2 (A) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 1.75 0.047
Source 5 2.33 0.064
Rel-8 SU-MIMO
(4 x 2, L=3)
1.45 0.043
ITU-R requirement 2.20 0.060
Table 10.2.3.1-3 shows the spectrum efficiency results of 8-by-2 MU-MIMO schemes with the eNB antenna
configuration (C/E). Obviously, increasing the number of eNB antennas contributes to further performance
enhancement.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 47 Release 9
Table 10.2.3.1-3 Performance of DL MU-MIMO 8 x 2 (C/E) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.77 0.091
Source 5 2.80 0.104
Source 10 2.55 0.070
Rel-8 SU-MIMO
(4 x 2, L=3)
1.61 0.050
ITU-R requirement 2.20 0.060
Tables 10.2.3.1-4 and 10.2.3.1-5 show the spectrum efficiency results of 4-by-2 CS/CB-CoMP schemes with eNB
antenna configuration (C) and 8-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C), respectively. Results
show that the CS/CB-CoMP schemes satisfy the ITU requirements and achieve significant gain compared with the Rel-
8 4-by-2 SU-MIMO scheme.
Table 10.2.3.1-4 Performance of DL CS/CB-CoMP 4 x 2 (C) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 2.68 0.067
Source 3 2.38 0.073
Source 11 2.44 0.067
Source 12 2.33 0.083
Source 13 2.30 0.063
Source 18 2.45 0.060
Rel-8 SU-MIMO
(4 x 2, L=3)
1.61 0.050
ITU-R requirement 2.20 0.060
Table 10.2.3.1-5 Performance of DL CS/CB-CoMP 8 x 2 (C) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 3.34 0.088
Source 4 2.70 0.085
Source 13 3.22 0.073
Rel-8 SU-MIMO
(4 x 2, L=3)
1.61 0.050
ITU-R requirement 2.20 0.060
Tables 10.2.3.1-6 and 10.2.3.1-7 show the spectrum efficiency results of 4-by-2 JP-CoMP schemes with eNB antenna
configuration (C) and (A), respectively. Tables show that the JP CoMP schemes provide additional performance
enhancement over MU-MIMO in Table 10.2.3.1-1 (compared with the same source).
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 48 Release 9
Table 10.2.3.1-6 Performance of DL JP-CoMP 4 x 2 (C) (UMa, FDD) 
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.41 0.064
Source 2 2.32 0.069
Rel-8 SU-MIMO
(4 x 2, L=3)
1.61 0.050
ITU-R requirement 2.20 0.060
Table10.2.3.1-7 Performance of DL JP-CoMP 4 x 2 (A) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 5 2.41 0.065
Rel-8 SU-MIMO
(4 x 2, L=3)
1.45 0.043
ITU-R requirement 2.20 0.060
Tables 10.2.3.1-8 and 10.2.3.1-9 show the spectrum efficiency results of 8-by-2 SU-MIMO schemes with eNB antenna
configuration (A) and (C), respectively. Results show that the use of 8 antenna elements gives additional degrees of
freedom, which leads to improved performance.
Table 10.2.3.1-8 Performance of DL SU-MIMO 8 x 2 (A) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 16 3.53 0.150
Rel-8 SU-MIMO
(4 x 2, L=3)
1.45 0.043
ITU-R requirement 2.20 0.060
Table 10.2.3.1-9 Performance of DL SU-MIMO 8 x 2 (C) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 16 3.54 0.149
Rel-8 SU-MIMO
(4 x 2, L=3)
1.61 0.050
ITU-R requirement 2.20 0.060
10.2.3.2 TDD, Downlink (UMa)
Tables 10.2.3.2-1 and 10.2.3.2-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna
configuration (C) and (A), respectively. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and
provide significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8 that cannot meet the requirements. Use
of MBSFN subframes (overhead reduction), IRC receiver, and short-term feedback tends to improve the cell- and cell-
edge spectrum efficiency performances.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 49 Release 9
Table 10.2.3.2-1 Performance of DL MU-MIMO 4 x 2 (C) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.34 0.062
Source 2 2.63 0.069
Source 7 2.27 0.062
Source 8 2.43 0.062
Source 9 2.62 0.080
Source 10 2.49 0.069
Source 12 2.32 0.071
Rel-8 SU-MIMO
(4 x 2, L=3)
1.34 0.040
ITU-R requirement 2.20 0.060
Table 10.2.3.2-2 Performance of DL MU-MIMO 4 x 2 (A) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 1.85 0.056
Rel-8 SU-MIMO
(4 x 2, L=3)
1.36 0.036
ITU-R requirement 2.20 0.060
Table 10.2.3.2-3 shows the spectrum efficiency results of 8-by-2 MU-MIMO schemes with the eNB antenna
configuration (C/E). Obviously, increasing the number of eNB antennas contributes to further performance
enhancement.
Table 10.2.3.2-3 Performance of DL MU-MIMO 8 x 2 (C/E) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 3.42 0.102
Source 2 3.14 0.077
Source 9 2.92 0.075
Source 10 2.89 0.077
Rel-8 SU-MIMO
(4 x 2, L=3)
1.34 0.040
ITU-R requirement 2.20 0.060
Tables 10.2.3.2-4 and 10.2.3.2-5 show the spectrum efficiency results of 4-by-2 CS/CB-CoMP schemes with eNB
antenna configuration (C) and 8-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C), respectively. Results
show that the CS/CB-CoMP schemes satisfy the ITU requirements and achieve significant gain compared with the Rel-
8 4-by-2 SU-MIMO scheme.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 50 Release 9
Table 10.2.3.2-4 Performance of DL CS/CB-CoMP 4 x 2 (C) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 2.74 0.075
Source 3 2.34 0.072
Source 12 2.21 0.069
Source 13 2.34 0.065
Rel-8 SU-MIMO
(4 x 2, L=3)
1.34 0.040
ITU-R requirement 2.20 0.060
Table 10.2.3.2-5 Performance of DL CS/CB-CoMP 8 x 2 (C) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 3.30 0.093
Source 4 2.77 0.092
Source 13 3.24 0.076
Rel-8 SU-MIMO
(4 x 2, L=3)
1.34 0.040
ITU-R requirement 2.20 0.060
Table 10.2.3.2-6 shows the spectrum efficiency results of 4-by-2 JP-CoMP schemes with eNB antenna configuration
(C). Results show that the JP-CoMP schemes satisfy the ITU requirements and achieve significant gain compared with
the Rel-8 4-by-2 SU-MIMO scheme.
Table 10.2.3.2-6 Performance of DL JP-CoMP 4 x 2 (C) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.95 0.073
Source 2 2.23 0.068
Rel-8 SU-MIMO
(4 x 2, L=3)
1.34 0.040
ITU-R requirement 2.20 0.060
10.2.3.3 FDD, Uplink (UMa)
Table 10.2.3.3-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration
(A). Results show that the 2-by-4 SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain
compared with the 1-by-4 SU-MIMO scheme in Rel-8.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 51 Release 9
Table 10.2.3.3-1 Performance of UL SU-MIMO 2 x 4 (A) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 1.77 0.086
Source 3 1.93 0.087
Source 5 1.75 0.082
Source 12 1.62 0.046
Source 14 2.03 0.072
Source 18 1.62 0.072
Rel-8 SIMO
(1 x 4)
1.55 0.067
ITU-R requirement 1.40 0.030
Tables 10.2.3.3-2 and 10.2.3.3-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna
configuration (A) and (C), respectively. Results show that the performance of CoMP schemes exceeds the ITU
requirements.
Table10.2.3.3-2 Performance of UL CoMP 1 x 4 (A) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 1.95 0.096
Source 18 1.51 0.075
Rel-8 SIMO
(1 x 4)
1.55 0.067
ITU-R requirement 1.40 0.030
Table10.2.3.3-3 Performance of UL CoMP 1 x 4 (C) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 1.89 0.095
Rel-8 SIMO
(1 x 4)
1.50 0.060
ITU-R requirement 1.40 0.030
Tables 10.2.3.3-4 and 10.2.3.3-5 show the spectrum efficiency results of 2-by-4 CoMP schemes with eNB antenna
configuration (A) and (C), respectively. Results show that additional transmit antenna provides further performance
enhancement.
Table10.2.3.3-4 Performance of UL CoMP 2 x 4 (A) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.11 0.102
Source 18 1.74 0.081
Rel-8 SIMO
(1 x 4)
1.55 0.067
ITU-R requirement 1.40 0.030
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 52 Release 9
Table 10.2.3.3-5 Performance of UL CoMP 2 x 4 (C) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.05 0.099
Rel-8 SIMO
(1 x 4)
1.50 0.060
ITU-R requirement 1.40 0.030
Table 10.2.3.3-6 shows the spectrum efficiency results of 2-by-4 MU-MIMO schemes with eNB antenna configuration
(A). Results show that MU-MIMO provides performance enhancement compared with SU-MIMO shown in Table
10.2.3.3-1 (compared with the same source).
Table 10.2.3.3-6 Performance of UL MU-MIMO 2 x 4 (A) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 12 1.91 0.069
Rel-8 SIMO
(1 x 4)
1.55 0.067
ITU-R requirement 1.40 0.030
10.2.3.4 TDD, Uplink (UMa)
Table 10.2.3.4-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration
(A). Results show that the 2-by-4 SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain
compared with the 1-by-4 SU-MIMO scheme in Rel-8.
Table 10.2.3.4-1 Performance of UL SU-MIMO 2 x 4 (A) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 1.73 0.083
Source 3 1.76 0.076
Source 12 1.54 0.059
Rel-8 SIMO
(1 x 4)
1.51 0.062
ITU-R requirement 1.40 0.030
Tables 10.2.3.4-2 and 10.2.3.4-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna
configuration (A) and (C), respectively. Results show that multi point reception provides performance enhancement
compared with the Rel-8 1-by-4 SIMO scheme.
Table10.2.3.4-2 Performance of UL CoMP 1 x 4 (A) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 1.89 0.093
Rel-8 SIMO
(1 x 4)
1.51 0.062
ITU-R requirement 1.40 0.030
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 53 Release 9
Table10.2.3.4-3 Performance of UL CoMP 1 x 4 (C) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 1.86 0.090
Rel-8 SIMO
(1 x 4)
1.45 0.059
ITU-R requirement 1.40 0.030
Tables 10.2.3.4-4 and 10.2.3.4-5 show the spectrum efficiency results of 2-by-4 CoMP schemes with eNB antenna
configuration (A) and (C), respectively. Results show that additional transmit antenna provides further performance
enhancement.
Table10.2.3.4-4 Performance of UL CoMP 2 x 4 (A) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.06 0.100
Rel-8 SIMO
(1 x 4)
1.51 0.062
ITU-R requirement 1.40 0.030
Table 10.2.3.4-5 Performance of UL CoMP 2 x 4 (C) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.00 0.097
Rel-8 SIMO
(1 x 4)
1.45 0.059
ITU-R requirement 1.40 0.030
Table 10.2.3.4-6 shows the spectrum efficiency results of 1-by-8 MU-MIMO schemes with eNB antenna configuration
(C/E). Results show that additional eNB receive antennas contributes to further performance enhancement.
Table 10.2.3.4-6 Performance of UL MU-MIMO 1 x 8 (C/E) (UMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 2.74 0.076
Rel-8 SIMO
(1 x 4)
1.45 0.059
ITU-R requirement 1.40 0.030
10.2.4 High speed (RMa channel model)
10.2.4.1 FDD, Downlink (RMa)
Tables 10.2.4.1-1 and 10.2.4.1-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna
configuration (C), and 8-by-2 MU-MIMO schemes with the eNB antenna configuration (C/E), respectively. Results
show that the MU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 4-
by-2 SU-MIMO scheme in Rel-8.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 54 Release 9
Table 10.2.4.1-1 Performance of DL MU-MIMO 4 x 2 (C) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 3.47 0.112
Source 8 2.94 0.077
Source 15 3.26 0.083
Rel-8 SU-MIMO
(4 x 2, L=3)
1.91 0.069
ITU requirement 1.10 0.040
Table 10.2.4.1-2 Performance of DL MU-MIMO 8 x 2 (C/E) (UMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 10 3.45 0.110
Rel-8 SU-MIMO
(4 x 2, L=3)
1.91 0.069
ITU requirement 1.10 0.040
10.2.4.2 TDD, Downlink (RMa)
Tables 10.2.4.2-1 and 10.2.4.2-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna
configuration (C), and 8-by-2 MU-MIMO schemes with the eNB antenna configuration (C/E), respectively. Results
show that the MU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 4-
by-2 SU-MIMO scheme in Rel-8.
Table 10.2.4.2-1 Performance of DL MU-MIMO 4 x 2 (C) (RMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 2.79 0.081
Source 8 2.87 0.075
Source 9 3.34 0.087
Source 10 2.86 0.086
Rel-8 SU-MIMO
(4 x 2, L=3)
1.60 0.053
ITU requirement 1.10 0.040
Table 10.2.4.2-2 Performance of DL MU-MIMO 8 x 2 (C/E) (RMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 3.36 0.092
Source 10 3.33 0.106
Rel-8 SU-MIMO
(4 x 2, L=3)
1.60 0.053
ITU requirement 1.10 0.040
Table 10.2.4.2-3 shows the spectrum efficiency results of 4-by-2 CS/CB-CoMP schemes with eNB antenna
configuration (C). Results show that the CS/CB-CoMP scheme satisfies the ITU-R requirements and provides
significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 55 Release 9
Table 10.2.4.2-3 Performance of DL CS/CB-CoMP 4 x 2 (C) (RMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 3 3.42 0.109
Rel-8 SU-MIMO
(4 x 2, L=3)
1.60 0.053
ITU requirement 1.10 0.040
10.2.4.3 FDD, Uplink (RMa)
Table 10.2.4.3-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration
(A). Results show that the 2-by-4 SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain
compared with the 1-by-4 SU-MIMO scheme in Rel-8.
Table 10.2.4.3-1 Performance of UL SU-MIMO 2 x 4 (A) (RMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.32 0.129
Source 3 2.45 0.129
Source 5 2.27 0.126
Source 12 1.90 0.062
Source 14 2.46 0.129
Source 18 1.90 0.086
Rel-8 SIMO
(1 x 4)
1.80 0.089
ITU requirement 0.70 0.015
Tables 10.2.4.3-2 and 10.2.4.3-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna
configuration (A) and 2-by-4 CoMP schemes with eNB antenna configuration (A), respectively. Results show that the
performance of the CoMP schemes far exceed the ITU requirements.
Table 10.2.4.3-2 Performance of UL CoMP 1 x 4 (A) (RMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.31 0.129
Source 18 1.75 0.086
Rel-8 SIMO
(1 x 4)
1.80 0.089
ITU requirement 0.70 0.015
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 56 Release 9
Table 10.2.4.3-3 Performance of UL CoMP 2 x 4 (A) (RMa, FDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.58 0.154
Source 18 2.02 0.098
Rel-8 SIMO
(1 x 4)
1.80 0.089
ITU requirement 0.70 0.015
10.2.4.4 TDD, Uplink (RMa)
Table 10.2.4.4-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration
(A). Results show that the 2-by-4 SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain
compared with the 1-by-4 SU-MIMO scheme in Rel-8.
Table 10.2.4.4-1 Performance of UL SU-MIMO 2 x 4 (A) (RMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.25 0.127
Source 3 2.27 0.118
Rel-8 SIMO
(1 x 4)
1.75 0.086
ITU requirement 0.70 0.015
Tables 10.2.4.4-2 and 10.2.4.4-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna
configuration (A) and 2-by-4 CoMP schemes with eNB antenna configuration (A), respectively. Results show that multi
point reception provides performance enhancement compared with the Rel-8 1-by-4 SIMO scheme.
Table 10.2.4.4-2 Performance of UL CoMP 1 x 4 (A) (RMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.23 0.125
Rel-8 SIMO
(1 x 4)
1.75 0.086
ITU requirement 0.70 0.015
Table 10.2.4.4-3 Performance of UL CoMP 2 x 4 (A) (RMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 1 2.52 0.153
Rel-8 SIMO
(1 x 4)
1.75 0.086
ITU requirement 0.70 0.015
Table 10.2.4.4-4 shows the spectrum efficiency results of 1-by-8 MU-MIMO schemes with eNB antenna configuration
(C/E). Results show that additional eNB receive antennas contributes to the further performance enhancement.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 57 Release 9
Table 10.2.4.4-4 Performance of UL MU-MIMO 1 x 8 (C/E) (RMa, TDD)
Cell spectral
efficiency
(bit/sec/Hz/cell)
Cell-edge user
spectral efficiency
(bit/sec/Hz/user)
Source 2 2.64 0.101
Rel-8 SIMO
(1 x 4)
1.73 0.077
ITU requirement 0.70 0.015
Annex A: Simulation model
A.1 Link simulation Scenarios
Link simulation analyses with (8x8) DL MIMO configuration can be conducted in relation to studies concerning the
peak data rate requirements set forth in [36.913].
A.2 System simulation Scenarios
A.2.1 System simulation assumptions
A.2.1.1 Reference system deployments
This section describes the reference system deployments to use for the different system evaluations.
A.2.1.1.1 Homogeneous deployments
The minimum set of simulation cases is given in Table A.2.1.1-1 along with additional assumptions related to carrier
frequency (CF), Inter-site distance (ISD), operating bandwidth (BW), penetration loss (PLoss) and UE speed.
For 3GPP cases only, the system simulation baseline parameters for the macro-cell deployment model are as specified
in [TR 25.814], with the modifications given in Table A.2.1.1-2.
For the ITU cases, simulation parameters should be aligned with the ITU guidelines in [IMT Eval], some of which are
reflected in Table A.2.1.1-1. Note that [IMT.EVAL] section-8 defines different antenna horizontal and vertical pattern
from those defined in Table A.2.1.1-2 which is for 3GPP case evaluation only.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 58 Release 9
Table A.2.1.1-1 – E-UTRA simulation case minimum set
Simulation CF ISD BW PLoss Speed Additional
Simulation
Cases (GHz) (meters) (MHz) (dB) (km/h) Parameters
3GPP case 1 2.0 500 FDD:10+10
TDD: 20
20 3 Table A.2.1.1-2
and 25.814
3GPP case 1 extended 2.0 500 FDD:80+40*
TDD: 80*
20 3 Table A.2.1.1-2
and 25.814
3GPP case 3 2.0 1732 FDD:10+10
TDD: 20
20 3 Table A.2.1.1-2
and 25.814
ITU Indoor,
Indoor hotspot scenario
3.4 60 FDD:20+20
TDD: 40*
N.A. 3 [IMT Eval]
Section 8
ITU Indoor extended,
Indoor hotspot scenario
3.4 60 FDD:80+40*
TDD: 80*
N.A. 3 [IMT Eval]
Section 8
ITU Microcellular,
Urban micro-cell scenario
2.5 200 FDD:10+10
TDD: 20
See [IMT
Eval]
Annex.1
3 [IMT Eval]
Section 8
ITU Base coverage urban,
Urban macro-cell scenario
2.0 500 FDD:10+10
TDD: 20
See [IMT
Eval]
Annex.1
30 [IMT Eval]
Section 8
ITU Base coverage urban
extended,
Urban macro-cell scenario
2.0 500 FDD:80+40*
TDD: 80*
See [IMT
Eval]
Annex.1
30 [IMT Eval]
Section 8
ITU High speed,
Rural macro-cell scenario
0.8 1732 FDD:10+10
TDD: 20
See [IMT
Eval]
Annex.1
120 [IMT Eval]
Section 8
(*) Pending availability of applicable channel model.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 59 Release 9
Table A.2.1.1-2 – 3GPP Case 1 and 3 (Macro-cell)
system simulation baseline parameters modifications as compared to TR 25.814
Parameter Assumption
Antenna pattern (horizontal)
(For 3-sector cell sites with fixed antenna
patterns)
( )
1
1
]
1

¸

,
_

¸
¸
− ·
m
dB
H
A A , 12 min
2
3
ϕ
ϕ
ϕ
dB 3
ϕ = 70 degrees, Am = 25 dB
Antenna pattern (vertical)
(For 3-sector cell sites with fixed antenna
patterns)
( )
1
1
]
1

¸

,
_

¸
¸ −
− ·
v
dB
etilt
V
SLA A , 12 min
2
3
θ
θ θ
θ
dB 3
θ = 10, SLAv = 20 dB
The parameter
etilt
θ is the electrical antenna downtilt. The value for
this parameter, as well as for a potential additional mechanical tilt, is
not specified here, but may be set to fit other RRM techniques used.
For calibration purposes, the values
etilt
θ = 15 degrees for 3GPP
case 1 and
etilt
θ = 6 degrees for 3GPP case 3 may be used.
Antenna height at the base station is set to 32m. Antenna height at
the UE is set to 1.5m.
Combining method in 3D antenna pattern
( ) ( ) ( ) [ ] { }
m V H
A A A A , min , θ ϕ θ ϕ + − − ·
Channel model 3GPP Spatial Channel Model (SCM) [TR 25.996]
For single transmit antenna evaluations, the Typical Urban (TU)
channel model may be used
Total BS TX power (Ptotal) 43dBm – 1.25, 5MHz carrier,
46/49dBm – 10, 20MHz carrier
Some evaluations to exploit carrier aggregation techniques may use
wider bandwidths e.g. 60 or 80 MHz (FDD). For these evaluations
[49 dBm] Total BS Tx power should be used.
UE power class 23dBm (200mW)
This corresponds to the sum of PA powers in multiple Tx antenna
case
In addition to the antenna bore-sight
orientation in TR25.814 (center direction
points to the flat side), an optional orientation
as shown can be used if needed in
Coordinated Multipoint study (i.e., point to
corners) for 3GPP internal evaluations

A.2.1.1.2 Heterogeneous deployments
Heterogeneous deployments consist of deployments where low power nodes are placed throughout a macro-cell layout.
A subset of the macro-cell layouts described in section A.2.1.1.1 could be used for heterogeneous network deployments
evaluation. For calibration purpose, the following cases should be used
- Case 1
- Case 3- Rural/high speed
To assess the benefit of adding low-power nodes to become a heterogeneous network, performance comparison should
be made to homogeneous macro-cell only deployment.
The categorization of the low power nodes is as described in Table A.2.1.1.2-1.
Table A.2.1.1.2-1. Categorization of new nodes
Backhaul Access Notes
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 60 Release 9
Remote radio
head (RRH)
Several µs latency to macro Open to all UEs Placed indoors or outdoors
Pico eNB (i.e.
node for
Hotzone cells)
X2 Open to all UEs
Placed indoors or outdoors. Typically
planned deployment.
HeNB (i.e. node
for Femto
cells)
No X2 as baseline (*)
Closed Subscriber
Group (CSG)
Placed indoors. Consumer deployed.
Relay nodes
Through air-interface with a
macro-cell (for in-band RN
case)
Open to all UEs Placed indoors or outdoors
Note: The reference to new nodes in this TR and its corresponding characteristics are applicable to evaluations in
this TR only.
(*): Baseline is in accordance to Rel-8/9 assumption. Evaluations with interference management for HeNBs (via
X2 or other means) allowed to assess interference management benefits
Table A.2.1.1.2-2 presents the baseline deployment scenario for Heterogeneous network.
Table A.2.1.1.2-2. Heterogeneous network deployment scenario
Case Environment Deployment Scenario Non-traditional node
5.1 Macro + Indoor Macro + femtocell femtocell
5.2 Macro + indoor relay Indoor relay
5.3 Macro + indoor RRH/Hotzone e.g. indoor pico
6.1 Macro + Outdoor Macro + outdoor relay Outdoor relay
6.2 Macro + outdoor RRH/Hotzone e.g., outdoor pico
Note 1: Priorities are as follows:
1. Indoor HeNB clusters
2. Outdoor Hotzone cells with configuration #1 and #4 (in Table A.2.1.1.2-4)
3. Indoor Hotzone scenario (RAN4 femto or pico models could be used)
4. Other scenarios can be studied with lower priority
Note 2: Relay deployment scenario (5.2, 6.1) are studied separately.
Table A.2.1.1.2-3 presents the baseline parameters for initial evaluations in heterogeneous networks. More
detailed modelling of new nodes propagation and channel model based on IMT.EVAL should be considered for
performance evaluation at a later stage.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 61 Release 9
Table A.2.1.1.2-3. Heterogeneous system simulation baseline parameters
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 62 Release 9
Parameter Assumption
RRH / Hotzone Femto Relay
Nodes per macro-cell RRH/Hotzone, and outdoor relay:
1, 2, 4 or 10 (nodes)
Femto and indoor relay:
1 (cluster)
Note: The number of HeNB and indoor relay nodes in each cluster is FFS.
Distance-dependent
path loss from new
nodes to UE*1
For outdoor RRH/Hotzone
Model 1:
Macro to UE:
L= 128.1+37.6log10(R)
Pico to UE:

R L
10
log 7 . 36 7 . 140 + ·
for 2GHz, R in km
Model 2:
Macro to UE:
PLLOS(R)= 103.4+24.2log10(R)
Model 1: 5x5 Grid
Femto to UEs
inside the same
cluster
R L
10
log 30 127 + ·
Other links
L=
128.1+37.6log10(R)
for 2GHz,
R in km, the
Macro to UE:
PLLOS(R)= 103.4+24.2log10(R)
PLNLOS(R)= 131.1+42.8log10(R)
For 2GHz, R in km.
Case 1:
Prob(R)=min(0.018/R,1)*(1-exp(-
R/0.063))+exp(-R/0.063)
Case 3 (Suburban):
Prob(R)=exp(-(R-0.01)/0.2)
Case 3 (Rural/ Suburban):
Prob(R)=exp(-(R-0.01)/1.0)
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 63 Release 9
PLNLOS(R)= 131.1+42.8log10(R)
For 2GHz, R in km.
Penetration loss 20dB
Case 1:
Prob(R)=min(0.018/R,1)*(1-exp(-
R/0.063))+exp(-R/0.063)
Case 3 (Suburban):
Prob(R)=exp(-(R-0.01)/0.2)
Case 3 (Rural/ Suburban):
Prob(R)=exp(-(R-0.01)/1.0)
Pico to UE:
PLLOS(R)=103.8+20.9log10(R)
PLNLOS(R)=145.4+37.5log10(R)
For 2GHz, R in km
Case 1: Prob(R)=0.5-
min(0.5,5exp(-0.156/R))+min(0.5,
5exp(-R/0.03))
Case 3: Prob(R)=0.5-
min(0.5,3exp(-0.3/R))+min(0.5,
3exp(-R/0.095))
For indoor RRH/Hotzone, see
Table 2.1.1.5-1/2
number of floors in
the path is
assumed to be 0.
Model 1: Dual strip
Model
See Table
A.2.1.1.2-7 and
A.2.1.1.2-8
Model 2
See Table
A.2.1.1.2-7 and
A.2.1.1.2-8
Macro to relay:
Relay with outdoor donor
antenna:
PLLOS(R)=100.7+23.5log10(R)
PLNLOS(R)= 125.2+36.3log10(R)
For 2GHz, R in km.
Prob(R) based on ITU models:
Case 1:
Prob(R)=min(0.018/R,1)*(1-exp(-
R/0.072))+exp(-R/0.072)
Case 3 (Suburban):
Prob(R)=exp(-(R-0.01)/0.23)
Case 3 (Rural/ Suburban)
Prob(R)=exp(-(R-0.01)/1.15)
Note 1: Bonus for donor macro
(from each of its sectors) to relay
for optimized deployment by site
planning optimization
methodology described in
[A.2.1.1.4].
Note 2: Higher probability of LOS
shall be reflected in consideration
of the height of RN antenna and
site planning optimization.
described in [A.2.1.1.4].
Note3: If link from donor Macro to
optimized relay site is LOS, the
links from other macros to
optimized relay site could be LOS
or NLOS, else all interference
links from other macros are
NLOS.
Relay with indoor donor antenna:
PLLOS(R)= 103.4+24.2log10(R)
PLNLOS(R)= 131.1+42.8log10(R)
For 2GHz, R in km
Case 1:
Prob(R)=min(0.018/R,1)*(1-exp(-
R/0.063))+exp(-R/0.063)
Case 3: Prob(R)=exp(-(R-
0.01)/1.0)
Note 4: Higher probability of LOS
shall be reflected in consideration
of the height of RN antenna
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 64 Release 9
Relay to UE:
Relay with outdoor coverage
antenna:
PLLOS(R)=103.8+20.9log10(R)
PLNLOS(R)=145.4+37.5log10(R)
For 2GHz, R in km
Case 1: Prob(R)=0.5-
min(0.5,5exp(-0.156/R))+min(0.5,
5exp(-R/0.03))
Case 3: Prob(R)=0.5-
min(0.5,3exp(-0.3/R))+min(0.5,
3exp(-R/0.095))
Note 1: this path loss model
assumes in-band relay.
Simulations for out-of-band relay
should re-examine this
assumption.
Relay with indoor coverage
antenna:
Model 1: 5x5 Grid
Relay to UEs inside the same
cluster
R L
10
log 30 127 + ·
Relay to UEs in different clusters
L= 128.1+37.6log10(R)
for 2GHz,
R in km, the number of floors in
the path is assumed to be 0.
Model 1: Dual strip Model
Use the HeNB to UE model in
Table A.2.1.1.2-7 and A.2.1.1.2-8
Model 2
Use the HeNB to UE model in
Table A.2.1.1.2-7 and A.2.1.1.2-8
Lognormal Shadowing Similar to UMTS 30.03, B 1.41.4 [ETSI TR 101 112]
Shadowing standard
deviation*2
For outdoor RRH/Hotzone
for pathloss model 1:
10 dB
for pathloss model 2 : ITU-R
M.2135 (i.e., according to LOS,
Model 1:
5x5 Grid
10dB for Link
between HeNB and
HeNB UE.
Macro to relay
Relay with outdoor donor
antenna: 6 dB
Relay with indoor donor antenna:
8 dB
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 65 Release 9
NLOS)
For indoor RRH/Hotzone, see
Table 2.1.1.5-1/2
8dB for other links
Dual strip:
4dB for Link
between HeNB and
HeNB UE.
8dB for other links.

Model 2 : 3dB for
LOS Link between
HeNB and HeNB
UE.
4dB for NLOS Link
between HeNB and
HeNB UE.
8dB for NLOS
other links. FFS for
LOS other links.
Relay to UE:
Relay with outdoor coverage
antenna:
10 dB
Relay with indoor coverage
antenna:
Model 1:
5x5 Grid
10 dB for link between relay and
relay UE.
8dB for other links
Dual strip:
4dB for link between relay and
relay UE.
8dB for other links.

Model 2: 3dB for LOS Link
between relay and relay UE.
4dB for NLOS Link between relay
and relay UE.
8dB for NLOS other links. FFS for
LOS other links.
Shadowing
correlation
Between
cells*3
For outdoor RRH/Hotzone
0.5
For indoor RRH/Hotzone
0
0 Relay with outdoor coverage
antenna:0.5
Relay with indoor coverage
antenna: 0
Between
sectors
N/A N/A N/A
Penetration Loss For outdoor RRH/Hotzone
20 dB for Case 1,3; See ITU.Eval
for ITU Rural
above for both "Macro to UE"
and "RRH/Pico to UE"
For indoor RRH/Hotzone, see
Table 2.1.1.5-1/2
Model 1:
5x5 Grid
Femto to UEs
inside the same
cluster: 0 dB
All other links: 20
dB.
Dual-strip: see
Table A.2.1.1.2-
7/8.
Model 2: see Table
A.2.1.1.2-7/8
Macro to UE
Outdoor relay: 20 dB.
Indoor relay
Model 1:
5x5 Grid: 20 dB.
Dual-strip: 20 dB if the UE is
indoors, 0 dB if the UE is
outdoors.
Model 2: 20 dB if the UE is
indoors, 0 dB if the UE is
outdoors.
Macro to relay:
Relay with outdoor donor
antenna: 0 dB
Relay with indoor donor antenna:
5 dB
Relay with outdoor coverage
antenna to UE: 20 dB for Case
1,3; See ITU.Eval for ITU Rural
Relay with indoor coverage
antenna to UE:
Model 1:
5x5 Grid
Relay to UEs inside the same
cluster:0 dB
Relay to UEs in different clusters:
20 dB.
Dual-strip: see the HeNB to UE
model in Table A.2.1.1.2-7/8.
Model 2: see the HeNB to UE
model in Table A.2.1.1.2-7/8.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 66 Release 9
Antenna pattern
(horizontal)
( ) 0 · θ A dB (omnidirectional) ( ) 0 · θ A dB
(omnidirectional)
See Table 2.1.1.4-3
See Table 2.1.1.4-3
Carrier Frequency CF= 2GHz for case 1 and case 3
CF = 0.8GHz for high speed rural
Channel model If fast fading modelling is disabled in system level simulations for relative evaluations, the
impairment of frequency-selective fading channels shall be captured in the physical layer
abstraction. For SIMO, the physical layer abstraction is based on TU link curves. For MIMO,
the physical layer abstraction is FFS.
UE speeds of interest Case 1 and Case 3: 3 km/h Rural high speed: 120 km/h for UEs served by macro, RRH,
hotzone or relay nodes. 3 km/h for UEs served by femto cells.
Doppler of relay-macro
link
N/A N/A Jakes spectrum with [5]Hz
for NLOS component. LOS
component [K=10dB].
Total BS TX power
(Ptotal)
Case1: 24, 30 dBm – 10MHz carrier
Case3: 24, 30, 37 dBm – 10MHz carrier
(37dBm is outdoor only)
20 dBm – 10MHz
carrier
See Table 2.1.1.4-3
See Table 2.1.1.4-3
UE power class 23dBm (200mW)
This corresponds to the sum of PA powers in multiple Tx antenna case
Inter-cell Interference
Modelling
UL: Explicit modelling (all cells occupied by UEs),
DL: Explicit modelling else cell power = Ptotal
Antenna configuration
2 tx , 2 rx antenna ports, or 4 tx , 4 rx
antenna ports
2 tx , 2 rx antenna
ports, or 4 tx , 4 rx
antenna ports
See Table 2.1.1.4-3
See Table 2.1.1.4-3
Antenna gain +
connector loss
[Motorola: reference
for these values?]
5dBi 5dBi
See Table 2.1.1.4-3
See Table 2.1.1.4-3
Placing of new nodes
and Ues
See Table A.2.1.1.2-4 See Table A.2.1.1.2-
6
See Table A.2.1.1.2-4
Minimum distance
between new node
and regular nodes
>=75m
Minimum distance
between UE and
regular node
>= 35m
Minimum distance
between UE and new
node (RRH/Hotzone,
Femto, Relay)
For outdoor RRH/Hotzone
> 10m
For indoor RRH/Hotzone
>= 3m
>= 3m Outdoor relay:> 10m
Indoor relay: >= 3m
Minimum distance
among new nodes
40 m 40 m cluster radious 40 m
*1 Outdoor RRH/Hotzone model 1 is based on TR 25.814 and IMT.EVAL UMi NLOS model, outdoor relay and
outdoor RRH/Hotzone model 2 LOS and NLOS path loss models are based on field measurements and LOS/NLOS
probability functions for macro to UE and macro to relay are based on ITU models and probability functions for relay
or pico to UE are based on field measurements; femto and indoor relay path loss is based on ITU-R M1225 single floor
indoor office model. Outdoor relay antenna height for both access and backhaul link is 5m. Path loss models for case 3
with 10m antenna height are FFS.
*2 Shadowing fading value for relays is applicable to NLOS component of the path loss. Value for LOS component is
FFS.
*3 Cells including macro cells of the overlay network and new nodes.
Fast fading may be modelled using any of the following:
- No fast fading as in current TR
- Fast fading with TU and fixed correlation matrix
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 67 Release 9
- Fast fading with ITU/SCM models or possible simplifications [ref. R4-091103] could also be used. (Detailed
proposals to be discussed, e.g., relevant propagation model to use with these.)
For preliminary simulation, all fast fading channel models can be applied. The baseline fast fading model to be used for
final evaluation results should be discussed in future (or in WI).
Table A.2.1.1.2-4. Placing of new nodes and UEs
Configuratio
n
UE density across
macro cells*
UE distribution
within a macro cell
New node distribution
within a macro cell
Comments
1 Uniform
25/macro cell
Uniform Uncorrelated Capacity enhancement
2 Non-uniform
[10 – 100]/macro cell
Uniform Uncorrelated Sensitivity to non-uniform
UE density across macro
cells
3 Non-uniform
[10 – 100]/macro cell
Uniform Correlated** Cell edge enhancement
4a, 4b Non-uniform*** Clusters Correlated** Hotspot capacity
enhancement
* New node density is proportional to the UE density in each macro cell. UE density is defined as the number of
UEs in the geographic area of a macro cell.
**Relay and hotzone nodes, often deployed by planning, see section A2.1.1.4.
*** Clustered UE Placement for Hotzone cells:
- Fix the total number of users, Nusers, dropped within each macro geographical area, where Nusers is 30 or 60 in
fading scenarios and 60 in non-fading scenarios.
- Randomly and uniformly drop the configured number of low power nodes, N, within each macro geographical
area (the same number N for every macro geographical area, where N may take values from {1, 2, 4, 10}).
- Randomly and uniformly drop Nusers_lpn users within a 40 m radius of each low power node, where
¸ ]
/N N P N
users
hotspot
users_lpn
⋅ · with P
hotspot
defined in Table A.2.1.1.2-5, where P
hotspot
is the fraction of all
hotspot users over the total number of users in the network.
- Randomly and uniformly drop the remaining users, Nusers - Nusers_lpn*N, to the entire macro geographical area of
the given macro cell (including the low power node user dropping area).
Table A.2.1.1.2-5. Configuration #4a and #4b parameters for clustered user dropping
Configuration Nusers N P
hotspot
Configuration #4a* 30 or 60 1
2
4
10
1/15
2/15
4/15
2/3
Configuration #4b 30 or 60 1
2*
4*
2/3
2/3*
2/3*
* Baseline for Configurations #4a and #4b.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 68 Release 9
Table A.2.1.1.2-6. Placing of femto cells and UEs
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 69 Release 9
Configuratio
n
Macro-femto
Deployment
Placing of nodes Placing of UEs
1 Independent channel Clustered Random placing of UEs within
X meters of the femto cell
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 70 Release 9
2 Co-channel Clustered Random placing of UEs within
X meters of the femto cell
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 71 Release 9
* Femto cell with 5x5 grid or dual-strip apartment blocks
* Non-uniform macro-UE drop in a femto cluster.
(Note that this does not preclude that one femto cluster may contain no UEs)
* Fraction of Macro UEs in Femto Cluster: 35% and/or 80%
Table A.2.1.1.2-7. Indoor femto Channel models (dual strip model): Suburban deployment
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 72 Release 9
Cases Path Loss (dB)
Fast Fading (when fast
fading in both frequency and
spatial domains is modeled)
UE to
macro
BS
(1) UE is
outside:
PL(R)
Model 1:
PL (dB) =15.3 + 37.6log10R
For 2GHz, R in m.
Model2:
PLLOS(R)= 30.8+24.2log10(R)
PLNLOS(R)= 2.7+42.8log10(R)
For 2GHz, R in m.
Prob(R)=exp(-(R-10)/1000)
RMa
(2) UE is
inside a
house
Model1:
PL (dB) =15.3 + 37.6log10R + Low, R in m
Model2:
PLLOS(R)= 30.8+24.2log10(R) + Low
PLNLOS(R)=2.7+42.8log10(R) + Low
For 2GHz, R in m.
Prob(R)=exp(-(R-10)/1000)
RMa
UE to
HeNB
(3) Dual-
stripe
model:
UE is
inside the
same
house as
HeNB
PL (dB) = 38.46 + 20 log10R + 0.7d2D,indoor+ 18.3 n
((n+2)/
(n+1)-0.46)

R and d2D,indoor are in m
n is the number of penetrated floors
In case of a single-floor house, the last term is not
needed
InH, LOS or NLOS depends
on whether line-of sight
from UE to HeNB;
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 73 Release 9
(4) Dual-
stripe
model:
UE is
outside
Model 1:
PL (dB) = max(15.3 + 37.6log10R, 38.46 + 20log10R) +
0.7d2D,indoor
+ 18.3 n
((n+2)/(n+1)-0.46)
+ Low
R and d2D,indoor are in m
Model 2:
PL (dB) = max(2.7+42.8 log10 R, 38.46 + 20log10R) +
0.7d2D,indoor
+ 18.3 n
((n+2)/(n+1)-0.46)
+ Low
R and d2D,indoor are in m
InH (NLOS)
(5) Dual-
stripe
model:
UE is
inside a
different
house
Model 1:
PL(dB) = max(15.3 + 37.6log10R, 38.46 + 20log10R) +
0.7d2D,indoor
+ 18.3 n
((n+2)/(n+1)-0.46)
+

Low,1 + Low,2
R and d2D,indoor are in m
Model 2:
PL(dB) = max(2.7+42.8 log10 R, 38.46 + 20log10R) +
0.7d2D,indoor
+ 18.3 n
((n+2)/(n+1)-0.46)
+

Low,1 + Low,2
R and d2D,indoor are in m
InH (NLOS)
R is the Tx-Rx separation
Low is the penetration loss of an outdoor wall, which is 20dB.
In Case (3), the path loss is modeled by free space loss, penetration loss due to internal walls and floors. The
loss due to internal walls is modeled as a log-linear value, equal to 0.7dB/m.
In Case (4), the path loss modeling takes account of case (2) and case (3). d2D,indoor is the distance inside the
house.
In Case (5), d2D,indoor is the total distance inside the two houses. Low,1 and Low,2 are the penetration losses of
outdoor walls for the two houses.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 74 Release 9
Table A.2.1.1.2-8. Indoor femto Channel models (dual strip model): Urban deployment
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 75 Release 9
Cases Path Loss (dB)
Fast Fading(when fast
fading in both
frequency and spatial
domains is modeled)
UE to
macro
BS
(1) UE is
outside
PL(R)
Model1:
PL (dB) =15.3 + 37.6log10R, R in m
Model2:
PLLOS(R)= 30.8+24.2log10(R)
PLNLOS(R)= 2.7+42.8log10(R)
For 2GHz, R in m.
Prob(R)=min(18/R,1)*(1-exp(-R/63))+exp(-R/63)
UMa
(2) UE is
inside an
apt
Model1:
PL (dB) =15.3 + 37.6log10R + Low, R in m
Model2:
PLLOS(R)= 30.8+24.2log10(R) + Low
PLNLOS(R)= 2.7+42.8log10(R) + Low
For 2GHz, R in m
Prob(R)=min(18/R,1)*(1-exp(-R/63))+exp(-R/63)
UMa
UE to
HeNB
(3) Dual-
stripe
model: UE
is inside
the same
apt stripe
as HeNB
PL (dB) = 38.46 + 20 log10R + 0.7d2D,indoor+ 18.3 n
((n+2)/(n+1)-
0.46)
+ q*Liw
R and d2D,indoor are in m
n is the number of penetrated floors
q is the number of walls separating apartments between
UE and HeNB
In case of a single-floor apt, the last term is not needed
InH, LOS or NLOS
depends on whether
line-of sight from UE
to HeNB;
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 76 Release 9
(4) Dual-
stripe
model: UE
is outside
the apt
stripe
Model 1:
PL (dB) = max(15.3 + 37.6log10R, 38.46 + 20log10R) +
0.7d2D,indoor
+ 18.3 n
((n+2)/(n+1)-0.46)
+ q*Liw + Low
Model 2:
PL (dB) = max(2.7+42.8 log10 R, 38.46 + 20log10R) +
0.7d2D,indoor
+ 18.3 n
((n+2)/(n+1)-0.46)
+ q*Liw + Low
R and d2D,indoor are in m
q is the number of walls separating apartments between
UE and HeNB
InH (NLOS)
(5) Dual-
stripe
model: UE
is inside a
different
apt stripe
Model 1:
PL(dB) = max(15.3 + 37.6log10R, 38.46 + 20log10R) +
0.7d2D,indoor
+ 18.3 n
((n+2)/(n+1)-0.46)
+

q*Liw + Low,1 + Low,2
Model 2:
PL(dB) = max(2.7+42.8 log10 R, 38.46 + 20log10R) +
0.7d2D,indoor
+ 18.3 n
((n+2)/(n+1)-0.46)
+

q*Liw + Low,1 + Low,2
R and d2D,indoor are in m
q is the number of walls separating apartments between
UE and HeNB
InH (NLOS)
Liw is the penetration loss of the wall separating apartments, which is 5dB.
The term 0.7d2D,indoor takes account of penetration loss due to walls inside an apartment.
Low is the penetration loss of an outdoor wall, which is 20dB.
Low,1 and Low,2 are the penetration losses of outdoor walls for the two houses.
A.2.1.1.3 Assumptions for Coordinated Multi point Transmission and Reception
Evaluations
Performance evaluations should at least provide details related to:
- Cooperating scheduler
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 77 Release 9
- CoMP category
- Feedback assumption and feedback impairment modelling
- Backhaul assumptions
- Time/frequency synchronization assumptions
- Transmission modes:
- MU-MIMO and/or SU-MIMO operation in conjunction with CoMP
- Selection of transmission mode (assumptions on how dynamic or semi-static the transmission mode can be
selected)
- Creation and maintenance of CoMP sets:
- Assumptions on CoMP sets definition and creation
- fixed vs. adaptive clusters, size of cluster…
Geometry cdf for the CoMP UE should be provided where appropriate, compared to the geometry cdf for a non-CoMP
UE.
The performance of downlink/uplink multi-point transmission and reception, and advanced ICIC techniques is sensitive
to the backhaul capacity and latency. In general, the backhaul latency could be classified into the following categories
- Minimal latency (in the order of μs) for eNB to RRH links
- Low latency (<1 ms) associated with co-located cells or cells connected with fibre links and only limited number
of routers in between
- Typical inter-cell latency associated with X2 interfaces.
The X2 backhaul latency, or more generally latency between new nodes, or new nodes and eNBs, or between eNBs, is
highly deployment dependent such as whether there is a dedicated X2 fibre network or a generic IP network.
The proponents should describe and justify the model assumed in particular studies.
A.2.1.1.4 Assumptions for Relay Evaluations
The evaluation scenarios for relay is summarized as two basic scenarios as follows,
Table A.2.1.1.4-1. Evaluation scenarios for relay
CASE Scenarios ISD[m] Carrier[GHz]
3GPP case 1.Relay Urban Macro 500 2.0
3GPP case 3.Relay Rural Area 1732 2.0
3GPP case 1.Indoor relay Urban Macro 500 2.0
3GPP case 3.Indoor relay Rural Area 1732 2.0
For 3GPP case 1&3.Relay scenarios, the placement of relay also regarded as site planning could be taken by two major
steps,
1) Virtual Relay Placement: A virtual relay is placed trying to enhance the cell edge throughput or overall cell
throughput.
[TBD]
2) Relay Site planning: Finding an optimal place among N candidate relay sites around the virtual relay which offers
optimization of shadow fading, LOS probability and etc.
Relays were placed on positions they are most needed in terms of SINR (geometry). Positions, where a relay
placement would result in the highest geometry gain, were found in an exhaustive search within N candidate relay
sites around the virtual relay.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 78 Release 9
The site planning procedure provides benefit on backhaul SINR (geometry). TWO alternatives to show this benefit
in simulation are considered with respect to
Alternative 1: Adding bonus to path loss formula.
This process offers optimization of shadow fading, LOS probability and etc.
The corrections of site planning with respect to macro-relay pathloss, LOS probability and shadowing standard
deviation are listed in Table 2.1.1.4-2.
Table A.2.1.1.4-2. Corrections of site planning (alternative 1)
No site planning Correction after site planning
Macro-relay Path Loss
For LOS: PLLOS(R)
For NLOS: PLNLOS(R)
For LOS: PLLOS(R)
For NLOS: PLNLOS(R)-B
Where B=5dB, for donor macro
(from each of its sectors) to relay,
otherwise, for non-donor cell and
non optimized deployment
B=0dB.
Macro-relay LOS probability Prob(R)
1-(1- Prob(R))^N
Where N=3, for donor macro
(from each of its sectors) to relay,
otherwise, for non-donor cell and
non optimized deployment N=1.
Alternative 2: Initialized in a system-level simulation by selecting best N relays according to a proposed site
planning optimization approach.
The site planning optimization should be taken into consideration in a relay placement procedure in step 2. It is a
process of finding an optimal place among N candidate relay sites around the virtual relay which offers benefit to
the performance.
The site planning optimization approach are described as follows,
- N=5 candidate relay sites are considered within a searching area of 50m radius around the virtual relay.
- For simplicity, the candidate relays are randomly placed in the searching area.
- The best relay site is selected based on SINR criteria on the backhaul link.
For 3GPP case 1&3.Indoor relay scenarios, 2 cases are distinguished, depending on the relay node configuration:
- if the relay donor antenna is outdoors, no planning is performed.
- if the relay node is made up of a donor module and a coverage module, both being placed indoors, the donor
module has to be placed near a window to optimize the backhaul link quality. This optimization is reflected by a
penetration loss of 5 dB in Table A.2.1.1.2-3. The coverage module has to be placed in the center of the house to
provide maximum coverage.
For a typical system level simulation, the following configurations are taken into consideration,
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 79 Release 9
Table A.2.1.1.4-3. Typical configuration for simulation
Parameter Description Case 1 Case 3 Case 1/3 Indoor
PRN Max Tx power
30 dBm @ 10 MHz
bandwidth
30 or 37 dBm @ 10 MHz
bandwidth
Downlink:
20 dBm @ 10 MHz
bandwidth
Uplink:
Indoor donor antenna:
23 dBm @ 10 MHz
bandwidth
Outdoor donor antenna:
30 dBm @ 10 MHz
bandwidth
HRN RS antenna height 5m 5m, 10m
Outdoor antenna:
case 1: 5 m
case 3: 5 m, 10 m

Indoor antenna: N/A
Antenna
Configuratio
n
One antenna set
5dBi antenna gain, Omni
2 tx , 2 rx antenna ports,
or 4 tx , 4 rx antenna ports
Use of antenna downtilt
and vertical antenna FFS
5dBi, Omni
2 tx , 2 rx antenna ports,
or 4 tx , 4 rx antenna ports
Use of antenna downtilt
and vertical antenna FFS
N/A
Two antenna sets
Relay-UE link:
5dBi antenna gain,
Omni
( ) 0 · θ A dB
2 tx , 2 rx antenna ports,
Macro-Relay link
7dBi, directional
( )
1
1
]
1

¸

,
_

¸
¸
− ·
m
dB
A A , 12 min
2
3
θ
θ
θ
dB 3
θ
= 70 degrees, Am
= 20 dB.
2 tx , 2 rx antenna ports,
or 4 tx , 4 rx antenna ports
Use of antenna downtilt
and vertical antenna FFS
Relay-UE link:
5dBi antenna gain,
Omni
( ) 0 · θ A dB
or
directional pointing away
from the donor cell
( )
1
1
]
1

¸

,
_

¸
¸
− ·
m
dB
A A , 12 min
2
3
θ
θ
θ
dB 3
θ
= 70 degrees, Am
= 20 dB.
2 tx , 2 rx antenna ports,
Macro-Relay link
7dBi, directional
( )
1
1
]
1

¸

,
_

¸
¸
− ·
m
dB
A A , 12 min
2
3
θ
θ
θ
dB 3
θ
= 70 degrees, Am
= 20dB.
2 tx , 2 rx antenna ports,
or 4 tx , 4 rx antenna ports
Use of antenna downtilt
and vertical antenna FFS
Relay-UE link:
5dBi antenna gain,
Omni
( ) 0 · θ A dB
2 tx , 2 rx antenna ports,
or 4 tx , 4 rx antenna ports
Macro-Relay link
Indoor donor antenna:
5dBi, directional
Outdoor donor antenna:
7dBi, directional
( )
1
1
]
1

¸

,
_

¸
¸
− ·
m
dB
A A , 12 min
2
3
θ
θ
θ
dB 3
θ
= 70 degrees, Am
= 20 dB.
2 tx , 2 rx antenna ports,
or 4 tx , 4 rx antenna ports
NFRN Noise figure 5 dB 5 dB 5 dB
HWRN Hardware loss/cable loss 0 dB 0 dB
Outdoor donor antenna:
2 dB
Indoor donor antenna:
0 dB
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 80 Release 9
A.2.1.1.5 Assumptions for indoor RRH/Hotzone Evaluations
The indoor hotspot scenario consists of single floor of a building as Figure.2.1.1.5-1 which is same with ITU sketch.
The height of the floor is 6 m. The floor contains 16 rooms of 15 m x 15 m and a long hall of 120 m x 20m. Two sites
are placed in the middle of the hall at 30m and 90m with respect to the left side of the building.
Figure.2.1.1.5-1. Sketch of indoor hotspot environment
The following two channel models are for evaluation purpose. Channel model 2 is a simplified channel model which
assumes all users are allocated within buildings.
Table A.2.1.1.5-1. Channel model 1 of indoor RRH/Hotzone
Cases Path Loss (dB)
Shadowing
standard
deviation
Penetration
Loss
Fast
Fading
(when fast
fading in
both
frequency
and spatial
domains is
modelled)*
UE to
macro BS
(1) UE is
outside
PLLOS(R)= 103.4+24.2log10(R)
PLNLOS(R)= 131.1+42.8log10(R)
For 2GHz, R in km.
Case 1: Prob(R)=min(0.018/R,1)*(1-exp(-
R/0.063))+exp(-R/0.063)
Case 3: Prob(R)=exp(-(R-0.01)/1.0)
10dB
0dB
ITU UMa
(2) UE is
inside
20dB
UE to
RRH/Hotz
one
(1) UE is
inside a
different
building
as the
indoor
hotzone
PL(dB) =Max(131.1+42.8log10(R),
147.4+43.3log10(R))
For 2GHz, R in km
10dB
40dB
ITU InH
(NLOS)
(2) UE is
outside
20dB
(3) UE is
inside
the same
building
as the
indoor
hotzone
PLLOS(R)= 89.5 + 16.9log10(R)
PLNLOS(R)= 147.4+43.3log10(R)
For 2GHz, R in km
Prob(R)=
1 0.018
exp( ( 0.018) / 0.027) , 0.018 0.037
0.5 0.037
R
R R
R
≤ ¹
¹
− − < <
'
¹

¹
LOS: 3dB
NLOS: 4dB
0dB ITU InH
* No Fast Fading orTU and fixed correlation matrix can also be used.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 81 Release 9
Table A.2.1.1.5-2. Channel model 2 of indoor RRH/Hotzone
Cases Path Loss (dB)
Shadowing
standard
deviation
Penetration
Loss
Fast
Fading
(when fast
fading in
both
frequency
and spatial
domains is
modelled)*
UE to macro BS
PLLOS(R)= 103.4+24.2log10(R)
PLNLOS(R)= 131.1+42.8log10(R)
For 2GHz, R in km.
Case 1: Prob(R)=min(0.018/R,1)*(1-exp(-
R/0.063))+exp(-R/0.063)
Case 3: Prob(R)=exp(-(R-0.01)/1.0)
10dB 20dB ITU UMa
UE to
RRH/Hotz
one
(1) UE is
outside
the same
building
as the
indoor
hotzone*
*
PL(dB) =Max(131.1+42.8log10(R),
147.4+43.3log10(R))
For 2GHz, R in km
10dB 20dB
ITU InH
(NLOS)
(2) UE is
inside
the same
building
as the
indoor
hotzone
PLLOS(R)= 89.5 + 16.9log10(R)
PLNLOS(R)= 147.4+43.3log10(R)
For 2GHz, R in km
Prob(R)=
1 0.018
exp( ( 0.018) / 0.027) , 0.018 0.037
0.5 0.037
R
R R
R
≤ ¹
¹
− − < <
'
¹

¹
LOS: 3dB
NLOS: 4dB
0dB ITU InH
* No Fast Fading orTU and fixed correlation matrix can also be used.
* * For UE is outside the same building as the indoor hotzone, UE are allocated within buildings which
either indoor hotzone nodes are deployed or not.
A.2.1.2 Channel models
Annex B describes the IMT-Advanced Channel Models, which are specified in the IMT.EVAL of ITU-R[15].
A.2.1.3 Traffic models
Traffic models for system performance evaluations are given in Table A.2.1.3-1. System throughput studies shall be
assessed using full-buffer traffic model capturing continuous traffic and non-varying interference. Additionally,
evaluations with time-varying interference shall be carried out using bursty traffic models. Table A.2.1.3-1 proposes
FTP traffic models to exercise system performance studies in bursty traffic.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 82 Release 9
Table A.2.1.3-1. Traffic Models
Traffic Models Model Applies to
Full buffer DL and UL.
Continuous traffic.
Non-full buffer
FTP models
DL and UL.
Bursty traffic.
VoIP DL and UL
Real time services
A.2.1.3.1 FTP traffic models
Two FTP traffic models are considered as non-full buffer traffic models. Tables A.2.1.3-2 and A.2.1.3-3 show the
parameters for FTP traffic model 1 and model 2, respectively. Figure A.2.1.3.1-1 and A.2.1.3.1-2 illustrate the user
arrival of traffic model 1 and 2, respectively. Baseline model is Model 1 with file size of 2 Mbytes, however Model 1
with file size of 0.5 Mbytes and Model 2 with file size of 0.5 Mbytes can be also evaluated.
Table A.2.1.3.1-1. FTP Traffic Model 1
Parameter Statistical Characterization
File size, S 2 Mbytes (0.5 Mbytes optional)
(one user downloads a single file)
User arrival rate λ Poisson distributed with arrival rate λ
- Small file size of 0.5 Mbytes can be chosen to speed-up the simulation.
- Simulations are run for various λ to find performance metrics covering at least the range of HM-NCT (See
A.2.1.3.2) that leads to [10%, 50%] of RU (See A.2.1.3.2) in non-CoMP SU-MIMO.
- Possible range of λ: [0.5, 1, 1.5, 2, 2.5] for 0.5 Mbytes, [0.12, 0.25, 0.37, 0.5, 0.625] for 2 Mbytes (See A.2.1.3.4
for more details). Range of λ can further be adjusted.
- The same traffic should be simulated for CoMP and non-CoMP schemes. The above range of λ will cover RU
from 10% to 50% for non-CoMP SU-MIMO
Time
Per cell traffic
S
user1
user2 user3 user4
Figure A.2.1.3.1-1: Traffic generation of FTP Model 1
Table A.2.1.3.1-2. FTP Traffic Model 2
Parameter Statistical Characterization
File Size, S 0.5 Mbytes
Reading Time, D Exponential Distribution, Mean= 5 seconds
PDF: 0 , ≥ ·

D e f
D
D
λ
λ λ = 0.2
Number of users, K Fixed
- Simulations are run for various K to find performance metrics covering at least the range of HM-NCT that leads
to [10%, 50%] of RU in non-CoMP SU-MIMO.
- Possible range of K: [2, 5, 8, 10, 14] (See A.2.1.3.4 for more details). Range of K can further be adjusted.
- The reading time D is the time interval between end of download of previous file and the user request for the
next file.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 83 Release 9
- The same traffic should be simulated for evaluating CoMP and non-CoMP schemes. The above range of K will
cover RU from 10% to 50% for non-CoMP SU-MIMO.
Time
Per-user traffic
S
D
Figure A.2.1.3.1-2: Traffic generation of FTP Model 2
A.2.1.3.2 Performance metrics
The following performance metrics are considered for non-full buffer traffic models.
- Mean, 5, 50, 95 % user throughput
- User throughput = amount of data (file size) / time needed to download data
- time needed to download data starts when the packet is received in the transmit buffer, and ends when the
last bit of the packet is correctly delivered to the receiver
- Served cell throughput
- Served cell throughput = total amount of data for all users / total amount of observation time / number of
cells
- Harmonic mean normalized cell throughput (HM-NCT)
- Harmonic mean normalized cell throughput = served cell throughput / harmonic mean of the full buffer cell
throughput
- Harmonic mean of the full buffer cell throughput = average number of full buffer users per cell / mean( 1 /
full buffer user throughputs)
- The 2% worst and 2% best users are excluded from the harmonic mean
- Full buffer simulations are run with 10 UEs / cell
- Normalized cell throughput (NCT)
- Normalized cell throughput = served cell throughput / full buffer served cell throughput
- Resource utilization (RU)
- Resource utilization = Number of RB per cell used by traffic during observation time / Total number of RB
per cell available for traffic over observation time
- In case of MU-MIMO, one RB allocated to N users within a cell is counted as used N times
A.2.1.3.2.1 Reference points
Possible reference points to compare CoMP, non-CoMP, SU-MIMO and MU-MIMO techniques:
- R1: X% HM-NCT value that leads to 50% RU for SU-MIMO (i.e. X value is derived from the SU-MIMO case,
and is then the same for all techniques)
- R2: Traffic load that gives 50% RU for SU-MIMO
- R3: 45% NCT
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 84 Release 9
Reference point R1 may be a baseline point. However, the results should be obtained at the other reference points to
gain understanding about how these new metrics relate to each other. R1 and R3 capture both the user throughput and
cell capacity gains in a reasonably well utilized network. R2 captures the user throughput gain at a given offered load.
Further refinement of adequate reference points can be considered.
A.2.1.3.3 File dropping criteria
A guideline of modelling file dropping for non-full buffer traffic simulation is given as follows.
- Files may be dropped from the simulation either to preserve stability at high loads, or because the radio
conditions are insufficient for reliable communication.
- Method for ensuring queue stability in high load conditions:
- drop a file if its transfer is not completed within a maximum transfer time T_drop
- T_drop = 8s for S = 0.5 Mbyte
- T_drop = 32s for S = 2 Mbyte
- Methods for handling HARQ retransmission failure:
- model RLC ACK
- if RLC ARQ is not modeled, drop the file after the maximum number of HARQ retransmissions is reached.
- Method to restrict the impact of users in very poor radio conditions:
- if RLC is modelled: Remove from the system the users who experiment a MAC error rate higher than 3%
- a MAC error is defined as a transport block being not correctly received after the maximum number of
retransmissions.
- do not include removed users' throughput nor contribution to the served cell throughput in the simulation
results.
- if RLC is not modeled: user outage is implicitly approximated by file dropping when the maximum number
of HARQ retransmissions is reached.
- Dropping a file means:
- the file is given zero user throughput
- the data in the dropped file is not included in the served cell throughput
- The amount of data dropped or not transmitted shall be recorded separately according to whether it arises from
users removed due to high MAC error rate or other causes, and the corresponding file dropping rates shall be
indicated in the simulation results.
Alternatives methods can be considered, and should be then described.
A.2.1.3.4 Estimation of range ofλ and K
For FTP Model 1
offered traffic =λ* S.
For LTE 2x2 10MHz, the [10%, 50%] range of resource utilization could be covered by the range of offered traffic [2 4
6 8 10] Mbps. For a file size of 500kbyte = 4 Mbit, this corresponds to:
λ= offered_traffic/S = [2, 4, 6, 8, 10]/4 = [0.5, 1, 1.5, 2, 2.5] users/s.
λvalues for file size of 2Mbytes can be computed in the same way.
For model2, the inter-arrival time is given by Tt + D with Tt file transfer time. In this caseλ= 1/(Tt+D). The total offered
traffic per cell is then given by
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 85 Release 9
offered traffic = λ* K * S = K * S / (Tt + D)
Assuming download data rate of 12.5 Mbps, the file transfer (download) time Tt is 0.32 sec for file size of 500 Kbytes =
4 Mbits. By using this rough estimate, offered traffic is K*4/(0.32+5). When the range of offered loads is [2 4 6 8 10],
the corresponding range of K is
K = [2 4 6 8 10]/4*5.32 ~= [2 5 8 10 14] users
A.2.1.4 System performance metrics
For evaluations with full-buffer traffic model, the following performance metrics need to be considered:
- Mean user throughput
- Throughput CDF
- Median and 5% worst user throughput
For evaluations with bursty traffic model, the following performance metrics need to be considered:
- User perceived throughput (during active time), defined as the size of a burst divided by the time between the
arrival of the first packet of a burst and the reception of the last packet of the burst
- Average perceived throughput of a user defined as the average from all perceived throughput for all bursts
intended for this user.
- Tail perceived throughput defined as the worst 5% perceived throughput among all bursts intended for a user
- User perceived throughput CDF (average and/or tail user perceived throughput).
- Percentage of users with [1]% or more dropped packets.
- Median and 5% worst user perceived throughput (average and/or tail user perceived throughput).
- Overall average user throughput defined as average over all users perceived throughput.
For VoIP capacity evaluations, the following performance metrics need to be considered:
- VoIP system capacity in form of the maximum number of satisfied users supported per cell in downlink and
uplink.
- System capacity is defined as the number of users in the cell when more than [95%] of the users are satisfied.
- A VoIP user is in outage (not satisfied) if [98%] radio interface tail latency of the user is greater than [50 ms].
This assumes an end-to-end delay below [200 ms] for mobile-to-mobile communications.
For heterogeneous network performance evaluation, the following performance metrics are the highest priority:
- Existing full buffer and bursty traffic performance metrics
- Throughput CDFs are for all UEs, i.e., macro UEs and HeNB/pico UEs
- Macro cell area throughput
- Fraction of throughput over low power nodes
- Macro and low power node serving UE throughput ratio
The following table should be included along with the simulation assumptions accompanying all results:
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 86 Release 9
Are Throughput Values
based solely on an
assumption of a number
of trials of independent
placing of UEs?
Comments
Yes/No If “Yes,” then state the number of trials, i.e., placing of a group of UEs in cells used. If
“No,” either state the methodology by which confidence interval is achieved as well as
confidence interval and confidence level, or justify the method of user placing in
different trials.
A.2.1.5 Scheduling and resource allocation
Different scheduling approaches have impacts on performance and signalling requirements.
Evaluations should include a high-level description of the scheduling and resource allocation schemes simulated,
including relevant parameter values. For frequency or carrier specific scheduling, and multipoint transmission schemes,
any feedback approach, delay, and feedback error assumptions should also be indicated. For uplink queue-related
scheduling, similar indications should be given for any buffer status feedback.
Evaluations should include fairness, as described in section A.2.1.4 above.
A.2.1.6 Antenna gain for a given bearing and downtilt angle
A.2.1.6.1 Polarized antenna modelling
In case of polarized antennas, the polarization is modelled as angle-independent in both azimuth and elevation, in the
antenna local coordinate system. For a linearly polarized antenna, the antenna element field pattern, in the horizontal
polarization and in the vertical polarization, are given by
( ) ( ) ( ) ( ) ζ θ φ ζ θ φ θ φ
φ φ
cos , , , , A F F · ·
and
( ) ( ) ( ) ( ) ζ θ φ ζ θ φ θ φ
θ θ
sin , , , , A F F · · ,
respectively, where
ζ
is the polarization slant angle
1
and ( ) θ φ, A is the 3D element antenna gain pattern as a
function of azimuth angle,
φ
and elevation angle,θ , which are defined in Figure A.2.1.6.1-1. Note that the vertical
and horizontal field direction are defined in terms of the spherical basis vectors,
θ
ˆ
and φ
ˆ
, respectively (i.e.,
) , ( ) , ( θ φ θ φ
φ
F F
H
·
and ) , ( ) , ( θ φ θ φ
θ
F F
V
· ).
For 2D element antenna gain pattern:
( ) ( ) ( ) ( ) ζ φ ζ φ φ
φ φ
cos , A F F · ·
and
( ) ( ) ( ) ( ) ζ φ ζ φ φ
θ θ
sin , A F F · · ,
1
Hence, an antenna with
0 · ζ
degrees corresponds to a purely horizontally polarized antenna.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 87 Release 9
Figure A.2.1.6.1-1 Definition of spherical angles and spherical unit vectors in a Cartesian coordinate
system, where nˆ is the given direction,
θ
ˆ
and φ
ˆ
are the spherical basis vectors.
A.2.1.6.2 Antenna gain for a given direction and mechanical tilt angle
A global coordinate system with Cartesian coordinates (x,y,z) and a local (“primed”) coordinate system with Cartesian
coordinates (x’,y’,z’) are defined with a common origin, with the y’-axis being parallel to the y-axis; that is, the x-axis
of the global coordinate system, should be aligned with the pointing direction of the sector
2
. The global coordinate
system is oriented with its z-axis along a vertical direction, thus having its xy-plane coinciding with a horizontal plane,
and all directions in space (angles of arrival and departure) are defined in the global coordinate system angular
coordinates θ and φ , see Figure A.2.1.6.1-1.
Assume that an antenna is installed with the antenna aperture normal direction (and antenna main beam peak for a
conventional sector antenna) in the xz-plane and that the antenna radiation pattern is defined in terms of angles ' θand
' φ
in the local coordinate system, with the local coordinate system antenna-fixed.
Mechanical tilt is modeled as a rotation of the antenna-fixed coordinate system around the y-axis. For zero mechanical
tilt the antenna-fixed coordinate system coincides with the global coordinate system. Conventionally, for sector
antennas with cylindrical shape, the antenna radiation pattern defined in the antenna-fixed coordinate system is
measured with the antenna cylinder axis installed to coincide with the z'-axis of the antenna-fixed coordinate system.
Note that electric tilt, and hence beam pointing direction, does not in general affect the choice of antenna-fixed
coordinate system.
By rotating the local coordinate system with respect to the global coordinate system the same effect as mechanical
tilting is attained. Since the antenna pattern is defined in the local coordinate system, which has been rotated with
respect to the global coordinate system, a transformation must be performed to allow evaluation of the tilted antenna
pattern as a function of coordinates in the global coordinate system. This transformation relates the spherical angles (θ
,
φ
) in the global coordinate system to spherical angles ( ' θ,
' φ
) in the local (antenna-fixed) coordinate system and
is defined as follows:
arccos( ' · θ β θ β θ φ cos cos sin sin cos +
), (1)
arg( ' · φ θ φ β θ β θ φ sin sin sin cos cos sin cos j + −
), (2)
where
β
is the mechanical tilt angle around the y-axis as defined in Figure A.2.1.6.2-1.
Having only the total gain pattern ) ' , ' ( ' φ θ A in the local system, the total gain in the global ( ) φ θ, -system is
simply given by the relation
2
Note that the global coordinate system should not be interpreted as a system wide global coordinate system, but a coordinate system centered
at the site antenna, with the xy-plane parallel to horizontal plane.
3GPP

θ
ˆ
φ
ˆ
θ
z
x
y
φ
3GPP TR 36.814 V9.0.0 (2010-03) 88 Release 9
) ' , ' ( ' ) , ( φ θ φ θ A A ·

with ' θ and
' φ
given by (1) and (2).

Figure A.2.1.6.2-1 Definition of angles and unit vectors when the local coordinate system has been
rotated an angle
β
around the y-axis of the global coordinate system.
For polarized fields a transformation of the field components is needed, in addition to the coordinate transformation. For
a mechanical tilt angle
β
, the global coordinate system field components ) , ( φ θ
V
F and ) , ( φ θ
H
F , (defined as
) , ( ) , ( φ θ φ θ
θ
F F
V
· and
) , ( ) , ( φ θ φ θ
φ
F F
H
·
) are calculated from the field components ) ' , ' (
'
φ θ
θ
F
and
) ' , ' (
'
φ θ
φ
F
of the radiation pattern in the local (antenna-fixed) coordinate system as
ψ φ θ ψ φ θ φ θ φ θ
φ θ θ
sin ) ' , ' ( cos ) ' , ' ( ) , ( ) , (
' '
F F F F
V
− · ·
(3)
ψ φ θ ψ φ θ φ θ φ θ
φ θ φ
cos ) ' , ' ( sin ) ' , ' ( ) , ( ) , (
' '
F F F F
H
+ · ·
(4)
where ' θ and
' φ
are defined as in (1) and (2), and
ψ
is defined as:
( ) β φ β θ φ β θ ψ sin sin sin cos cos cos sin arg j + − · . (5)
As an example, in the horizontal cut, i.e., for

90 · θ
, equations (1), (2) and (5) become
arccos( ' · θ β φsin cos
),
arg( ' · φ φ β φ sin cos cos j +
),
( ) β φ β ψ sin sin cos arg j + · .
The notation ) (φ
H
F and ) (φ
V
F , used below, should be interpreted as the field components for the specificθ that
is given by the direction of the propagation; that is, the angle between the global z-axis and the line connecting the site
antenna and the UE.
If

45 t
polarized antennas are used, the field components
' θ
F and
' φ
F
are related to the co-polarized and cross-
polarized field components in the local (antenna-fixed) coordinate system for the

45 −
polarized antenna as
3GPP
y
z
x

β
β
'
ˆ
φ
'
ˆ
θ
θ
ˆ
φ
ˆ
' x
' θ
' y
' z
' φ

θ
ˆ
φ
ˆ
ψ
'
ˆ
φ
'
ˆ
θ
3GPP TR 36.814 V9.0.0 (2010-03) 89 Release 9
) (
2
1
45 45 ' − −
+ ·
cross co
F F F
θ
, (6)
) (
2
1
45 45 ' − −
+ − ·
cross co
F F F
φ
, (7)
and similarly for the

45 +
polarized antenna.
A.2.1.7 Advanced receivers modeling
A.2.1.7.1 Iterative soft interference cancellation receivers
Advanced receivers based on iterative soft interference cancellation receivers (e.g. Turbo SIC) are non-linear receivers
whose performance improves with the reliability of the interference reconstruction as the number of iterations increases.
Modeling this interference reconstruction reliability or not in system-level simulations has a significant impact on the
accuracy of the performance evaluation, especially if only a small number of iterations is performed in order to limit the
receiver complexity.
Therefore, for system-level simulations employing a link-to-system abstraction of iterative soft interference cancellation
receivers, the interference reconstruction reliability should be modeled. In order to ease the comparison of results from
different sources, the simulation conditions should briefly describe the used modeling method.
In addition, the number of iterations should be indicated, as this parameter impacts the receiver performance as well as
its complexity.
A.2.1.8 Effective IoT
According to the ITU evaluation guidelines [4] an IoT (Interference over Thermal noise) measure reflecting the
‘effective’ interference received by the base station should be used. No definition of this measure is provided however.
This section contains a simple effective IoT measure, taking the interference suppression capabilities of the receiver into
account.
A linear receiver, e.g. an MRC or MMSE, with M antennas is assumed. The received signal at antenna m is
m m m m
n e x h r + + ·
where hm is the channel between the (single) transmitter and receive antenna m, x is the transmitted symbol, em is the
sum of the interfering signals (e.g. aimed at other users), and nm is the thermal noise at antenna m. The receiver
processing constitutes of estimating the transmitted symbol x by appropriately weighting the received signal at each
antenna with a complex-valued weight wm
H
, and then summing the weighted signals. The sum-weighted output signal is
given by
( )
∑ ∑
· ·
+ + · ·
M
m
m m m
H
m
M
m
m
H
m
n e x h w r w y
1 1
Alternatively, using vector and matrix notation, where bold variables are column vectors spanning the antenna domain:
( ) n e h w y + + · x
H
The effective SINR at the receiver output is given by the ratio of the power of the desired part of the signal, w
H
hx, to the
power of the non-desired part of the signal, w
H
(e+n) which under the assumption that noise and interference is
independent can be written as
( )w N Q w
w hh w
+
·
H
H H
eff
SINR :
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 90 Release 9
where
{ }
H
E ee Q ·
is the covariance matrix of the interference, and
{ }
H
E nn N ·
is the covariance matrix of the noise. Note that typically N


2
I, where σ
2
is the noise variance per antenna, and I is the
identity matrix of size MxM.
The denominator in the effective SINR expression now contains the effective power of the interference and thermal
noise. Hence:
( )w N Q w + ·
H
INeff
P .
The effective IoT can now be defined as the ratio between the total effective noise and interference power to the
effective thermal noise power
( )
Nw w
w N Q w
H
H
Neff
INeff
eff
+
· ·
P
P
IoT
.
Since receive weights typically vary over frequency and time, the effective IoT should be calculated per frequency and
time instant, and then averaged (i.e. do not average PINeff and PNeff separately). The average is taken over allocated
resources, non-used resources are excluded.
A.2.2 System level simulator calibration
To facilitate LTE-A evaluations simulators have been calibrated to ensure that they produce comparable results. In a
first step (1a), downlink wideband SINR (also denoted ‘geometry’) and coupling loss distributions for the ITU scenarios
and 3GPP case 1 have been evaluated and compared. In a second step (1c) downlink and uplink spectral efficiencies,
user throughput distributions, and SINR distributions for a basic LTE configuration have been evaluated and compared.
The parameters used are listed in Table A.2.2-1.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 91 Release 9
Table A.2.2-1. Parameters for calibration of system level simulators
Parameter Value
General Parameters and assumptions not explicitly stated here according to ITU
guidelines M.2135 and 3GPP specifications
Duplex method FDD
Network synchronization Synchronized
Handover margin 1dB
Downlink transmission scheme 1x2 SIMO
Downlink scheduler Round robin with full bandwidth allocation
Downlink link adaptation Wideband CQI, no PMI on PUCCH (mode 1-0)
5ms periodicity,
6ms delay total (measurement in subframe n is used in subframe n+6)
CQI measurement error: None
MCSs based on LTE transport formats [5]
Downlink HARQ Maximum four transmissions
Downlink receiver type MRC
Uplink transmission scheme 1x2 SIMO
Uplink scheduler Frequency Domain Multiplexing – non-channel dependent, share
available bandwidth between users connected to the cell, all users get
resources in every uplink subframe.
With M users and Nrb PRBs available, Mh=mod(Nrb,M) users get
floor(Nrb/M)+1 PRBs whereas Ml=M-Mh users get floor(Nrb/M) PRBs
Uplink Power control P0 = -106dBm, alpha = 1.0
Uplink Link adaptation Based on delayed measurements. Ideal channel estimate from UL
transmission in subframe n can be used for rate adaptation in subframe
n+7
MCSs based on LTE transport formats [5]
Uplink HARQ Maximum four transmissions
Proponent to specify IR or CC
Uplink receiver type MMSE in frequency domain,
MRC over antennas
(no intercell interference rejection)
Antenna configuration Vertically polarized antennas
0.5 wavelength separation at UE,
10 wavelength separation at basestation
Channel estimation Ideal, both demodulation and sounding
Control Channel overhead,
Acknowledgements etc.
LTE: L=3 symbols for DL CCHs, M=4 resource blocks for UL CCH,
overhead for demodulation reference signals,
BS antenna downtilt ITU Indoor, indoor hotspot scenario (InH): N/A
ITU Microcellular, urban micro-cell scenario (Umi): 12deg
ITU Base coverage urban, Urban macro-cell scenario (Uma): 12deg
ITU High speed, Rural macro-cell scenario (Rma): 6 deg
Case 1 3GPP 3D: 15 deg
Case 1 3GPP 2D: N/A
Feeder loss 0dB, except for the ITU scenarios in step 1a where a feeder loss of 2dB is
used.
Channel model According to ITU for ITU scenarios
SCM urban macro high spread for 3GPP case 1
Intercell interference modeling Explicit
For step 1a, a summary of the results are presented in Figure A.2.2-1 and are based on averaging independent results
from 17 different simulators. For the downlink wideband SINR, the results from different simulators are typically
within 0.5dB of the average SINR.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 92 Release 9
Calibration
0
10
20
30
40
50
60
70
80
90
100
-140 -120 -100 -80 -60 -40
Coupling gain (Prx-Ptx) [dB]
C
.
D
.
F
.

[
%
]
InH
UMi
UMa
RMa
Case13D
Case12D
Calibration
0
10
20
30
40
50
60
70
80
90
100
-10 0 10 20 30 40
Downlink wideband SINR (geometry) [dB]
C
.
D
.
F
.

[
%
]
InH
UMi
UMa
RMa
Case13D
Case12D
Figure A.2.2-1. Distributions of coupling gain and downlink wideband SINR (geometry).
Results for step 1c in terms of cell spectral efficiency, cell-edge user spectral efficiency, normalized user throughput
distributions, and post antenna combination signal-to-noise-and-interference ratio distributions (with linear averaging
over time and frequency) from 16 different simulators were available. A summary is given in Figure A.2.2-2, Figure
A.2.2-3 and Table A.2.2-2, where the results have been obtained by averaging the results of the different simulators.
Downlink Calibration
0
10
20
30
40
50
60
70
80
90
100
-10 0 10 20 30 40 50
DL Per UE Average SINR [dB]
C
.
D
.
F
.

[
%
]
InH
UMi
UMa
RMa
Case13D
Case12D
Uplink Calibration
0
10
20
30
40
50
60
70
80
90
100
-5 0 5 10 15 20
UL Per UE Average SINR [dB]
C
.
D
.
F
.

[
%
]
InH
UMi
UMa
RMa
Case13D
Case12D
Figure A.2.2-2. Distributions of downlink and uplink SINR after antenna combination.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 93 Release 9
Downlink Calibration
0
10
20
30
40
50
60
70
80
90
100
0 0,1 0,2 0,3 0,4 0,5
Normalized User Throughput [bps/Hz]
C
.
D
.
F
.

[
%
]
InH
UMi
UMa
RMa
Case 1 3D
Case 1 2D
Uplink Calibration
0
10
20
30
40
50
60
70
80
90
100
0,0 0,1 0,2 0,3 0,4 0,5
Normalized User Throughput [bps/Hz]
C
.
D
.
F
.

[
%
]
InH
UMi
UMa
RMa
Case 1 3D
Case 1 2D
Figure A.2.2-3. Distributions of downlink and uplink normalized user throughput.
Table A.2.2-2. Spectral efficiencies for calibration in the different environments
Direction Metric InH UMi UMa RMa Case 1
3D
Case 1
2D
Downlink Cell spectral efficiency 2.3 1.2 1.0 1.2 1.5 1.1
Cell-edge user spectral efficiency 0.082 0.028 0.022 0.027 0.035 0.026
Uplink Cell spectral efficiency 1.77 0.91 0.68 0.86 0.99 0.74
Cell-edge user spectral efficiency 0.084 0.033 0.026 0.034 0.036 0.031
A certain spread of the results exists. The magnitude of this is summarized in Table A.2.2-3, in terms of coefficients of
variation (standard deviation divided by average) for the spectral efficiencies.
Table A.2.2-3. Coefficients of variation for the different environments, directions and metrics.
Direction Metric InH UMi UMa RMa Case 1
3D
Case 1
2D
Downlink Cell spectral efficiency 3% 9% 8% 6% 5% 5%
Cell-edge user spectral efficiency 9% 19% 17% 14% 15% 15%
Uplink Cell spectral efficiency 5% 7% 5% 5% 5% 4%
Cell-edge user spectral efficiency 13% 16% 15% 11% 7% 9%
A.2.3 Downlink CoMP evaluation assumptions for intra-NodeB CoMP
The objectives of downlink CoMP evaluation for intra-eNodeB CoMP is to clarify the performance gain of CoMP
schemes over single-cell schemes, and clarify the need of enhanced feedback or SRS.
The intra-site CoMP with up to 3 coordinated co-located cells is the baseline scenario for performance evaluation, as
shown in Figure A.2.3-1, with both 2 transmit antennas per cell and 4 transmit antenna per cell. For full buffer traffic
model, the evaluations include 2 users per cell and 10 users per cell. Alternative scenario (without X2 interface) can be
evaluated, e.g. Intra-eNodeB CoMP with distributed Radio Remote Heads.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 94 Release 9
Figure A.2.3-1: Baseline evaluation scenario: 3 cell co-located Intra-site CoMP
Table A.2.3-1-system simulation parameters for downlink intra-site CoMP Evaluation
Parameter Values used for evaluation
Deployment scenarios Homogeneous deployments
Simulation case 3GPP-case1 SCM-UMa (high spread)
System bandwidth 10 MHz (FDD), 20MHz (TDD)
Possible transmission schemes
 SU-MIMO
 MU-MIMO
 SU-MIMO with intra-eNB CS/CB
 MU-MIMO with intra-eNB CS/CB
 SU-MIMO with intra-eNB JP-CoMP
 MU-MIMO with intra-eNB JP-CoMP
- Model the impairments of JP-CoMP
- Collision between CRS and PDSCH
- Different control regions
- Describe the way to handle the impairments
Network synchronization Synchronized
Antenna configuration
Following priority:
Config.1 eNB: Cross-polarized (0.5 l spacing)
UE: Cross-polarized antennas
Config.2 eNB: Grouped co-polarized (0.5 l within group, 10 l between group)
UE: co-polarized antennas
Config.3 eNB: co-polarized (0.5 l spacing)
UE: co-polarized antennas
Antenna pattern Follow Annex A 2.1.1.1 Table A.2.1.1-2
eNB Antenna tilt
Follow Annex A 2.1.1.1 Table A.2.1.1-2
3D as baseline
2D as additional
Channel estimation
Non-ideal.
Clarify in detail the following on CoMP evaluation:
- CSI knowledge of eNB
- Feedback scheme and/or UL sounding scheme
- Accuracy of CSI
. Quantization error
. Channel estimation error based on CRS/CSI-RS
- Channel estimation error based on DMRS
DL overhead assumption
Should be clarified for each transmission scheme
(e.g., follow ITU evaluation)
Placing of UEs Uniform distribution
Traffic model Full buffer and Non-full-buffer
A.3 Evaluation assumptions for IMT-A
The discussions during the study item phase have raised a lot of different choices. In order to have a meaningful set of
evaluation results for ITU-R submission, a common baseline of the key L1 parameters was proposed; in order that the
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 95 Release 9
results from different parties would not deviate too much so that forming any common view of the expected benefits of
LTE-A is less difficult. These parameters are not exclusive or restrictive to additional results.
Table A.3-1 Simulation assumption for Cell/cell-edge spectrum efficiency
Parameter Values used for evaluation
Deployment scenario • Indoor hotspot
• Urban micro-cell
• Urban macro-cell
• Rural macro-cell
Parameters and assumptions not shown here for each scenario are shown in
ITU guidelines [ITU-R Report M.2135].
Duplex method and bandwidths FDD: 10+10 MHz except Indoor hotspot with 20+20 MHz
TDD: 20 MHz
Baseline asymmetry during 5 subframes period:
2 full DL subframes,
Special subframe: DwPTS 11symbol, GP 1 symbol, UpPTS 2 symbol,
2 full UL subframes
Alternative special subframe configurations may be used if stated.
Network synchronization Synchronized
Handover margin 1.0 dB
Downlink transmission scheme Baseline transmission scheme (LTE Rel.8)
• MIMO closed loop precoded spatial multiplexing (transmission mode 4
[36.213]): Baseline: 4x2 MIMO
• MIMO single stream beamforming (transmission mode 7 [36.213])
Advanced scheme (LTE-A)
• MU-MIMO without coordination
• MU-MIMO with intercell coordination
• Joint processing CoMP
(SU-MIMO is possible for all cases.)
Downlink scheduler For baseline transmission scheme (LTE Rel.8):
Proportional fair in time and frequency
For advanced transmission scheme (LTE-A)
Aligned with transmission scheme
Downlink link adaptation Non-ideal based on non-ideal CQI/PMI/RI reports and/or non-ideal sounding
transmission, reporting mode: and period selected according to scheduler and
MIMO transmission schemes; reporting delay and MCS based on LTE
transport formats according to [36.213].
Baseline (LTE Rel.8):
A) Non-frequency selective PMI and frequency selective CQI report with 5ms
periodicity, subband CQI with measurement error: N(0,1) per PRB
B) Sounding-based precoding, frequency selective CQI report with 5ms
periodicity, subband CQI with measurement error: N(0,1) per PRB
Advanced transmission scheme (LTE-A)
Following scheme(s) is assumed for the base station(s) to access the channel
state information, to all UEs served by the CoMP cooperating set, which can
range from a single base station, to multiple base stations.
• The long-term, wideband, spatial channel information such as transmit
channel covariance matrices or angular information can be obtained, for
example, from uplink measurements (based on e.g., SRS) or by
feedback with an average feedback overhead similar to that in Rel-8.
• The short-term and/or narrowband channel state information (PMI,
transmit covariance matrix, channel transfer function, etc.) can be
obtained, for example, from uplink measurements (based on e.g., SRS),
and/or by feedback with the average feedback overhead to be
described in the evaluation results.
Downlink HARQ scheme Incremental redundancy or Chase combining
Downlink receiver type Baseline scheme
MMSE
Advanced scheme
MMSE-SIC, MLD based receiver
Each company should report a description on interference rejection and
cancellation capabilities.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 96 Release 9
Uplink transmission scheme Baseline transmission scheme (LTE Rel.8)
• SIMO with and without MU-MIMO
Baseline: 1 x 4 SIMO
Advanced transmission scheme (LTE-A)
• SU-MIMO
• UL CoMP
Uplink scheduler Channel dependent
Uplink Power control Baseline: Fractional power control.
Alternative: Other Rel.8 specified
Power control parameters (P0 and alpha) are chosen according to the
deployment scenario. (IoT reported with simulation results.)
Uplink link adaptation Non-ideal based on delayed SRS-based measurements: MCS based on LTE
transport formats and SRS period and bandwidths according to [36.213].
Uplink HARQ scheme Incremental redundancy or Chase combining
Uplink receiver type MMSE or MMSE-SIC (MU-MIMO)
Antenna configuration
base station
Baseline: 4 or 8 Tx antennas with the following configurations:
A) Uncorrelated co-polarized:
Co-polarized antennas separated 4 wavelengths
(illustration for 4 Tx: | | | |)
B) Grouped co-polarized:
Two groups of co-polarized antennas. 10 wavelengths between center of
each group. 0.5 wavelength separation within each group
(illustration for 4 Tx: || || )
C) Correlated: co-polarized:
0.5 wavelengths between antennas
(illustration for 4 Tx: |||| )
D) Uncorrelated cross-polarized:
Columns with +-45deg linearly polarized antennas
Columns separated 4 wavelengths
(illustration for 4 Tx: X X)
E) Correlated cross-polarized
Columns with +-45deg linearly polarized antennas
Columns separated 0.5 wavelengths
(illustration for 8Tx: XXXX)
Baseline mappings between deployment scenario and antenna configurations:
For downlink:
• Indoor Hotspot: A (Rel.8)
• Urban Micro: C or E (MU-MIMO), B or E (CoMP, Rel.8)
• Urban Macro: C or E (MU-MIMO), C or E (CoMP)
• Rural Macro: C or E (MU-MIMO, Rel.8)
Note: MU-MIMO = MU-MIMO with / without coordination
CoMP = Joint processing CoMP
For uplink:
• Indoor Hotspot: A
• Urban Micro: A or B or C or E
• Urban Macro: A or C or E
• Rural Macro: A or C or E
Antenna configuration UE Baseline:
Vertically polarized antennas with 0.5 wavelengths separation at UE
Alternative:
Columns with linearly polarized orthogonal antennas with 0.5 wavelengths
spacing between columns
Channel estimation
(Uplink and downlink)
Recommended: Non-ideal
(For non-ideal case, consider both estimation errors both for demodulation
reference signals and sounding reference signals)
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 97 Release 9
Control channel and reference
signal overhead,
Acknowledgements etc.
For baseline transmission (LTE) schemes:
• Overhead for CRS and antenna port 5 according to DL transmission
schemes
• SRS overhead according to UL (and DL) scheduler and transmission
scheme
• Overhead for UL CCH according to CQI/PMI reporting mode and
periodicity used for DL simulation for the same scenario
• Overhead for DL CCH of [1-3] OFDM symbols
For advanced transmission schemes (LTE-A)
• UL overhead for CSI feedback and SRS transmission according to
transmission scheme factored into the uplink results for the same
scenarios.
• DL overhead for additional DRS if used.
Feedback and control channel
errors
None
Table A.3-2. Mobility parameters and recommended baseline values
Parameter Baseline value
Power control Fractional (α = 0.8), P0 fitted to environment
Scheduling bandwidth 4-5 RB
Table A.3-3. VoIP parameters and recommended baseline values
Parameter Range Baseline value
Number of base station antennas 2-8 4
Scheduling Dynamic, Semi-persistent Both
PDDCH limitation Included, excluded Included
TDD configuration Configuration 0 and 1,
DwPTS 12, GP 1, UpPTS 1
A.4 Detailed simulation results
Detailed simulation results of cell spectral efficiency and cell edge spectral efficiency for full-buffer traffic in Section
10.2, VoIP capacity and mobility are contained in [6]. The following 18 corporate entities (listed below alphabetically)
participated in these simulations:
Alcatel-Lucent/Alcatel-Lucent Shanghai Bell, CATT, CMCC, Ericsson/ST-Ericsson, Fujitsu, Hitachi, Huawei, LGE,
Motorola, NEC, Nokia/Nokia Siemens Networks, NTT DOCOMO, Panasonic, Qualcomm, RITT, Samsung, Texas
Instruments, ZTE
Annex B:
Channel models (Informative)
The ITU-R IMT-Advanced channel model is a geometry-based stochastic model. It can also be called double
directional channel model. It does not explicitly specify the locations of the scatters, but rather the directions of the rays,
like the well-known spatial channel model (SCM) [B1]. Geometry-based modelling of the radio channel enables
separation of propagation parameters and antennas.
The channel parameters for individual snapshots are determined stochastically based on statistical distributions
extracted from channel measurements. Antenna geometries and radiation patterns can be defined properly by the user of
the model. Channel realizations are generated through the application of the geometrical principle by summing
contributions of rays (plane waves) with specific small-scale parameters like delay, power, angle-of-arrival (AoA) and
angle-of-departure (AoD). Superposition results to correlation between antenna elements and temporal fading with
geometry dependent Doppler spectrum.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 98 Release 9
A number of rays constitute a cluster. In the terminology of this document we equate the cluster with a propagation path
diffused in space, either or both in delay and angle domains. Elements of the MIMO channel, e.g., antenna arrays at
both link ends and propagation paths, are illustrated in Figure B-1. The generic MIMO channel model is applicable for
all scenarios, e.g. indoor, urban and rural.
Figure B-1 The MIMO channel
Path N
Array 1
(S Tx elements)
Array 2
(U Rx elements)

1 , rx
r
U rx,
r

S tx,
r
1 , tx
r
Path 1
The time variant impulse response matrix of the U x S MIMO channel is given by
( ) ( )

·
·
N
n
n
t t
1
; ; τ τ H H
, (3)
where t is time, τ is delay, N is the number of paths, and n is path index. It is composed of the antenna array response
matrices Ftx and Frx for the transmitter (Tx) and the receiver (Rx) respectively, and the dual-polarized propagation
channel response matrix hn for cluster n as follows
( ) ( ) ( ) ( )
∫∫
· ϕ φ φ ϕ φ τ ϕ τ d d t t
T
tx n rx n
, , ; ; F h F H
. (4)
The channel from Tx antenna element s to Rx element u for cluster n is expressed as

( )
( )
( )
( )
( )
( ) ( ) ( ) ( )
( ) ( )
m n m n
s tx m n u rx m n
m n H s tx
m n V s tx
HH m n HV m n
VH m n VV m n
T
m n H u rx
m n V u rx
M
m
n s u
t j
r j r j
F
F
a F
F
t H
, ,
, ,
1
0 , ,
1
0
, , ,
, , ,
, , , ,
, , , ,
, , ,
, , ,
1
, ,
2 exp
2 exp 2 exp
;
τ τ δ πυ
φ πλ ϕ πλ
φ
φ
α
α α
ϕ
ϕ
τ
− ×
⋅ ⋅ ×
1
]
1

¸

1
]
1

¸

1
]
1

¸

·
− −
·

(5)
where Frx,u,V and Frx,u,H are the antenna element u field patterns for vertical and horizontal polarizations respectively,
α n,m,VV and α n,m,VH are the complex gains of vertical-to-vertical and horizontal-to-vertical polarizations of ray n,m
respectively, λ 0 is the wave length of the carrier frequency,
m n.
φ
is the AoD unit vector,
m n.
ϕ
is the AoA unit
vector,
s tx
r
,
and
u rx
r
,
are the location vectors of element s and u respectively, and ν n,m is the Doppler frequency
component of ray n,m. If the radio channel is modelled as dynamic, all the above mentioned small-scale parameters are
time variant, i.e., they are functions of t [B8].
The IMT-Advanced channel model for the evaluation of IMT-Advanced candidate RITs consists of a Primary Module
and an Extension Module. In this part, only the primary module is included, which is mandatory for IMT-Advanced
evaluation. The Primary Module covers the mathematical framework, which is called generic model, a set of parameters
as well as path loss models.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 99 Release 9
The generic channel model is a double-directional geometry-based stochastic model. It is a system level
3
model in the
sense that is employed, e.g., in the SCM model [B1]. It can describe an unlimited number of propagation environment
realizations for single or multiple radio links for all the defined scenarios and for arbitrary antenna configurations, with
one mathematical framework by different parameter sets. The generic channel model is a stochastic model with two (or
three) levels of randomness. First, large-scale (LS) parameters like shadow fading, delay, and angular spreads are drawn
randomly from tabulated distribution functions. Next, small-scale (SS) parameters like delays, powers, and directions of
arrival and departure are drawn randomly according to tabulated distribution functions and random LS parameters. At
this stage the geometric setup is fixed and the only free variables are the random initial phases of the scatters. By
picking (randomly) different initial phases, an infinite number of different realizations of the model can be generated.
When the initial phases are also fixed, there is no further randomness left.
Figure B-2 shows the overview of the channel model creation. The first stage consists of two steps. First, the
propagation scenario is selected. Then, the network layout and the antenna configuration are determined. In the second
stage, large-scale and small-scale parameters are defined. In the third stage, channel impulse responses (ChIRs) are
calculated.
The generic model is based on the drop concept. When using the generic model, the simulation of the system behaviour
is carried out as a sequence of “drops”, where a “drop” is defined as one simulation run over a certain time period. A
drop (or snapshot or channel segment) is a simulation entity where the random properties of the channel remain
constant except for the fast fading caused by the changing phases of the rays. The constant properties during a single
drop are, e.g., the powers, delays, and directions of the rays. In a simulation the number and the length of drops have to
be selected properly by the evaluation requirements and the deployed scenario. The generic model allows the user to
simulate over several drops to get statistically representative results. Consecutive drops are independent.
Figure B-2 Channel model creation process
ChIR
generation
Propagation parameter
generation
User defined parameters
Scenario
selection
-urban macro
-urban micro
-indoor
-out2in
-etc.
Network
Layout
-BS & MS
locations
-velocities
Antennas
-# elements
-orientations
-field patterns
Large scale
parameters
-DS, AS, K
-XPR
-shadowing
-path loss
Multi-path
parameters
-power,
delay, AoA,
AoD, etc.
Channel
coefficient
generation
ChIR
B.1 Channel models
This section provides the reference channel model for each test environment. These test environments are intended to
cover the range of IMT-Advanced operating environments.
The test operating environments are considered as a basic factor in the evaluation process of the candidate RITs. The
reference models are used to estimate the critical aspects, such as the spectrum, coverage and power efficiencies.
B.1.1 Void
B.1.2 Primary module
The following sections provide the details of the channel models, including the path loss models, for the terrestrial
component. For terrestrial environments, the propagation effects are divided into three distinct types: These are the path
loss, the slow variation due to shadowing and scattering, and the rapid variation in the signal due to multipath effects.
3
The term system-level means here that the model is able to cover multiple links, cells and terminals.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 100 Release 9
The channel models are specified in the frequency range from 2 GHz to 6 GHz. For the rural macro-cell scenario
(RMa), the channel model can be used for lower frequencies down to 450 MHz. The channel models also cover MIMO
aspects as all desired dimensions (delay, AoA, AoD and polarisation) are considered. The channel models are targeted
for up to 100 MHz RF bandwidth.
B.1.2.1 Path loss models
Path loss models for the various propagation scenarios have been developed based on measurement results carried out
in [B2, B7, B9-B14], as well as results from the literature. The models can be applied in the frequency range of 2 – 6
GHz and for different antenna heights. The rural path-loss formula can be applied to the desired frequency range from
450 MHz to 6 GHz. The path loss models have been summarized in Table B.1.2.1-1. Note that the distribution of the
shadow fading is log-normal, and its standard deviation for each scenario is given in the table.
Table B.1.2.1-1 Summary table of the primary module path loss models
Scenario Path loss [dB]
Note: fc is given in GHz and distance in meters!
Shadow
fading
std [dB]
Applicability range, antenna
height default values
I
n
d
o
o
r

H
o
t
s
p
o
t

(
I
n
H
)

LOS
PL = 16.9log10(d) + 32.8 + 20log10(fc) σ = 3 3 m < d < 100 m
hBS =3-6 m
hUT =1-2.5 m
NLOS
PL = 43.3log10(d) + 11.5 + 20log10(fc)
σ = 4
10 m < d < 150 m
hBS = 3-6 m
hUT = 1-2.5 m
U
r
b
a
n

M
i
c
r
o

(
U
M
i
)
LOS
PL = 22.0log10(d) + 28.0 + 20log10(fc)
PL = 40log10(d1) + 7.8 – 18log10(h’BS) –18log10(h’UT) +
2log10(fc)
σ = 3
σ = 3
10 m < d1 < d’BP
1)
d’BP < d1 < 5000 m
1)
hBS = 10 m
1)
, hUT = 1.5 m
1)
NLOS Manhattan grid layout:
( ) ) , ( ), , ( min
1 2 2 1
d d PL d d PL PL ·
where
( )
) ( log 3
) ( log 10 5 . 12 9 . 17 ) (
,
10
10
c
l j j k LOS
l k
f
d n n d PL
d d PL
+ + − +
·
and
( ) 84 . 1 , 0024 . 0 8 . 2 max
k j
d n − ·
and
PLLOS is the path loss of scenario UMi LOS and k,l ∈
{1,2}.
Hexagonal cell layout:
PL = 36.7log10(d) + 22.7 + 26log10(fc)
σ = 4
σ = 4
10 m < d
1
+ d
2
< 5 000 m,
w/2 < min(d
1
,d
2
)
2)
w = 20 m (street width)
hBS = 10 m, hUT = 1.5 m
When 0 < min(d
1
,d
2
)

< w/2 , the
LOS PL is applied.
10 m < d < 2 000 m
hBS = 10 m
hUT =1-2.5 m
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 101 Release 9
O-to-I
in tw b
PL PL PL PL + + ·
Manhattan grid layout (θ known):
¹
¹
¹
'
¹
·
− + ·
+ ·
i n i n
t w
i n o u t B b
d P L
P L
d d P L P L
5 . 0
) ) c o s ( 1 ( 1 5 1 4
) (
2
1
θ
For hexagonal layout (θ unknown):
PLtw = 20, other values remain the same.
σ = 7
10 m < dout+din< 1 000 m,
0 m < din< 25 m,
hBS=10m, hUT=3(nFl -1)+1.5m,
nFl=1,
Explanations: see
3)
U
r
b
a
n

M
a
c
r
o

(
U
M
a
)
LOS
PL = 22.0log10(d) + 28.0 + 20log10(fc)
( )
c UT
BS
f h
h d PL
10 10
10 1 10
log 0 . 2 ) ' ( log 0 . 18
) ' ( log 0 . 18 8 . 7 ) ( log 0 . 40
+ −
− + ·
σ = 4
σ = 4
10 m < d < d’BP
1)
d’BP < d < 5 000 m
1)
hBS = 25 m
1)
, hUT = 1.5 m
1)
NLOS
PL = 161.04 – 7.1 log10 (W) + 7.5 log10 (h)
– (24.37 – 3.7(h/hBS)
2
) log10 (hBS)
+ (43.42 – 3.1 log10 (hBS)) (log10 (d)-3) +
20 log10(fc) – (3.2 (log10 (11.75 hUT))
2
- 4.97)
σ = 6
10 m < d < 5 000 m
h = avg. building height
W = street width
hBS = 25 m, hUT = 1.5 m,
W = 20 m, h = 20 m
The applicability ranges:
5 m < h < 50 m
5 m < W < 50 m
10 m < h
BS
< 150 m
1 m < h
UT
< 10 m
S
u
b
u
r
b
a
n

M
a
c
r
o

(
S
M
a
,

o
p
t
i
o
n
a
l
)
LOS
PL1 = 20log10(40π dfc /3)
+ min(0.03h
1.72
,10)log10(d)
– min(0.044h
1.72
,14.77)
+ 0.002log10(h)d
PL2 = PL1 (dBP) + 40 log10(d/dBP)
σ = 4
σ = 6
10 m < d < dBP
4)
dBP < d < 5 000 m
hBS = 35 m, hUT = 1.5 m,
W = 20 m, h = 10 m
(The applicability ranges of h, W,
hBS, hUT are same as in UMa
NLOS)
NLOS
PL = 161.04 – 7.1 log10 (W) + 7.5 log10 (h)
– (24.37 – 3.7(h/hBS)
2
) log10 (hBS)
+ (43.42 – 3.1 log10 (hBS)) (log10 (d)-3) +
20 log10(fc) – (3.2 (log10 (11.75 hUT))
2
- 4.97)
σ = 8 10 m < d < 5 000 m
hBS = 35 m, hUT = 1.5 m,
W = 20 m, h = 10 m
(Applicability ranges of h, W,
hBS, hUT are same as in UMa
NLOS)
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 102 Release 9
R
u
r
a
l

M
a
c
r
o

(
R
M
a
)
LOS
PL1 = 20log10(40π dfc /3)
+ min(0.03h
1.72
,10)log10(d)
– min(0.044h
1.72
,14.77)
+ 0.002log10(h)d
PL2 = PL1 (dBP) + 40 log10(d/dBP)
σ = 4
σ = 6
10 m < d < dBP,
4)
dBP < d < 10 000 m,
hBS = 35 m, hUT = 1.5 m,
W = 20 m, h = 5 m
(Applicability ranges of h, W,
hBS, hUT are same as UMa NLOS)
NLOS
PL = 161.04 – 7.1 log10 (W) + 7.5 log10 (h)
– (24.37 – 3.7(h/hBS)
2
) log10 (hBS)
+ (43.42 – 3.1 log10 (hBS)) (log10 (d)-3) +
20 log10(fc) – (3.2 (log10 (11.75 hUT))
2
- 4.97)
σ = 8
10 m < d < 5 000 m,
hBS = 35 m, hUT = 1.5 m,
W = 20 m, h = 5 m
(The applicability ranges of h, W,
hBS, hUT are same as UMa NLOS)
1) Break point distance d’BP

= 4 h’BS h’UT fc/c, where fc is the centre frequency in Hz, c = 3.0× 10
8
m/s is the propagation velocity in
free space, and h’BS and h’UT are the effective antenna heights at the BS and the UT, respectively. The effective antenna heights
h’BS and h’UT are computed as follows: h’BS = hBS – 1.0 m, h’UT = hUT – 1.0 m, where hBS and hUT are the actual antenna heights,
and the effective environment height in urban environments is assumed to be equal to 1.0 m.
2) The distances d1 and d2 are defined below in Figure B.1.2.1-1.
3) PLb = basic path-loss, PL B1 = Loss of UMi outdoor scenario, PLtw = Loss through wall, PLin = Loss inside, dout = distance from
BS to the wall next to UT location, din = perpendicular distance from wall to UT (assumed evenly distributed between 0 and 25
m), θ = angle between LOS to the wall and a unit vector normal to the wall.
4) Break point distance dBP

= 2π hBS hUT fc/c, where fc is the centre frequency in Hz, c = 3.0× 10
8
m/s is the propagation velocity in
free space, and hBS and hUT are the antenna heights at the BS and the UT, respectively.
The line-of-sight (LOS) probabilities are given in Table B.1.2.1-2. Note that probabilities are used only for system level
simulations.
Table B.1.2.1-2
Scenario LOS probability as a function of distance d [m]
InH ( ) ( )
¹
¹
¹
'
¹

< < − −

·
37 , 5 . 0
37 18 , 27 18 exp
18 , 1
d
d d
d
P
LOS
UMi
( ) ) 36 / exp( ) 36 / exp( 1 ) 1 , / 18 min( d d d P
LOS
− + − − ⋅ ·
(for outdoor users only)
UMa ( ) ) 63 / exp( ) 63 / exp( 1 ) 1 , / 18 min( d d d P
LOS
− + − − ⋅ ·
RMa
10 ,
1000
10
exp
10 , 1
>

,
_

¸
¸ −


·
d
d
d
P
LOS
The NLOS path loss model for scenario UMi is dependent on two distances, d1 and d2 in the case of the Manhattan grid.
These distances are defined with respect to a rectangular street grid, as illustrated in Figure B.1.2.1-1, where the UT is
shown moving along a street perpendicular to the street on which the BS is located (the LOS street). d1 is the distance
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 103 Release 9
from the BS to the centre of the perpendicular street, and d2 is the distance of the UT along the perpendicular street,
measured from the centre of the LOS street.
Figure B.1.2.1-1 Geometry for d1 - d2 path-loss model
BS
d 1
d
d
2
2
MS
+
-
B.1.2.1.1 Autocorrelation of shadow fading
The long-term (log-normal) fading in the logarithmic scale around the mean path loss PL (dB) is characterized by a
Gaussian distribution with zero mean and standard deviation. Due to the slow fading process versus distance ∆ x,
adjacent fading values are correlated. Its normalized autocorrelation function R(∆ x) can be described with sufficient
accuracy by the exponential function [B6]
( )
cor
d
x
x R


· ∆ e
(6)
with the correlation length d
cor
being dependent on the environment, see the correlation parameters for shadowing and
other large scale parameters in Table B.1.2.2.1-4 (Channel model parameters).
B.1.2.2 Primary module channel model parameters
B.1.2.2.1 Generic model
The radio channels are created using the parameters listed in Table B.1.2.2.1-4. The channel realizations are obtained by
a step-wise procedure [B2] illustrated in Figure B.1.2.2.1-1 and described below. It has to be noted that the geometric
description covers arrival angles from the last bounce scatterers and respectively departure angles to the first scatterers
interacted from the transmitting side. The propagation between the first and the last interaction is not defined. Thus, this
approach can model also multiple interactions with the scattering media. This indicates also that e.g., the delay of a
multipath component cannot be determined by the geometry. In the following steps, downlink is assumed. For uplink,
arrival and departure parameters have to be swapped.
Figure B.1.2.2.1-1 Channel coefficient generation procedure
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 104 Release 9
Coefficient generation:
Small scale parameter:
General parameters :
Set scenario, network
layout and antenna
parameters
Assign propagation
condition (NLOS/
LOS)
Calculate path loss
Generate correlated
large scale paramters
(DS, AS, SF, K)
Generate delays
Generate cluster
powers
Generate arrival &
departure angles
Perform random
coupling of rays
Draw random initial
phases
Generate channel
coefficient
Apply path loss &
shadowing
General parameters:
Step 1: Set environment, network layout, and antenna array parameters
a. Choose one of the scenarios (InH, UMi, …)
b. Give number of BS and UT
c. Give locations of BS and UT, or equally distances of each BS and UT and relative directions and
φ LOS and ϕ LOS of each BS and UT
d. Give BS and UT antenna field patterns Frx and Ftx and array geometries
e. Give BS and UT array orientations with respect to north (reference) direction
f. Give speed and direction of motion of UT
g. Give system centre frequency
Large scale parameters:
Step 2: Assign propagation condition (LOS/NLOS).
Step 3: Calculate path loss with formulas of Table B.1.2.1-1for each BS-UT link to be modelled.
Step 4: Generate correlated large scale parameters, i.e. delay spread, angular spreads, Ricean K factor and shadow
fading term like explained in [B2, section 3.3.1] (Correlations between large scale parameters). Limit random rms
arrival and departure azimuth spread values to 104 degrees, i.e., σ
ϕ = min(σ
ϕ ,104°).
Small scale parameters:
Step 5: Generate delays τ.
Delays are drawn randomly from the delay distribution defined in Table B.1.2.2.1-4. With exponential delay
distribution calculate
( )
n n
X r ln '
τ τ
σ τ − ·
, (7)
where r
τ
is the delay distribution proportionality factor, Xn ~ Uni(0,1), and cluster index n = 1,…,N. With uniform
delay distribution the delay values τ n’ are drawn from the corresponding range. Normalise the delays by
subtracting the minimum delay and sort the normalised delays to descending order:
( ) ( ) ' min ' sort
n n n
τ τ τ − ·
. (8)
In the case of LOS condition, additional scaling of delays is required to compensate for the effect of LOS peak
addition to the delay spread. The heuristically determined Ricean K-factor dependent scaling constant is
3 2
000017 . 0 0002 . 0 0433 . 0 7705 . 0 K K K D + + − · , (9)
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 105 Release 9
where K [dB] is the Ricean K-factor defined in Table B.1.2.2.1-4.. The scaled delays
D
n
LOS
n
/ τ τ ·
, (10)
are not to be used in cluster power generation.
Step 6: Generate cluster powers P.
Cluster powers are calculated assuming a single slope exponential power delay profile. Power assignment depends
on the delay distribution defined in Table B.1.2.2.1-4. With exponential delay distribution the cluster powers are
determined by
10
'
10
1
exp
n
r
r
P
n n
Ζ −

,
_

¸
¸ −
− ·
τ τ
τ
σ
τ
(11)
where Ζ n ~ N(0, ζ ) is the per cluster shadowing term in [dB]. Average the power so that the sum power of all
cluster powers is equal to one, i.e.,

·
·
N
n
n
n
n
P
P
P
1
'
'
(12)
Assign the power of each ray within a cluster as Pn / M, where M is the number of rays per cluster.
Remove clusters with less than -25 dB power compared to the maximum cluster power.
Step 7: Generate arrival angles ϕ and departure angles φ .
As the composite PAS of all clusters is modelled as wrapped Gaussian (see Table B.1.2.2.1-4), except indoor
hotspot scenario (InH) as Laplacian, the AoAs are determined by applying the inverse Gaussian function (13) or
inverse Laplacian function (14) with input parameters Pn and RMS angle spread σ
ϕ
( ) ( )
C
P P
n n
n
max ln 2
'
AoA

·
σ
ϕ
. (13)
( ) ( )
C
P P
n n
n
max ln
'
ϕ
σ
ϕ − · (14)
In the equation (13)
4 . 1
AoA ϕ
σ σ ·
is the standard deviation of the arrival angles (the factor 1.4 is the ratio of
Gaussian std and the corresponding “RMS spread”). Constant C is a scaling factor related to total number of clusters
and is given in the table below:
Table B.1.2.2.1-1
#
cluster
s
4 5 8 10 11 12 14 15 15
(InH)
16 19 19
(InH)
20
C 0.77
9
0.86
0
1.01
8
1.09
0
1.12
3
1.14
6
1.19
0
1.21
1
1.434 1.22
6
1.27
3
1.501 1.289
In the LOS case, constant C is dependent also on the Ricean K-factor. Constant C in (13) and (14) is substituted by
C
LOS
. Additional scaling of the angles is required to compensate for the effect of LOS peak addition to the angle
spread. The heuristically determined Ricean K-factor dependent scaling constant is
( )
3 2
0001 . 0 002 . 0 028 . 0 1035 . 1 K K K C C
LOS
+ − − ⋅ · , (15)
As for indoor hotspot scenario, the scaling constant is:
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 106 Release 9
( )
3 2
0002 . 0 0071 . 0 0439 . 0 9275 . 0 K K K C C
LOS
+ − + ⋅ · (16)
where K [dB] is the Ricean K-factor defined in Table B.1.2.2.1-4.
Assign positive or negative sign to the angles by multiplying with a random variable Xn with uniform distribution to
the discrete set of {1,–1}, and add component
( ) 7 , 0 N ~
ϕ
σ
n
Y
to introduce random variation
LOS n n n n
Y X ϕ ϕ ϕ + + · '
, (17)
where ϕ LOS is the LOS direction defined in the network layout description, see Step1c.
In the LOS case, substitute (17) by (18) to enforce the first cluster to the LOS direction ϕ LOS
( ) ( )
LOS n n n n
Y X Y X ϕ ϕ ϕ ϕ − + − + ·
1 1 1
' '
. (18)
Finally add offset angles α m from Table B.1.2.2.1-2 to the cluster angles
m AoA n m n
c α ϕ ϕ + ·
, , (19)
where cAoA is the cluster-wise rms azimuth spread of arrival angles (cluster ASA) in Table B.1.2.2.1-4.
Table B.1.2.2.1-2 Ray offset angles within a cluster, given for 1 rms angle spread
Ray number
m
Basis vector of offset angles
α m
1,2 ± 0.0447
3,4 ± 0.1413
5,6 ± 0.2492
7,8 ± 0.3715
9,10 ± 0.5129
11,12 ± 0.6797
13,14 ± 0.8844
15,16 ± 1.1481
17,18 ± 1.5195
19,20 ± 2.1551
For departure angles φ n the procedure is similar.
Step 8: Random coupling of rays within clusters.
Couple randomly departure ray angles φ n,m to arrival ray angles ϕ n,m within a cluster n, or within a sub-cluster in
the case of two strongest clusters (see Step 10a and Table B.1.2.2.1-2).
Coefficient generation:
Step 9: Draw random initial phase { }
hh
,
hv
,
vh
,
vv
,
, , ,
m n m n m n m n
Φ Φ Φ Φ for each ray m of each cluster n and for four
different polarisation combinations (vv,vh,hv,hh). The distribution for initial phases is uniform within (-π, π ).
In the LOS case, draw also random initial phases { }
hh vv
,
LOS LOS
Φ Φ for both VV and HH polarisations.
Step 10a: Generate channel coefficients for each cluster n and each receiver and transmitter element pair u,s.
For the N – 2 weakest clusters, say n = 3,4,…,N, and uniform linear arrays (ULA), the channel coefficients are
given by:
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 107 Release 9
( )
( )
( )
( ) ( )
( ) ( )
( )
( )
( ) ( ) ( ) ( ) ( ) t j jd jd
F
F
j j
j j
F
F
P t
m n m n u m n s
m n H s tx
m n V s tx
hh
m n
hv
m n
vh
m n
vv
m n
T
m n H u rx
m n V u rx
M
m
n n s u
, ,
1
0 ,
1
0
, , ,
, , ,
, ,
1
,
1
,
, , ,
, , ,
1
, ,
2 exp sin 2 exp sin 2 exp
exp exp
exp exp
π υ ϕ π λ φ π λ
φ
φ
κ
κ
ϕ
ϕ
− −


·

1
]
1

¸

1
1
]
1

¸

Φ Φ
Φ Φ
1
]
1

¸

·


H
(20)
where Frx,u,V and Frx,u,H are the antenna element u field patterns for vertical and horizontal polarisations respectively,
ds and du are the uniform distances [m] between transmitter elements and receiver elements respectively, κ is the
cross polarisation power ratio in linear scale, and λ 0 is the wavelength of the carrier frequency. If polarisation is not
considered, the 2x2 polarisation matrix can be replaced by the scalar
( )
m n
j
,
exp Φ
and only vertically polarised
field patterns are applied.
The Doppler frequency component is calculated from the angle of arrival (downlink), UT speed v and direction of
travel θ v
( )
0
,
,
cos
λ
θ ϕ
υ
v m n
m n
v −
·
, (21)
For the two strongest clusters, say n = 1 and 2, rays are spread in delay to three sub-clusters (per cluster), with fixed
delay offset {0,5,10 ns} (see Table B.1.2.2.1-3). The delays of the sub-clusters are
ns 10
ns 5
ns 0
3 ,
2 ,
1 ,
+ ·
+ ·
+ ·
n n
n n
n n
τ τ
τ τ
τ τ
(22)
Twenty rays of a cluster are mapped to sub-clusters as presented in Table B.1.2.2.1-3 below. The corresponding
offset angles are taken from Table B.1.2.2.1-2 with mapping of Table B.1.2.2.1-3.
Table B.1.2.2.1-3 Sub-cluster information for intra cluster delay spread clusters
sub-cluster # mapping to rays power delay offset
1 1,2,3,4,5,6,7,8,19,2
0
10/20 0 ns
2 9,10,11,12,17,18 6/20 5 ns
3 13,14,15,16 4/20 10 ns
In the LOS case, define
n s u n s u , , , ,
' H H ·
and determine the channel coefficients by adding a single line-of-sight
ray and scaling down the other channel coefficient generated by (20). The channel coefficients are given by:
( ) ( )
( )
( )
( )
( )
( )
( )
( )
( ) ( ) ( ) ( ) ( ) t j jd jd
F
F
j
j
F
F
K
K
n
t
K
t
LOS LOS u LOS s
LOS H s tx
LOS V s tx
hh
LOS
vv
LOS
T
LOS H u rx
LOS V u rx
R
R
n s u
R
n s u
π υ ϕ π λ φ π λ
φ
φ
ϕ
ϕ
δ
2 exp sin 2 exp sin 2 exp
exp 0
0 exp
1
1
'
1
1
1
0
1
0
, ,
, ,
, ,
, ,
, , , ,
− −

1
]
1

¸

1
]
1

¸

Φ
Φ
1
]
1

¸

+
− +
+
·

H H
(23)
where δ (
.
) is the Dirac’s delta function and KR is the Ricean K-factor defined in Table B.1.2.2.1-4 converted to
linear scale.
If non-ULA arrays are used, the equations must be modified. For arbitrary array configurations on the horizontal
plane, see Figure B.1.2.2.1-2, the distance term du in equations (23) and (20) is replaced by:
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 108 Release 9
( ) ( )
m n
m n u u u u
m n u
x y y x
d
,
,
2 2
'
, ,
sin
arctan cos
ϕ
ϕ − +
· , (24)
where (xu,yu) are the co-ordinates of the uth element Au and A0 is the reference element.

ϕ
n,m

d
u

y
u

x
u

A
0

A
u
y
x
'
, , m n u
d
Figure B.1.2.2.1-2 Modified distance of antenna element u with non-ULA array
Step 10b: Generate channel coefficients for each cluster n and each receiver and transmitter element pair u,s.
Alternatively to Step 10a channel coefficients can be generated by applying the so called correlation matrix based
method. Temporal correlations (Doppler effect) are introduced either by filtering independent and identically distributed
complex Gaussian sequences with a proper Doppler spectrum shaping filter or by applying some equivalent method.
MIMO antenna correlations are introduced by performing a linear transformation to the temporally correlated
sequences. Unique covariance matrices are determined for each drop and radio link based on the parameters defined in
the previous steps. The MIMO covariance matrix can be composed of transmitter and receiver spatial correlation
matrices, and by the polarization covariance matrices with the Kronecker product of matrices. Spatial correlation
matrices can be derived for each cluster directly from the ChIRs of Step 10a [B3, B4]. Alternatively, spatial correlation
matrices and the polarization covariance matrices can be derived from the antenna configuration and the model
parameters (AoA, AoD) determined in Step7 with XPR parameter from Table B.1.2.2.1-4 [B5].
Step 11: Apply path loss and shadowing for the channel coefficients. This is valid only for system level simulations.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 109 Release 9
Table B.1.2.2.1-4 Channel model parameters
• In the table below DS = rms delay spread, ASD = rms azimuth spread of departure angles, ASA = rms azimuth spread of arrival angles, SF
= shadow fading, and K = Ricean K-factor.
• The sign of the shadow fading is defined so that positive SF means more received power at UT than predicted by the path loss model.
Scenarios
InH UMi UMa RMa
LOS NLOS LOS NLOS O-to-I LOS NLOS LOS NLOS
Delay spread (DS)
log10([s])
µ -7.70 -7.41 -7.19 -6.89 -6.62 -7.03 -6.44 -7.49 -7.43
σ 0.18 0.14 0.40 0.54 0.32 0.66 0.39 0.55 0.48
AoD spread (ASD)
log10([°])
µ 1.60 1.62 1.20 1.41 1.25 1.15 1.41 0.90 0.95
σ 0.18 0.25 0.43 0.17 0.42 0.28 0.28 0.38 0.45
AoA spread (ASA)
log10([°])
µ 1.62 1.77 1.75 1.84 1.76 1.81 1.87 1.52 1.52
σ 0.22 0.16 0.19 0.15 0.16 0.20 0.11 0.24 0.13
Shadow fading (SF)
[dB]
σ 3 4 3 4 7 4 6 4 8
K-factor (K) [dB]
µ 7 N/A 9 N/A N/A 9 N/A 7 N/A
σ 4 N/A 5 N/A N/A 3.5 N/A 4 N/A
Cross-Correlations *
ASD vs DS 0.6 0.4 0.5 0 0.4 0.4 0.4 0 -0.4
ASA vs DS 0.8 0 0.8 0.4 0.4 0.8 0.6 0 0
ASA vs SF -0.5 -0.4 -0.4 -0.4 0 -0.5 0 0 0
ASD vs SF -0.4 0 -0.5 0 0.2 -0.5 -0.6 0 0.6
DS vs SF -0.8 -0.5 -0.4 -0.7 -0.5 -0.4 -0.4 -0.5 -0.5
ASD vs ASA 0.4 0 0.4 0 0 0 0.4 0 0
ASD vs Κ 0 N/A -0.2 N/A N/A 0 N/A 0 N/A
ASA vs Κ 0 N/A -0.3 N/A N/A -0.2 N/A 0 N/A
DS vs Κ -0.5 N/A -0.7 N/A N/A -0.4 N/A 0 N/A
SF vs Κ 0.5 N/A 0.5 N/A N/A 0 N/A 0 N/A
Delay distribution Exp Exp Exp Exp Exp Exp Exp Exp Exp
AoD and AoA distribution Laplacian Wrapped Gaussian
Wrapped
Gaussian
Wrapped
Gaussian
Delay scaling parameter r
τ
3.6 3 3.2 3 2.2 2.5 2.3 3.8 1.7
XPR [dB] µ 11 10 9 8.0 9 8 7 12 7
Number of clusters 15 19 12 19 12 12 20 11 10
Number of rays per cluster 20 20 20 20 20 20 20 20 20
Cluster ASD 5 5 3 10 5 5 2 2 2
Cluster ASA 8 11 17 22 8 11 15 3 3
Per cluster shadowing std ζ [dB] 6 3 3 3 4 3 3 3 3
Correlation distance [m]
DS 8 5 7 10 10 30 40 50 36
ASD 7 3 8 10 11 18 50 25 30
ASA 5 3 8 9 17 15 50 35 40
SF 10 6 10 13 7 37 50 37 120
Κ 4 N/A 15 N/A N/A 12 N/A 40 N/A
Table B.1.2.2.1-5 Expectation (median) output values for large scale parameters
Scenario DS (ns) AS at BS (º) AS at UT (º)
InH
LOS 20 40 42
NLOS 39 42 59
UMi
LOS 65 16 56
NLOS 129 26 69
O-to-I 49 18 58
UMa
LOS 93 14 65
NLOS 365 26 74
RMa
LOS 32 8 33
NLOS 37 9 33
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 110 Release 9
B.2 References
[B1] 3GPP TR25.996 V6.1.0 (2003-09) “Spatial channel model for multiple input multiple output (MIMO)
simulations” Release 6.
[B2] IST-WINNER II Deliverable 1.1.2 v.1.2, “WINNER II Channel Models”, IST-WINNER2, Tech. Rep., 2007
(http://www.ist-winner.org/deliverables.html).
[B3] J. P. Kermoal, L. Schumacher, K. I. Pedersen, P. E. Mogensen, and F. Frederiksen, "A stochastic MIMO
radio channel model with experimental validation," IEEE Journal on Selected Areas in Communications, vol.
20, no. 6, pp. 1211-1226, 2002.
[B4] P. J. Smith and M. Shafi, "The impact of complexity in MIMO channel models," in Proceedings of IEEE
International Conference on Communications (ICC '04), vol. 5, pp. 2924-2928, Paris, France, June 2004.
[B5] WiMAX Forum, WiMAX Forum Mobile Release 1.0 Channel Model, June 2008, available at
http://www.wimaxforum.org/technology/documents/wimax_forum_mobile_release_1_0_channel_model_v100
.pdf.
[B6] Rec. ITU-R P.1816, ” The prediction of the time and the spatial profile for broadband land mobile services
using UHF and SHF bands”.
[B7] T. Fujii, “Path-loss prediction formula in mobile communication -an expansion of “SAKAGAMI” path-loss
prediction formula-,” Trans. IEICE, Japan, J86-B, 10, pp. 2264-2267, 2003.
[B8] M. Steinbauer, A. F. Molisch, and E. Bonek, “The double-directional radio channel,” IEEE Antennas and
Propagation Mag., pp. 51-63, August 2001.
[B9] “Proposed Propagation Models for Evaluation Radio Transmission Technologies in IMT-Advanced,” Doc.
5D/88, January 2008, Geneva, Switzerland.
[B10] Gudmundson, M. Correlation Model for Shadow Fading in Mobile Radio Systems. Electron. Lett, Vol. 27,
23, 2145-2146). November, 1991.
[B11] H. Omote and T. Fujii, “Empirical arrival angle profile prediction formula for mobile communication
systems” Proceeding of IEEE 2007 VTC Spring, Dublin, 2007.
[B12] H. Okamoto and S, Ichitsubo “Investigations of outdoor-to-indoor propagation loss on 800MHz-8GHz at
urban area,” Proc. The 2005 IEICE General Conference, Japan, pp. S23-S24, March 2005 (in Japanese).
[B13] Jianhua Zhang, Di Dong, Yanping Liang, Xin Nie, Xinying Gao,Yu Zhang, Chen Huang and Guangyi Liu,
“Propagation characteristics of wideband MIMO channel in urban micro- and macrocell,” IEEE PIMRC 2008,
Invited paper.
[B14] Dong, Weihui; Zhang, Jianhua; Gao, Xinyin; Zhang, Ping; Wu, Yufei;, “Cluster Identification and Properties
of Outdoor Wideband MIMO Channel”, Vehicular Technology Conference, 2007. VTC -2007 Fall. 2007 IEEE
66th, Sept. 30 2007-Oct. 3 2007 Page(s): 829 – 833.
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 111 Release 9
Annex C:
Change history
Table C.1: Draft History
Change history
Date TSG # TSG Doc. Subject/Comment Old New
R1#54bis Draft skeleton TR
TP for clause 5,6,7,8 and 9 are captured
- 0.1.0
R1#55 TP for simulation models are captured - 0.2.0
R1#55bis R1-090510 TP for winder bandwidth are captured - 0.3.0
R1#56 R1-091009 TP for UL scheme and channel models are captured - 0.4.0
R1#56 R1-091125 TP for UL/DL MIMO, DL RS, Relay are captured - 1.0.0
18/03/10 RAN_47 RP-100357 TR36.814 ver.2.0.1 for approval by plenary 2.0.0 2.01
18/03/10 RAN_47 - TR is put under change control as per RAN_47 decision 2.0.1 9.0.0
3GPP
3GPP TR 36.814 V9.0.0 (2010-03) 112 Release 9

Release 9

2

3GPP TR 36.814 V9.0.0 (2010-03)

Keywords
UMTS, Radio

3GPP Postal address 3GPP support office address
650 Route des Lucioles - Sophia Antipolis Valbonne - FRANCE Tel.: +33 4 92 94 42 00 Fax: +33 4 93 65 47 16

Internet
http://www.3gpp.org

Copyright Notification No part may be reproduced except as authorized by written permission. The copyright and the foregoing restriction extend to reproduction in all media.
© 2010, 3GPP Organizational Partners (ARIB, ATIS, CCSA, ETSI, TTA, TTC). All rights reserved. UMTS™ is a Trade Mark of ETSI registered for the benefit of its members 3GPP™ is a Trade Mark of ETSI registered for the benefit of its Members and of the 3GPP Organizational Partners LTE™ is a Trade Mark of ETSI currently being registered for the benefit of its Members and of the 3GPP Organizational Partners GSM® and the GSM logo are registered and owned by the GSM Association

3GPP

Release 9

3

3GPP TR 36.814 V9.0.0 (2010-03)

Contents
Foreword................................................................................................................................................6 1 Scope...................................................................................................................................................7 2 References...........................................................................................................................................7 3 Definitions, symbols and abbreviations...............................................................................................7
3.1 Definitions.............................................................................................................................................................7 3.2 Symbols.................................................................................................................................................................7 3.3 Abbreviations........................................................................................................................................................8

4 Introduction.........................................................................................................................................8 5 Support of wider bandwidth................................................................................................................8
5.1 MAC-PHY interface.............................................................................................................................................8 5.2 DL control signalling............................................................................................................................................8 5.3 UL control signalling............................................................................................................................................9

6 Uplink transmission scheme..............................................................................................................10
6.1 Uplink spatial multiplexing.................................................................................................................................10 6.2 Uplink transmit diversity.....................................................................................................................................11 6.2.1 Transmit Diversity for Uplink Control Channel..............................................................................................11 6.3 Uplink multiple access........................................................................................................................................12 6.4 Uplink reference signals......................................................................................................................................12 6.5 Uplink power control..........................................................................................................................................12

7 Downlink transmission scheme.........................................................................................................12
7.1 Physical channel mapping...................................................................................................................................12 7.2 Downlink spatial multiplexing............................................................................................................................12 7.2.1 Feedback in support of downlink spatial multiplexing....................................................................................14 7.3 Enhanced multiuser MIMO transmission...........................................................................................................15 7.4 Downlink reference signals.................................................................................................................................15 7.4.1 Demodulation reference signal........................................................................................................................15 7.4.2 CSI reference signal.........................................................................................................................................16 7.5 Downlink transmit diversity................................................................................................................................16

8 Coordinated multiple point transmission and reception.....................................................................16
8.1 Downlink coordinated multi-point transmission.................................................................................................16 8.1.1 Terminology and definitions ...........................................................................................................................16 8.1.2 Radio-interface specification areas..................................................................................................................17 8.1.3 Feedback in support of DL CoMP...................................................................................................................18 8.1.3.1 Explicit Feedback in support of DL CoMP....................................................................................................18 8.1.3.2 Implicit Feedback in support of DL CoMP....................................................................................................19 8.1.4 Overhead in support of DL CoMP operation...................................................................................................20 8.2 Uplink coordinated multi-point reception...........................................................................................................20

9 Relaying functionality.......................................................................................................................20
9.1 Relay-eNodeB link for inband relay...................................................................................................................21 9.1.1 Resource partitioning for relay-eNodeB link...................................................................................................21 9.1.2 Backward compatible backhaul partitioning ...................................................................................................21 9.1.3 Backhaul resource assignment.........................................................................................................................22 9.2 Relay-eNodeB link for outband relay.................................................................................................................23 9.3 Relay-eNodeB link for inband relay Type 1b.....................................................................................................23

9A Heterogeneous deployments...........................................................................................................23
9A.1 Rel-8/9 schemes...............................................................................................................................................23 9A.2 Non-Rel-8/9 schemes.......................................................................................................................................24 9A.2.1 CA-based scheme..........................................................................................................................................24 9A.2.2 Non-CA based schemes................................................................................................................................24

3GPP

...................47 10.....2 TDD................................................. Downlink.............................................................1...........................1....................82 A................................3 File dropping criteria.......................................................1.............86 A...39 10...................................................0......................................26 10.....................77 A............................ Uplink........................................2.. Uplink (UMi) .....98 3GPP .............................. Downlink (UMa) .................................................2......................................................2 TDD.........................1.........................................................87 A...........3.........3 FDD.................1..........................................................................2 Microcellular (UMi channel model)..........54 10..............1...................................2........................................6....................1 Polarized antenna modelling....1.......................................................................................... Downlink (InH) ..............1 FDD...........................2..................................2..............24 10...................2 Performance metrics.............2....2.....................54 10....2..1.............................................88 A..........................................................................4 Detailed simulation results.....................2..............2................1...1.............37 10..57 Annex A: Simulation model..............................2 TDD............2.......................................................................2..........................................6 Antenna gain for a given bearing and downtilt angle......................................1 Link simulation Scenarios........87 A..............................................................1.........................................1 Indoor (InH channel model).............2..............1.............................4...............2................................................................58 A......2..3 FDD..........3 Evaluation assumptions for IMT-A.......................................2 System simulation Scenarios.............................................. Downlink (UMa) ...........47 10.................................1 Iterative soft interference cancellation receivers .................................................3 Assumptions for Coordinated Multi point Transmission and Reception Evaluations .................................1.........................................2.........................3............2...................................58 A............................................2......1...............4...........................3 FDD....1...........2 Cell spectral efficiency and cell-edge spectral efficiency for ITU-R requirements..............4......................................................................6.....................2 TDD............................................2..........2...........................4 TDD...............................................................38 10.....1.................................................................2.................................................................................................1..........................3 Base coverage urban (UMa channel model).....56 10............................................8 Effective IoT.................................................................................58 A.... Uplink (InH) ......2..............................................1 Cell spectral efficiency and cell-edge spectral efficiency against 3GPP target......................................................... Downlink...............................1......................................2..............................................................3 FDD...............................................................36 10..........................................................................1 3GPP Case1 (3GPP spatial channel model)...........1..1 Reference system deployments............ Uplink (UMi) .................2 TDD..........................................................................................................................................2.............53 10.....1...........7.......2.........................1...2.................................................................33 10..................3.............58 A.........................................................................................2...........................................1..........................................1.......................2.......................................................................2 Antenna gain for a given direction and mechanical tilt angle.2..................1 System simulation assumptions....94 A. Uplink (InH) .......................1....1............................2......2...... Uplink (UMa) ..2...................................2.............29 10........................2.....................40 10...44 10......Release 9 4 3GPP TR 36..................................................91 A...............................1..........................................90 A.78 A.....1...........3 Downlink CoMP evaluation assumptions for intra-NodeB CoMP..........................31 10..................................2.....60 A....................................2.......2..............1...........................................2..............................3 Traffic models....2.........2....................90 A................................... Uplink (RMa) ......90 A...........................1 FDD..................42 10...........................................................................................................3...........................1 Reference points...4 High speed (RMa channel model).............................1.............................................................................1...........................................................1................. Downlink (UMi) ..............................81 A....................................36 10..................58 A...................... Downlink (UMi) ..........1 FDD.......1.....................................2 System level simulator calibration...............55 10.........................2..2 Channel models.................................................85 A.....................................................................................34 10..............4 Assumptions for Relay Evaluations ...........4 TDD..................................95 A.58 A.......................3........3...............2............................................5 Scheduling and resource allocation..........................4 System performance metrics...2..........1........4 Estimation of range ofλ and K..............................83 A....85 A....................................................................................... Downlink (RMa) ...................................................1.............................1 Homogeneous deployments.................................................25 10......................................................51 10.........................................40 10........................1.........0 (2010-03) 10 Evaluation of techniques for Advanced E-UTRA.....................................................3..............................1.2................................................3. Uplink (UMa) .............4 TDD...........5 Assumptions for indoor RRH/Hotzone Evaluations..............................................49 10...........................................................................................................2.............................................84 A....3.................2............................814 V9............................... Downlink (InH) ...........................................82 A............................2.....................45 10...................................26 10..........................87 A.................................... Downlink (RMa) .................................................................. Uplink (RMa) ......................2 Heterogeneous deployments..........................................1....................................................................................3 FDD............84 A.........1 FTP traffic models................4 TDD....................1............................................1 FDD.......................... Uplink...2.1..........................................2..............4 TDD................................................................1 FDD........................4.........7 Advanced receivers modeling ......

...............................................2 Primary module channel model parameters...1..................................................1........1..........100 B......2........................................1.............................2...........................................................1 Autocorrelation of shadow fading.........................104 B.......1 Generic model.........112 3GPP .................................................................................1..........................1............................................2 Primary module.........................104 B..................................................................101 B................0 (2010-03) Annex B: Channel models (Informative)...98 B...............................0..................1 Void 100 B..1 Channel models .........................................1 Path loss models.................................................................Release 9 5 3GPP TR 36................................100 B...................................................................................111 Annex C: Change history...........................................1.2........................2..........2 References....2.......................104 B.......................................814 V9.................................................................................................................................................................

0.y.0 (2010-03) Foreword This Technical Report has been produced by the 3rd Generation Partnership Project (3GPP). updates.Release 9 6 3GPP TR 36. 2 presented to TSG for approval. y the second digit is incremented for all changes of substance.z where: x the first digit: 1 presented to TSG for information. technical enhancements. The contents of the present document are subject to continuing work within the TSG and may change following formal TSG approval. z the third digit is incremented when editorial only changes have been incorporated in the document.e. i. it will be re-released by the TSG with an identifying change of release date and an increase in version number as follows: Version x. corrections. 3 or greater indicates TSG approved document under change control. etc. 3GPP .814 V9. Should the TSG modify the contents of the present document.

int/publ/R-REP-M. along with the complexity evaluation of each technique.905: "Vocabulary for 3GPP Specifications".e. 3GPP TR 21. it is permanently updated and presented to TSG-RAN meetings.) or non-specific. 3 Definitions.0 (2010-03) 1 Scope This document is related to the technical report for physical layer aspect of the study item “Further advancements for EUTRA” [1]. i. The purpose of this TR is to help TSG RAN WG1 to define and describe the potential physical layer evolution under consideration and compare the benefits of each evolution techniques.814". Physical layer procedures (Release 8)” 3GPP TSG RAN1 R1-094954: "Annex A4 Self-evaluation results for TR36. a non-specific reference implicitly refers to the latest version of that document in the same Release as the present document. 3GPP TR 36. complexity.814 V9. edition number.Release 9 7 3GPP TR 36. version number.itu. 2 References The following documents contain provisions which. • References are either specific (identified by date of publication. 2008-11. symbols and abbreviations 3.2135. • For a specific reference. [1] [2] [3] [4] [5] [6] 3GPP TD RP-080137: "Proposed SID on LTE-Advanced".2 Void 3GPP .1 Void Definitions Symbols 3. the latest version applies. Guidelines for evaluation of radio interface technologies for IMTAdvanced.2135-2008/en 3GPP TS 36. and draw a conclusion on way forward. etc. This activity involves the Radio Access work area of the 3GPP studies and has impacts both on the Mobile Equipment and Access Network of the 3GPP systems. constitute provisions of the present document. through reference in this text. http://www. subsequent revisions do not apply.913: “Requirements for Evolved UTRA (E-UTRA) and Evolved UTRAN (EUTRAN) ITU-R Report M.213: “Evolved Universal Terrestrial Radio Access (E-UTRA). This document is intended to gather all information in order to compare the solutions and gains vs. In the case of a reference to a 3GPP document (including a GSM document). This document is a ‘living’ document. • For a non-specific reference.0.

Independent control region size is applied for each component carrier.1 MAC-PHY interface From a UE perspective.905 [2] and the following apply: 4 Introduction At the 3GPP TSG RAN #39 meeting. If possible. at least when the aggregated numbers of component carriers in the UL and the DL are same. This technical report covers the physical-layer aspects of these technology components. the Study Item description on "Further Advancements for E-UTRA (LTEAdvanced)" was approved [1]. . Consideration of non-backwardcompatible configurations of LTE-A component carriers is not precluded The L1 specification shall support carrier aggregation for both contiguous and non-contiguous component carriers with each component carrier limited to a maximum of 110 Resource Blocks using the Release 8 numerology For contiguous carrier aggregation. In typical TDD deployments. provided that the structure of the component carrier follows the Rel-8 specifications.814 V9. A UE may be scheduled over multiple component carriers simultaneously. up to 100MHz and for spectrum aggregation. An LTE Rel-8 terminal can receive and transmit on a single component carrier only. and downlink HARQ ACK/NACK indication are described below.3 Abbreviations For the purposes of the present document. 5 Support of wider bandwidth Carrier aggregation. Each transport block is mapped to a single component carrier only. The study item covers technology components to be considered for the evolution of EUTRA. 3GPP . 5. where two or more component carriers are aggregated. On any carrier with a control region. 5.0 (2010-03) 3. A terminal may simultaneously receive or transmit one or multiple component carriers depending on its capabilities: An LTE-Advanced terminal with reception and/or transmission capabilities for carrier aggregation can simultaneously receive and/or transmit on multiple component carriers. there is one transport block (in absence of spatial multiplexing) and one hybrid-ARQ entity per scheduled component carrier. e. the abbreviations defined in 3GPP TS 21. uplink and downlink resource assignments. coding.g. the needed frequency spacing between the contiguous component carriers will be studied by RAN WG4.2 DL control signalling The design principles for downlink control signalling of control region size.. It shall be possible to configure all component carriers LTE Release 8 compatible. to fulfil the requirements on IMT-Advanced. This study should include the supported number of RBs per component carrier and the needed guard bands between and at the edges for a certain aggregation case.Release 9 8 3GPP TR 36. Rel-8 design (modulation. - It will be possible to configure a UE to aggregate a different number of component carriers of possibly different bandwidths in the UL and the DL. mapping to resource elements) for PCFICH is reused.0. the same solution should be used in the L1 specifications for contiguous and non-contiguous aggregation. RAN WG4 will study the supported combinations of aggregated component carrier and bandwidths. the number of component carriers and the bandwidth of each component carrier in UL and DL will be the same.g. is supported by LTE-Advanced in order to support wider transmission bandwidths e.

following principles are applied. same CCE-based resource mapping) and DCI formats are used on each component carrier. mapping to resource elements) are reused. scheduling request and channel state information (CSI) on PUCCH are described below. Rel-8 PDCCH structure (same coding. not system-specific or cellspecific) CIF is a fixed 3-bit field. and Rel-8 PDCCH structure (same coding. -The scheduling request is transmitted on PUCCH and is semi-statically mapped onto one UE specific UL component carrier. PHICH physical transmission aspects from Rel-8 (orthogonal code design.For signalling of HARQ ACK/NACK on PUCCH for downlink (PDSCH) transmission. The CSI is semi-statically mapped onto one UE specific UL component carrier and the design follows the Rel-8 principles for CQI/PMI/RI. considering ways to reduce reporting overhead or to extend CSI payload. following mechanisms are supported: . while not aiming to optimise for the case of large number of UEs being simultaneously scheduled on multiple DL component carriers. -Periodic CSI reporting on PUCCH is supported for up to five DL component carriers.All HARQ ACK/NACK for a UE can be transmitted on PUCCH in absence of PUSCH transmission.814 V9.PDCCH on a component carrier assigns PDSCH resources on the same component carrier and PUSCH resources on a single linked UL component carrier. transmission of one ACK/NACK for each DL component carrier transport block is supported. 5.e.Release 9 9 3GPP TR 36. following mechanisms are supported. PHICH is transmitted only on the downlink component carrier that was used to transmit the UL grant At least in case that the number of downlink component carriers are more than or equal to that of uplink component carriers and no carrier indicator field is used.The Rel-10 PUCCH design supports up to five DL component carriers. where Configuration for the presence of CIF is semi-static and UE specific (i. limited transmission of ACK/NACK for the DL component carrier transport blocks is supported. .PDCCH on a component carrier can assign PDSCH or PUSCH resources in one of multiple component carriers using the carrier indicator field(CIF).0.For signalling of resource assignments for downlink (PDSCH) and uplink (PUSCH) transmission. same CCE-based resource mapping) is reused. 3GPP . . .3 UL control signalling The design principles for uplink control signalling of HARQ ACK/NACK. The design of the ACK/NACK resource allocation should consider performance and power control aspects.0 (2010-03) . In general. CIF location is fixed irrespective of DCI format size Cross-carrier assignments can be configured both when the DCI formats have the same or different sizes Explicit CIF at least for the case of same DCI format size There will be an upper limit on the total number of blind decodes For signalling of downlink HARQ ACK/NACK indication. modulation. In case of power limitation. . scrambling sequence. the Rel-8 PHICH resource mapping rule is reused.

For FDD and TDD.0 (2010-03) 6 6. It is possible to configure the uplink single user spatial multiplexing transmission with or without the layer shifting. If layer shifting is configured.1-2. the number of precoding matrices is 20.1-1: 3-bit precoding codebook for uplink spatial multiplexing with two transmit antennas Codebook index 0 1 2 3 4 5 Number of layers 1 υ 2 1 1   1 2  1 1    2 −1 1 1    2  j 1 1    2 − j  1 1   2 0 1 0   2 1  1 1  2 0 0 1  - For uplink spatial multiplexing with four transmit antennas. Codebook 3GPP . Depending on the number of transmission layers. For uplink spatial multiplexing with two transmit antennas. Otherwise. up to two transport blocks can be transmitted from a scheduled UE in a subframe per uplink component carrier. a 6-bit precoding codebook is used.1-3. In case of the layer shifting. For 3-layer transmission. The subset of the precoding codebook used for 1-layer transmission is defined in Table 6. precoding is applied after the layer mapping.0. the HARQ-ACKs for all transport blocks are bundled into a single HARQ-ACK.1-2: 6-bit precoding codebook for uplink spatial multiplexing with four transmit antennas: precoding matrices for 1-layer transmission. precoding is performed according to a predefined codebook. The baseline for the subset of the precoding codebook used for 2-layer transmission is defined in Table 6.1-1 is used. In case of full-rank transmission.1 Uplink transmission scheme Uplink spatial multiplexing Uplink spatial multiplexing of up to four layers is supported by LTE-Advanced. and only BPSK or QPSK alphabets are used for non-zero elements in precoding matrices. The transmission rank can be adapted dynamically. shifting in time domain is supported.814 V9. Each transport block has its own MCS level. Table 6. Application of a single precoding matrix per uplink component carrier is supported. precoding is applied after the layer shifting operation. In the uplink single user spatial multiplexing.Release 9 10 3GPP TR 36. each transport block has its own HARQ-ACK feedback signalling. only identity precoding matrix is supported. 3bit precoding codebook as defined in Table 6. one-bit NACK is transmitted to the UE. If layer shifting is not configured. Table 6. One-bit ACK is transmitted to the UE if all transport blocks are successfully decoded by the eNodeB. If layer shifting is not configured. the modulation symbols associated with each of the transport blocks are mapped onto one or two layers according to the same principle as in Rel-8 E-UTRA downlink spatial multiplexing. If layer shifting is configured.

For a given UE. For the UE with four transmit antennas.814 V9.Release 9 11 3GPP TR 36. PUSCH and SRS transmissions. Codebook Index 0 to 7 1  1 1 2 0  0 1  1 0 2 1  0 0  0 1  1 2  − j 1 1  0  0 0 1 0  1 − j   1 2  0   j 0 0  1 1  0 2  −1 1 0  −1  0 0 0 1  1 2  1 0 1 1  0 2  1 1 − j  0  0 1 0  −1  0 0 0 1  1 2  −1 0 1 1  0 2  −1 1 −1  0  0 1 0  0  1 0  0  1  1  2  − j 0 1  1  0 1 −1  0  0 0 0 1  1 2  j 0  1 1  −1 2  0  1 j  0  0 1 0  0  −1 0 1 0 1  j   1  2 0   1 0 0 1 1  1 2  0 1 0  0  −1 0 0  1  −1 0 1   −1  0 Index 8 to 15 0 1  1  1 0  0  2 1   1 0 1  1 0 2 0  1 6.0 (2010-03) Index 0 to 7 1    11  21    −1 1    1 −1 21    1  1   1 0 2 1   0 1    1 1  2  j    j 1    1  −1  2 j    − j  1    10  2 −1   0  1    1 1  2 −1   1  1    1 −1 2 −1   −1 1    1 0  2  j   0  1    11  2 − j    − j  1    1  −1  2 − j     j  1    10  2 − j    0  1    1  j 2 1     j 1    1 − j  2 1    − j  0   1 1 2 0   1 1    1  j 2  j   1  1    1 − j  2 j     −1  0    11  20    −1 1    1 j  2  −1    − j  1    1 − j  2  −1     j  0    1 1  2 0     j 1    1 j  2 − j     −1  1    1 − j  2 − j    1   0    11  2 0    − j  Index 8 to 15 Index 16 to 23 Table 6. the 2-tx transmit diversity scheme is applied. 3GPP . the uplink Single Antenna Port Mode can be independently configured for its PUCCH. on two separate orthogonal resources. the spatial orthogonal-resource transmit diversity (SORTD) scheme is supported for transmissions with two antenna ports.1 Transmit Diversity for Uplink Control Channel For uplink control channels with Rel-8 PUCCH format 1/1a/1b. The uplink Single Antenna Port Mode is the default mode before eNodeB is aware of the UE transmit antenna configuration.2.0.1-3: 6-bit precoding codebook for uplink spatial multiplexing with four transmit antennas: precoding matrices for 2-layer transmission. 6. where the UE behaviour is same as the one with single antenna from eNodeB’s perspective. the same modulation symbol from the uplink channel is transmitted from two antenna ports.2 Uplink transmit diversity For UEs with multiple transmit antennas. In this transmit diversity scheme. an uplink Single Antenna Port Mode is defined.

0 (2010-03) 6.3 Uplink multiple access DFT-precoded OFDM is the transmission scheme used for PUSCH both in absence and presence of spatial multiplexing.5 - Uplink power control UL power control mainly compensates for slow-varying channel conditions while reducing the interference generated towards neighboring cells Fractional path-loss compensation or full path-loss compensation is used on PUSCH and full path-loss compensation on PUCCH Scope of uplink power control in LTE-Advanced is similar to Rel’8: LTE-Advanced supports component carrier specific UL power control for both contiguous and non-contiguous carrier aggregation for closed-loop case.814 V9. Rel-8 principles are reused. there is one DFT per component carrier. and for open loop at least for the cases that the number of downlink component carriers is more than or equal to that of uplink component carriers.2 Downlink spatial multiplexing Downlink spatial multiplexing of up to eight layers is supported byLTE-Advanced. Each transport block is assigned its own modulation and coding scheme. 7 7. In case of multiple component carriers. same CP length is used for control and data Relation between CP length of normal and MBSFN subframes in the control region is the same as in Rel-8 7. 6. one bit is used for each transport block.0. both normal and extended cyclic prefix can be used for control and data region.Release 9 12 3GPP TR 36. Simultaneous transmission of uplink L1/L2 control signalling and data is supported through two mechanisms Control signalling is multiplexed with data on PUSCH according to the same principle as in Rel-8 Control signalling is transmitted on PUCCH simultaneously with data on PUSCH 6. two types of uplink reference signals are supported by LTE-Advanced: The precoding applied for the demodulation reference signal is the same as the one applied for the PUSCH. In the downlink 8-by-X single user spatial multiplexing. up to two transport blocks can be transmitted to a scheduled UE in a subframe per downlink component carrier. The baseline for sounding reference signal in LTE-Advanced operation is non-precoded and antenna-specific. Both frequencycontiguous and frequency-non-contiguous resource allocation is supported on each component carrier.4 - Uplink reference signals Demodulation reference signal Sounding reference signal Similar to LTE. 3GPP .1 Downlink transmission scheme Physical channel mapping LTE-Advanced supports the PDSCH to be mapped also to MBSFN (non-control) region of MBSFN subframes that are not used for MBMS In case of PDSCH mapping to MBSFN subframes. For multiplexing of the sounding reference signals. For HARQ ACK/NAK feedback on uplink. Cyclic shift separation is the primary multiplexing scheme of the demodulation reference signals.

2-1... Complex-valued modulation symbols (q) d ( q ) (0).3. d ( q ) ( M sym −1) for code word q shall be mapped onto the layers b T lay er 1 x(i ) = x ( 0) (i ) ...211. i =0.. the layer mapping shall be done according to Table 7.0 (2010-03) A transport block is associated with a codeword. [ ] υ lay er is the number of layers and M symb is 3GPP .2 of TS 36. For up to four layers.0.1..814 V9.. M symb − where the number of modulation symbols per layer..3. the codeword-to-layer mapping is according to section 6. which is for retransmission of one out of two codewords that were initially transmitted with more than four layers... For above four layers as well as the case of mapping one codeword to three or four layers..Release 9 13 3GPP TR 36. x (υ−1) (i ) .

0 (2010-03) Table 7...Release 9 14 3GPP TR 36.814 V9...2.1.0. 3GPP . M symb − 1 3 1 x ( 0 ) (i ) = d ( 0 ) (3i ) layer (0) x (1) (i ) = d ( 0 ) (3i + 1)M symb = M symb 3 ( 2) ( 0) x (i ) = d (3i + 2) 4 1 x ( 0 ) (i ) = d ( 0 ) (4i ) x (1) (i ) = d ( 0 ) (4i + 1)M layer symb x ( 2 ) (i ) = d ( 0 ) (4i + 2) x (3) (i ) = d ( 0 ) (4i + 3) x ( 0) (i ) = d ( 0 ) (2i ) x (1) (i ) = d ( 0 ) (2i +1) (0) = M symb 4 5 2 layer ( 0) (1) x ( 2) (i ) = d (1) (3i ) M symb = M symb 2 = M symb 3 x (3) (i ) = d (1) (3i + 1) x ( 4) (i ) = d (1) (3i + 2) 6 2 x ( 0 ) (i ) = d ( 0) (3i ) x (1) (i ) = d ( 0 ) (3i + 1) x ( 2) (i ) = d ( 0 ) (3i + 2) x (i ) = d (3i ) x ( 4 ) (i ) = d (1) (3i + 1) x ( 5) (i ) = d (1) (3i + 2) ( 3) (1) layer ( 0) (1) M symb = M symb 3 = M symb 3 x ( 0 ) (i ) = d ( 0) (3i ) x (1) (i ) = d ( 0 ) (3i + 1) x ( 2) (i ) = d ( 0 ) (3i + 2) 7 2 x (3) (i ) = d (1) (4i ) M symb x ( 4 ) (i ) = d (1) (4i + 1) x (5) (i ) = d (1) ( 4i + 2) x ( 6 ) (i ) = d (1) ( 4i + 3) x ( 0 ) (i ) = d ( 0 ) (4i ) x (1) (i ) = d ( 0 ) (4i + 1) x ( 2) (i ) = d ( 0) (4i + 2) x (3) (i ) = d ( 0) (4i + 3) x (i) = d (4i ) x (5) (i) = d (1) ( 4i + 1) x ( 6) (i) = d (1) ( 4i + 2) x ( 7 ) (i ) = d (1) (4i + 3) ( 4) (1) layer ( 0) (1) = M symb 3 = M symb 4 8 2 layer ( 0) (1) M symb = M symb 4 = M symb 4 7.1 Feedback in support of downlink spatial multiplexing The baseline for feedback in support of downlink single-cell single-user spatial multiplexing is codebook-based precoding feedback.2-1: Codeword-to-layer mapping for above four layers and the case of mapping one codeword to three or four layers Number of layers Number of code words Codeword-to-layer mapping layer i = 0.

For rank 2. etc are FFS).Release 9 15 3GPP TR 36. Figure 7.1-2. Present only in resource blocks and layers scheduled by the eNodeB for transmission.4 - Downlink reference signals Reference signals targeting PDSCH demodulation Reference signals targeting CSI estimation (for CQI/PMI/RI/etc reporting when needed) Two types of downlink reference signals are considered for LTE-Advanced: The reference signal structure can be used to support multiple LTE-Advanced features. The design principle for the demodulation reference signals is an extension to multiple layers of the concept of Rel-8 UE-specific reference signals used for beamforming (details on pattern. Basic principle of MU-MIMO in LTEAdvanced is as follows: Switching between SU. CoMP and spatial multiplexing. spreading and scrambling) as in LTE Rel-9 is used.0 (2010-03) 7.1 - Demodulation reference signal Reference signals targeting PDSCH demodulation are: UE-specific. DM-RS for 1st and 2nd layers and that for 3rd and 4th layers are multiplexed by means of frequency division multiplexing (FDM).0. DM-RS for 1st layer and that for 2nd layer are multiplexed by means of code division multiplexing (CDM) by using orthogonal cover code (OCC) over two consecutive resource elements (blue) in time domain. as illustrated in Figure 7. DM-RS for 1st layer and that for 2nd layer are multiplexed by means of CDM by using OCC over two consecutive resource elements (blue) in time domain. i.1-1: DM-RS pattern for rank 1 and 2 For rank 3 and 4. the same DM-RS structure (including patterns. Normal subframes with normal CP: For rank 1and 2. e.3. 3GPP .3. location. 7. Reference signals transmitted on different layers are mutually orthogonal at the eNodeB.and MU-MIMO transmission is possible without RRC reconfiguration 7. Complementary use of Rel-8 cell-specific reference signals by the UE is not precluded.g.4.e.4.814 V9. the PDSCH and the demodulation reference signals intended for a specific UE are subject to the same precoding operation.3 Enhanced multiuser MIMO transmission Enhanced multi-user MIMO transmission is supported in LTE-Advanced.1-1. the DM-RS pattern is illustrated in Figure 7. DM-RS for 3rd layer and that for 4th layer are multiplexed by means of CDM by using OCC over two consecutive resource elements (green) in time domain.

and PDSCH in nonMBSFN subframes. the cell-edge throughput and/or to increase system throughput in both high load and low load scenarios. the Rel-8 transmit diversity scheme is used.1 8.0. This is the serving cell of Rel-8 (concept that already exists) CoMP categories: Joint Processing (JP): data is available at each point in CoMP cooperating set (definition below) Joint Transmission: PDSCH transmission from multiple points (part of or entire CoMP cooperating set) at a time 3GPP frequency Coordinated multi-point (CoMP) transmission/reception is considered for LTE-Advanced as a tool to improve the coverage of high data rates.4. 8 Coordinated multiple point transmission and reception 8. General terminology: Serving cell: Cell transmitting PDCCH assignments (a single cell).4.1.1 Downlink coordinated multi-point transmission Terminology and definitions Downlink coordinated multi-point transmission implies dynamic coordination among multiple geographically separated transmission points.5 Downlink transmit diversity For the downlink transmit diversity with more than four transmit antennas applied to PDCCH.1-2: DM-RS pattern for rank 3 and 4 7.814 V9.2 - CSI reference signal Reference signals targeting CSI estimation are cell specific sparse in frequency and time. .Release 9 16 3GPP TR 36. punctured into the data region of normal/MBSFN subframe 7.0 (2010-03) Figure 7.

a single point is the transmission point at every subframe. information needs to be signalled between eNodeBs. channel reciprocity may be exploited Reporting to facilitate the decision on the set of participating transmission points For TDD.Release 9 17 3GPP TR 36. we have: RRM measurement set: in support of RRM measuremets (already in Rel-8) and therefore not CoMP specific 8. the CoMP transmission points are the points in the CoMP cooperating set For Dynamic cell selection.3 The CoMP measurement set may be the same as the CoMP cooperating set The actual UE reports may down-select cells for which actual feedback information is transmitted (reported cells) In addition.814 V9. Note that this set may or need not be transparent to the UE. CoMP transmission point(s): point or set of points actively transmitting PDSCH to UE CoMP transmission point(s) is a subset of the CoMP cooperating set For Joint transmission.1. to (coherently or non-coherently) improve the received signal quality and/or cancel actively interference for other UEs Dynamic cell selection: PDSCH transmission from one point at a time (within CoMP cooperating set) Coordinated Scheduling/Beamforming (CS/CB): data is only available at serving cell (data transmission from that point) but user scheduling/beamforming decisions are made with coordination among cells corresponding to the CoMP cooperating set. This transmission point can change dynamically within the CoMP cooperating set. For Coordinated scheduling/beamforming. CoMP sets: CoMP cooperating set Set of (geographically separated) points directly or indirectly participating in PDSCH transmission to UE. there is no difference from the UE perspective if the cells belong to the same eNodeB or different eNodeBs.g.1.0 (2010-03) - data to a single UE is simultaneously transmitted from multiple transmission points. From a radio-interface perspective. channel reciprocity may be exploited - Preprocessing schemes Joint processing prior to transmission of the signal over the multiple transmission points Downlink control signaling to support the transmission scheme - Reference signal design 3GPP .0. Potential impact on the radio-interface specifications is foreseen in mainly three areas: Feedback and measurement mechanisms from the UE Reporting of dynamic channel conditions between the multiple transmission points and the UE For TDD. e. the CoMP transmission point corresponds to the “serving cell” CoMP measurement set: set of cells about which channel state/statistical information related to their link to the UE is reported as discussed in section 8.2 Radio-interface specification areas Downlink coordinated multi-point transmission should include the possibility of coordination between different cells. If inter-eNodeB coordination is supported.

UE CoMP feedback reports target the serving cell (on UL resources from serving cell) as baseline when X2 interface is available and is adequate for CoMP operation in terms of latency and capacity. 8. or 3GPP . Do not have to confine the CoMP studies to payload sizes currently supported by PUCCH operation.. Complementary inter-cell feedback might be needed. and for cases where feedback reports to the serving cell causes large interference (e.1 Explicit Feedback in support of DL CoMP This section lists different forms of explicit feedback in support of DL CoMP.3.g. For the CoMP schemes that require feedback. Combinations of full or subset of above three are possible.0 (2010-03) - Depending on the transmission scheme.Any additional feedback designed for CoMP shall be consistent with the feedback framework for SU/MU-MIMO.Release 9 18 3GPP TR 36. In this case.814 V9. UL overhead (number of bits) associated with each specific feedback mechanism needs to be identified. specification of additional reference signals may be required. They are all characterized by having a channel part and a noise-and-interference part.As baseline. or JP with MBSFN subframes) is the same as that for non CoMP (SU/MU-MIMO). without assuming any transmission or receiver processing Implicit channel state/statistical information feedback feedback mechanisms that use hypotheses of different transmission and/or reception processing.0.g. Channel part: For each cell in the UE’s measurement set that is reported in a given subframe. .. the network need not explicitly signal to the UE the CoMP transmission point(s) and the UE reception/demodulation of CoMP transmissions (CS/CB. Look at these types of feedback mechanisms for the evaluations. if found needed.1.” . a solution needs to be identified..1. CQI/PMI/RI - UE transmission of SRS can be used for CSI estimation at eNB exploiting channel reciprocity. DL performance tradeoff should be assessed with the goal to target minimum overhead for a given performance. for example. The feedback reporting for cases with X2 interface not available or not adequate (latency and capacity).3 - Feedback in support of DL CoMP The three main categories of CoMP feedback mechanisms have been identified to be: Explicit channel state/statistical information feedback Channel as observed by the receiver. the reception of UE reports at cells other than the serving cell is a network implementation choice. e. CSI RS design should take potential needs of DL CoMP into account New forms of feedback and signaling may be needed to support CoMP that are. Two possibilities should be studied the “container” of the DL CoMP feedback: Expand the supported PUCCH payload sizes Use periodic/a-periodic reports on PUSCH 8. one or several channel properties are reported Channel properties include (but are not limited to) the following (‘i‘ is the cell index): Channel matrix (Hi) – short term (instantaneous) The full matrix Hi. The feedback overhead (UL) vs. individual per-cell feedback is considered as baseline. in heteronegenous deployment scenarios) for CoMP operation needs to be discussed and. configured by RRC for a given UE.

Multi user MIMO Single cell vs.1.g. j=0. e. or main eigen component(s) 8. there may be a semi-static hypothesis configuration e. tx weights) JP: multiple single-cell or multi-cell PMI capturing coherent or non-coherent channel across reported cells CB/CS: Single-cell or multiple single-cell PMIs capturing channel from the reported cell(s) to the UE Transmit precoder based on or derived from the PMI weight Other types of feedbacks. main Multi-cell eigen-component.2 - Implicit Feedback in support of DL CoMP This section lists different forms of implicit feedback in support of DL CoMP. instead of PMI are being considered - Receive processing (i. e. Interference outside the cells reported by the UE CoMP transmission points Total receive power (Io) or total received signal covariance matrix Covariance matrix of the noise-and-interference the full matrix.J-1.1.3 UE transmission of SRS in support of DL CoMP 3GPP . (‘j’ is span over time or frequency) The full matrix Ri. multi-point (JP) transmission Within Joint processing CoMP: Subsets of transmission points or subsets of reported cells (Joint Transmission) CoMP transmission point(s) (Dynamic Cell Selection) Transmit precoder (i.0 (2010-03) - main eigen component(s) of Hi Transmit channel covariance (Ri).g.: Single vs. precoded RS may be used to allow UE to generate refined CQI/RI feedback 8. where Ri = (sum{Hij†Hij})/J.e. rx weights) Interference based on particular tx/rx processing There may be a need for the UE to convey to the network the hypothesis or hypotheses used (explicit signalling of hypothesis to eNB). grouping of hypotheses (explicit signalling of hypothesis to the UE). There are hypotheses at the UE and the feedback is based on one or a combination of two or more of the following.…. Coordinated transmission Within coordinated transmission: Single point (CB/CS) vs.814 V9.1.2.3.3.0.e.g. or main eigen component(s) of Ri - Inter-cell channel properties may also be reported Noise-and interference part. e. And/or.Release 9 19 3GPP TR 36. And/or.g..

Depending on the relaying strategy. geographically separated points. 3GPP . and Type 2 relay are examples of this type of relaying. “Type 1a relay nodes”. Rel-8 UEs should be able to connect to the donor cell in this case. “Type 1 relay nodes”. In the case the relay is in control of cells of its own. Uplink CoMP reception can involve joint reception (JR) of the transmitted signal at multiple reception points and/or coordinated scheduling (CS) decisions among cells to control interference and may have some RAN1 specification impact. relays can be classified into transparent. 9 Relaying functionality Relaying is considered for LTE-Advanced as a tool to improve e. while parts of the RRM may be located in the relay.g.0 (2010-03) This section lists UE transmission of SRS related feedback in support of DL CoMP. With respect to the relay node’s usage of spectrum. non-transparent.Release 9 20 3GPP TR 36. group mobility.2). 8. Uplink coordinated multi-point reception is expected to have very limited impact on the RAN1 specifications. At least part of the RRM is controlled by the eNodeB to which the donor cell belongs. The same RRM mechanisms are available and from a UE perspective there is no difference in accessing cells controlled by a relay and cells controlled by a “normal” eNodeB. and “Type 1b relay nodes” use this type of relaying.814 V9.2 Uplink coordinated multi-point reception Coordinated multi-point reception implies coordination among multiple. outband. Smart repeaters. UE transmission of SRS can be used for CSI estimation at multiple cells exploiting channel reciprocity. Enhanced SRS schemes may be considered. Self-backhauling (L3 relay). For both inband and outband relaying. In this case. a relay should preferably support also LTE Rel-8 UEs. it shall be possible to operate the eNB-to-relay link on the same carrier frequency as eNB-to-UE links With respect to the knowledge in the UE.0. At least “Type 1” and “Type 1a” relay nodes are part of LTE-Advanced. The cells controlled by the relay should support also LTE Rel-8 UEs. the cell-edge throughput and/or to provide coverage in new areas. the relay controls one or several cells and a unique physical-layer cell identity is provided in each of the cells controlled by the relay. Studies on CSI RS impact on PDSCH transmissions to Rel-8 UEs for various RS densities needed There should be no impact from CSI RS transmission on transmission of PBCH/PSS/SSS 8. its operation can be classified into: inband. different types of L2 relays. a relay may be part of the donor cell control cells of its own In the case the relay is part of the donor cell. temporary network deployment. The need for extended CP operation in certain UL subframes should be further investigated. the relay does not have a cell identity of its own (but may still have a relay ID). Rel8 UEs should be able to connect to the donor cell in this case. the coverage of high data rates. in which case the UE is aware of whether or not it is communicating with the network via the relay.1. in which case the eNB -relay link shares the same carrier frequency with relay-UE links. decode-andforward relays. in which case the UE is not aware of whether or not it communicates with the network via the relay.4 - Overhead in support of DL CoMP operation DL CoMP operation overhead is very much related to the DL-RS structure (see section 7. in which case the eNB-relay link does not operate in the same carrier frequency as relay-UE links. The relay node is wirelessly connected to the radio-access network via a donor cell.

… In the context of single-cell operation. some resources in the time-frequency space are set aside for the backhaul link (Un) and cannot be used for the access link (Uu).0.1 9.2 Backward compatible backhaul partitioning Due to the relay transmitter causing interference to its own receiver. 9. a Rel-8 UE is not aware of the presence of a Type 2 relay node. It is transparent to Rel-8 UEs. each of which appears to a UE as a separate cell distinct from the donor cell The cells shall have their own Physical Cell ID (defined in LTE Rel-8) and the relay node shall transmit its own synchronization channels.1.Release 9 21 3GPP TR 36.814 V9.1. At least the following scheme will be supported for this resource partitioning: General principle for resource partitioning at the relay: eNB → RN and RN → UE links are time division multiplexed in a single carrier frequency (only one is active at any time) RN → eNB and UE → RN links are time division multiplexed in a single carrier frequency (only one is active at any time) Multiplexing of backhaul links in FDD: eNB → RN transmissions are done in the DL frequency band RN → eNB transmissions are done in the UL frequency band Multiplexing of backhaul links in TDD: eNB → RN transmissions are done in the DL subframes of the eNB and RN RN → eNB transmissions are done in the UL subframes of the eNB and RN 9.0 (2010-03) A “Type 1” relay node is an inband relaying node characterized by the following: It control cells. except “Type 1a” operates outband and “Type 1b” operates inband with adequate antenna isolation. “Type 1a” and “Type 1b” relay nodes are characterised by the same set of features as the “Type 1” relay node above. simultaneous eNodeB-to-relay and relay-to-UE transmissions on the same frequency resource may not be feasible unless sufficient isolation of the outgoing and incoming signals is provided e.1 Relay-eNodeB link for inband relay Resource partitioning for relay-eNodeB link In order to allow inband relaying. reference symbols.g. it should be possible for a relay node to appear differently than Rel-8 eNodeB to allow for further performance enhancement. at the relay it may not be possible to receive UE transmissions simultaneously with the relay transmitting to the eNodeB. At least. the UE shall receive scheduling information and HARQ feedback directly from the relay node and send its control channels (SR/CQI/ACK) to the relay node It shall appear as a Rel-8 eNodeB to Rel-8 UEs (i. it does not transmit CRS and PDCCH. A “Type 1a” relay node is expected to have little or no impact on RAN1 specifications. It can transmit PDSCH. by means of specific. 3GPP . well separated and well isolated antenna structures. be backwards compatible) To LTE-Advanced UEs. A “Type 2” relay node is an inband relaying node characterized by the following: It does not have a separate Physical Cell ID and thus would not create any new cells. Similarly.e.

use of common search space. or implicitly derived from the downlink backhaul subframes using the HARQ timing relationship A new physical control channel (here referred to as the “R-PDCCH”) is used to dynamically or “semipersistently” assign resources. The R-PDCCH may assign uplink resources in one or more later subframes. but allow removing some unnecessary procedure or bandwidthwasting procedure by considering the relay property. etc. to create “gaps” in the relay-to-UE transmission. interleaving. the following is valid At the RN. except for possible adjustment to allow for RN transmit/receive switching The set of downlink backhaul subframes. 9.1. Within the PRBs semi-statically assigned for R-PDCCH transmission.2-1. multiplexing.1. If the “search space” approach of R8 is used for the backhaul link.2-1: Example of relay-to-UE communication using normal subframes (left) and eNodeB-torelay communication using MBSFN subframes (right). is the baseline. Relay-to-eNodeB transmissions can be facilitated by not allowing any terminal-to-relay transmissions in some subframes.0 (2010-03) One possibility to handle the interference problem is to operate the relay such that the relay is not transmitting to terminals when it is supposed to receive data from the donor eNodeB. is semistatically assigned The set of uplink backhaul subframes.3 - Backhaul resource assignment In case of downlink backhaul in downlink resources. If RN-specific search space is configured. “R-PDSCH” and “R-PDCCH” can be transmitted within the same PRBs or within separated PRBs. a subset of the resources is used for each R-PDCCH.0. it could be implicitly or explicitly known by RN.Release 9 22 3GPP TR 36. within the semi-statically assigned sub-frames. The R-PDCCH may assign downlink resources in the same and/or in one or more later subframes. eNB-to-relay transmission One subframe transmission gap (“MBSFN subframe”) Ctrl Data Ctrl No relay-to-UE transmission Figure 9. The detailed R-PDCCH transmitter processing (channel coding.1. can be semistatically assigned. These “gaps” during which terminals (including Rel-8 terminals) are not supposed to expect any relay transmission can be created by configuring MBSFN subframes as exemplified in Figure 9. during which downlink backhaul transmission may occur. the access link downlink subframe boundary is aligned with the backhaul link downlink subframe boundary. for the downlink backhaul data (corresponding to the “R-PDSCH” physical channel). - - - - - - 3GPP . These resources may correspond to the full set of OFDM symbols available for the backhaul link or be constrained to a subset of these OFDM symbols. i. The actual overall set of resources used for R-PDCCH transmission within the above mentioned semi-statically assigned PRBs may vary dynamically between subframes. The resources that are not used for R-PDCCH within the above mentioned semi-statically assigned PRBs may be used to carry R-PDSCH or PDSCH. The “R-PDCCH” is also used to dynamically or “semi-persistently” assign resources for the uplink backhaul data (the “R-PUSCH” physical channel).e. during which uplink backhaul transmission may occur. The R-PDCCH is transmitted starting from an OFDM symbol within the subframe that is late enough so that the relay can receive it.) should reuse Rel-8 functionality to the extent possible. which can be semi-statically configured (and potentially includes entire system bandwidth).814 V9.

921].2 The interference characteristics in a heterogeneous deployment can be significantly different than in a homogeneous deployment. 9A. in case (b) a macro user causes severe interference towards the HeNB and in case (c).g. case (d). Such methods may operate in time. path-loss based cell association (e.2 Relay-eNodeB link for outband relay If relay-eNB and relay-UE links are isolated enough in frequency (possibly with help of additional means such as antenna separation). a macro user with no access to the CSG cell will be interfered by the HeNB. a CSG user is interfered by another CSG HeNB. On the right hand side. by restricting the transmission power during part of the time for at least one cell layer to reduce interference to control signalling on other cell layers or through different power control schemes as discussed in [TR36. the eNB→ RN and RN→ UE (RN→ eNB and UE→ RN) links can be activated simultaneously without the need for the time division multiplexing. Enhancements to these types of schemes can be considered in releases beyond Rel-9. e.1. 9A Heterogeneous deployments Heterogeneous deployments consist of deployments where low power nodes are placed throughout a macro-cell layout.0. then there is no interference issue in activating both links simultaneously.Release 9 23 3GPP TR 36. if the same carrier is used across different cell layers. by appropriate arrangement of the respective antennas for the Un and Uu links. it becomes possible for relay-eNodeB link to reuse the channels designed for UE-eNodeB link. preliminary results indicate that methods for handling the uplink and downlink interference towards data as well as L1/L2 control signaling. synchronization signals and reference signals are important.1. In case (a).2. Examples hereof are given in Figure 9A-1.1 Rel-8/9 schemes Several methods for handling the above-mentioned interference scenarios using the functionality present in Rel-8/9 can be envisioned. or. 3GPP . Different scenarios under consideration that could occur in heterogeneous deployments are discussed in Appendix A. Therefore. In these scenarios. frequency and/or spatial domains.0 (2010-03) - No “R-PCFICH” exists for indication of backhaul control region size in time and frequency domains 9. Examples hereof include using different carrier frequencies for different cell layers. by using biased RSRP reports) may improve the uplink but at the cost of increasing the downlink interference of non-macro users at the cell edge.814 V9.g..3 Relay-eNodeB link for inband relay Type 1b If the outgoing and incoming signals at the relay are adequately isolated in the spatial domain. Figure 9A-1: Examples of interference scenarios in heterogeneous deployments. The operation of Type 1b relay nodes may not be supported in all deployment scenarios. 9.

2.2. One example is illustrated in Figure 9A.2 Non-Rel-8/9 schemes A common property of the non-Rel-8/9 schemes below is that they provide means for coordination of the control channel interference between cell layers.2 are in support of the submission of the 3GPP "LTE Release 10 & beyond (LTEAdvanced)" to the ITU-R as a candidate technology for the IMT-Advanced.0 (2010-03) 9A. Rel-8/9 terminals can be scheduled on one component carrier while Rel-10 terminal capable of carrier aggregation can be scheduled on multiple component carriers. Downlink interference for control signaling can be handled by partitioning component carriers in each cell layer into two sets. Time synchronization between the cell layers is assumed in this example. possibly in combination with time shifting. For the data part.814 V9. can be used for heterogeneous deployments. the downlink control signalling is transmitted over part of the carrier bandwidth. Time synchronization between the cell layers is assumed in this approach. including. Macr • Con • Data 10 Evaluation of techniques for Advanced E-UTRA Section 10.2.2. but not limited to. Examples of some mechanism that might be used for interference handling can be found in Section 8.0. In a time-domain scheme. downlink interference coordination techniques can be used.2 Non-CA based schemes Several schemes for control-channel interference handling not relying on carrier aggregation can be envisioned. The evaluation results in Section 10. described in Section 5.1 CA-based scheme Carrier aggregation (CA) with cross-carrier scheduling using CIF . for at least one cell layer to reduce interference to control signalling on other cell layers.1-1. time-domain and frequency-domain schemes.1 presents the evaluation results of the techniques for advanced E-UTRA against the 3GPP targets [3]. 9A. one set used for data and control and one set used mainly for data and possibly control signaling with reduced transmission power.2 and agreed to be part of Rel-10. Figure 9A. Interference management can be achieved by using different parts of the carrier bandwidth for control signalling in different cell layers.Release 9 24 3GPP TR 36. In a frequency-domain scheme.1-1: One example of carrier aggregation applies to heterogeneous deployments. 9A. Time synchronization between the cell layers is assumed. interference handling for downlink control signalling can be achieved by restricting the transmission during part of the time. 3GPP .

3. In the tables for downlink.Release 9 25 3GPP TR 36. uplink FDD.1-2 Simulation parameters for 3GPP targets fulfillment (DL. and 4 show the source specific simulation parameters that characterize the performances of downlink FDD. there are no CRS in data regions.e. FDD)   CCH duration (L symbols) Num of MBSFN subframe Channel estimation Receiver type CSI assumption at eNB Short-term for CS/CB-CoMP Long-term + Short-term for JP-CoMP Long-term Short-term Long-term + Short-term Long-term + Short-term Short-term Short-term Long-term Source 1 3 6 Real MMSE Source 3 Source 4 Source 5 Source 7 Source 11 Source 15 Source 18 3 3 3 3 3 3 3 6 6 6 6 6 6 6 Real Real Real Ideal Real Real Real MMSE with IRC MMSE with IRC MMSE MMSE MMSE with IRC MMSE with IRC MMSE with IRC Table 10. CSI-assumption at eNB is the factor that characterizes the transmit signal processing at eNB for MUMIMO and CoMP schemes.814 V9. In the MBSFN subframes. downlink and uplink CoMP. Note. Tables 10. 2.1 Cell spectral efficiency and cell-edge spectral efficiency against 3GPP target Cell spectral efficiency and cell-edge spectral efficiency are evaluated through extensive simulations conducted by a number of companies.0 (2010-03) 10. In the tables for uplink.1-1 Simulation parameters for 3GPP targets fulfillment (DL.0. control channel (CCH) duration (L OFDM symbols) and the number of the MBSFN subframes are the factors that affect the overhead.. The tables in the following subsections show the simulation results focusing on LTE-Advanced configurations. the PUCCH bandwidth is the factor that affects the overhead. and uplink SU-MIMO. which effectively reduce the overhead. i. Table 10.1-1. however. TDD)   Source 1 Source 2 Source 3 Source 4 Source 5 Source 10 Source 19 CCH duration (L symbols) 3 3 3 3 3 3 3 Num of MBSFN subframe 2 0 2 2 0 2 2 Channel estimation Real Real Real Real Real Real Real Receiver type MMSE MMSE with IRC MMSE with IRC MMSE with IRC MMSE MMSE with IRC MMSE with IRC CSI assumption at eNB Short-term Short-term Long-term Short-term Short-term Short-term Short-term 3GPP . Channel estimation and receiver types are the factors that affect the demodulation performance. downlink and uplink MU-MIMO. and uplink TDD. downlink TDD. respectively. that the performance differences among sources could also be explained by other factors such as detailed signal processing algorithms at the transmitter and the receiver.

0.56 2.1-3 Simulation parameters for 3GPP targets fulfillment (UL.1.091 0. and 1-by-4 MU-MIMO for uplink) assuming real channel estimation.23 2.814 V9.1-4 Simulation parameters for 3GPP targets fulfillment (UL. Downlink Tables 10.74 2.1-3 show the spectrum efficiency results of 2-by-2.1.1-1. 4-by-2.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. FDD)   Source 1 Source 3 Source 4 Source 5 Source 15 Source 18 PUCCH bandwidth (RB/10 MHz) 6 4 4 4 4 6 Channel estimation Real Real Real Real Real Real Receiver type MMSE MMSE with IRC MMSE with IRC MMSE MMSE with IRC MMSE with IRC Table 10.Release 9 26 3GPP TR 36.77 2.1.0 (2010-03) Table 10. and 10. and 1-by-2/1-by-4 SIMO. As performance references.1. and 4-by-4 MUMIMO schemes with eNB antenna configuration (C).1.110 0.1.1.1-2.1. 10.1. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.1-1 Performance of DL MU-MIMO 2 x 2 (C) (3GPP Case1. 10.070 Source 1 Source 5 Source 15 Rel-8 SU-MIMO (2 x 2. TDD)   Source 1 Source 2 Source 3 Source 4 Source 5 Source 10 PUCCH bandwidth (RB/10 MHz) 4 4 4 4 4 4 Channel estimation Real Real Real Real Real Real Receiver type MMSE MMSE with IRC MMSE with IRC MMSE with IRC MMSE with IRC MMSE with IRC Tables in the following subsections capture the spectrum efficiency results from individual sources. L=3) 3GPP target 3GPP . Results show that the MU-MIMO schemes satisfy the 3GPP target and provide significant gain compared with the Rel-8 SU-MIMO scheme. the 3GPP targets and the averaged results of Rel-8 schemes (2-by-2/4-by-2/4-by-4 SU-MIMO with L=3 for downlink.070 0.079 0. Table 10.1.1 10.1.1 3GPP Case1 (3GPP spatial channel model) FDD.

FDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.223 0.Release 9 27 3GPP TR 36.1.53 2.1-6 show the spectrum efficiency results of 2-by-2.63 4.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.182 0.1.41 3.1.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.1.1.088 0.1-5.114 0.070 Source 3 Source 15 Rel-8 SU-MIMO (2 x 2.45 3.42 4.1-4 Performance of DL CS/CB-CoMP 2 x 2 (C) (3GPP Case1.127 0.196 0.54 2.23 2. and 10. L=3) 3GPP target Tables 10.49 4.210 0.070 0. 4-by-2.205 0.1-3 Performance of DL MU-MIMO 4 x 4 (C) (3GPP Case1.100 0.104 0.1-2 Performance of DL MU-MIMO 4 x 2 (C) (3GPP Case1.1.1.58 2.42 2.83 3.118 0. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 5.1. Results show that the CS/CB schemes satisfy the 3GPP target and achieve significant gain compared with the Rel-8 SU-MIMO scheme. Table 10.1. 10.120 Source 1 Source 4 Source 5 Source 11 Source 15 Rel-8 SU-MIMO (4 x 4.143 0.079 0.15 3.55 3. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.123 0.120 0. and 4-by-4 CS/CBCoMP schemes with eNB antenna configuration (C).23 3.814 V9.1-4.45 4. L=3) 3GPP target 3GPP .41 3.1. L=3) 3GPP target Table 10.1.090 Source 1 Source 4 Source 5 Source 7 Source 11 Source 15 Rel-8 SU-MIMO (4 x 2.0.1.0 (2010-03) Table 10.

Table 10.1-7.120 0.1. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 4. Results show that the JP-CoMP schemes provide additional performance enhancement over MU-MIMO in Table 10. L=3) 3GPP target 3GPP .090 Source 1 Source 15 Rel-8 SU-MIMO (4 x 2.0 (2010-03) Table 10.23 2.92 2.1-8 Performance of DL JP-CoMP 4 x 2 (C) (3GPP Case1.1.1.1-7 Performance of DL JP-CoMP 2 x 2 (C) (3GPP Case1. and 4-by-4 JPCoMP schemes with eNB antenna configuration (C).53 2.121 0.1.1.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.28 3.Release 9 28 3GPP TR 36.100 0.34 2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 3. 4-by-2.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.40 3.1-1-2.24 2. and 10.1-9 show the spectrum efficiency results of 2-by-2.53 2. 10. and Table 10.1-1-1.51 3.143 0.1.34 3.1.180 0.1-6 Performance of DL CS/CB-CoMP 4 x 4 (C) (3GPP Case1.090 Source 1 Source 15 Rel-8 SU-MIMO (2 x 2. L=3) 3GPP target Tables 10. L=3) 3GPP target Table 10.100 0.1.1. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 4.150 0.41 3.814 V9.1-1-3 (compared with the same source).144 0.1.1.0.1-5 Performance of DL CS/CB-CoMP 4 x 2 (C) (3GPP Case1.143 0. L=3) 3GPP target Table 10.1.079 0.120 Source 3 Source 15 Source 18 Rel-8 SU-MIMO (4 x 4.180 0.1-8.1.44 3.090 Source 1 Source 3 Source 15 Source 18 Rel-8 SU-MIMO (4 x 2.97 4.1.088 0.1.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.100 0. 10.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.239 0.58 4.195 0.1.1.47 2.

1.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.173 0.2 TDD.1.48 3.1-9 Performance of DL JP-CoMP 4 x 4 (C) (3GPP Case1. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 5. and 4-by-4 MUMIMO schemes with eNB antenna configuration (C).1. Table 10.1.227 0.1.1.0 (2010-03) Table 10.1. L=3) 3GPP target Table 10.143 0.52 3.16 3.2-2. L=3) 3GPP target 10.0.1.1.1.1.096 0.2-4 shows the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna configuration (E). and 10.121 0. 10.2-1 Performance of DL MU-MIMO 2 x 2 (C) (3GPP Case1.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.1.083 0.2-3 show the spectrum efficiency results of 2-by-2. Downlink Tables 10.090 Source 1 Source 2 Source 4 Source 5 Source 10 Source 19 Rel-8 SU-MIMO (4 x 2.070 Source 1 Source 5 Rel-8 SU-MIMO (2 x 2.52 2.2-2 Performance of DL MU-MIMO 4 x 2 (C) (3GPP Case1.1. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.814 V9.17 2.Release 9 29 3GPP TR 36.1.1.68 3.45 4.41 3.80 2.106 0.150 0.1. Results show that the MU-MIMO schemes satisfy the 3GPP target and provide significant gain compared with the Rel-8 SU-MIMO scheme.96 2.5 2. Table 10.23 3. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 4.172 0.120 Source 1 Source 15 Rel-8 SU-MIMO (4 x 4. 4-by-2.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.31 0.2-1.131 0.89 4.160 0. L=3) 3GPP target 3GPP .120 0.

160 0.814 V9.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.16 4.1.2-4 Performance of DL MU-MIMO 4 x 2 (E) (3GPP Case1.194 0.70 3.1. Results show that the CS/CB schemes satisfy the 3GPP target and achieve significant gain compared with the Rel-8 SU-MIMO scheme.2-3 Performance of DL MU-MIMO 4 x 4 (C) (3GPP Case1. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.1.1. and 10.250 0.1.120 Source 3 Rel-8 SU-MIMO (4 x 4.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2-6.088 0.070 Source 3 Rel-8 SU-MIMO (2 x 2.2-7 show the spectrum efficiency results of 2-by-2.096 0.1.1.1.38 2.2-5.1.06 3.090 Source 3 Rel-8 SU-MIMO (4 x 2.1.154 0.52 2. 4-by-2.150 0.154 0.07 4.42 2. Table 10. L=3) 3GPP target Table 10. 10.090 Source 10 Rel-8 SU-MIMO (4 x 2. L=3) 3GPP target Table 10.67 4.073 0.1.1.120 Source 1 Source 2 Source 5 Source 19 Rel-8 SU-MIMO (4 x 4.2-5 Performance of DL CS/CB-CoMP 2 x 2 (C) (3GPP Case1.0. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.2-7 Performance of DL CS/CB-CoMP 4 x 4 (C) (3GPP Case1. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 5. L=3) 3GPP target Table 10. L=3) 3GPP target Tables 10.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.244 0. L=3) 3GPP target 3GPP .35 2.1.1. and 4-by-4 CS/CBCoMP schemes with eNB antenna configuration (C).28 3. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.146 0.2-6 Performance of DL CS/CB-CoMP 4 x 2 (C) (3GPP Case1.0 (2010-03) Table 10. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 6.1.17 2.182 0.Release 9 30 3GPP TR 36.083 0.1.54 2.28 3.

1. 10.2-3.1.090 Source 1 Source 10 Rel-8 SU-MIMO (4 x 2.Release 9 31 3GPP TR 36.1.175 0.1. 4-by-2. Uplink Tables 10.1. L=3) 3GPP target Table 10.1.1.2-4 (compared with the same source).61 3. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 4.1.1.2-11 shows the spectrum efficiency results of 4-by-2 JP-CoMP schemes with eNB antenna configuration (E). and 10.3-1 and 10.1. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.2-10 Performance of DL JP-CoMP 4 x 4 (C) (3GPP Case1.42 2.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2-9.1.1.083 0.1.1.28 3. 3GPP .1.070 Source 1 Rel-8 SU-MIMO (2 x 2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 6.096 0. L=3) 3GPP target 10.130 0.1.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.1.1. and 10.154 0.330 0.1.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 4. Results show that the 2-by-4 SU-MIMO schemes satisfy the 3GPP target and provide significant gain compared with the Rel-8 SIMO scheme. Results show that the JP-CoMP schemes provide additional performance enhancement over MU-MIMO in Table 10.1.0 (2010-03) Tables 10.15 2. respectively.2-1.2-8.0. L=3) 3GPP target Table 10.210 0. L=3) 3GPP target Table 10.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. Table 10.1.1. Table 10.1.1.17 2.1.05 2.52 2.89 4.2-8 Performance of DL JP-CoMP 2 x 2 (C) (3GPP Case1.2-2.3 FDD.38 2.1. 10.1.1.073 0.814 V9.1. and 4-by-4 JPCoMP schemes with eNB antenna configuration (C).2-10 show the spectrum efficiency results of 2-by-2.3-2 show the spectrum efficiency results of 2-by-4 SU-MIMO with eNB antenna configuration (A) and (C).2-9 Performance of DL JP-CoMP 4 x 2 (C) (3GPP Case1.2-11 Performance of DL JP-CoMP 4 x 2 (E) (3GPP Case1.188 0.1.120 Source 1 Rel-8 SU-MIMO (4 x 4. 10.090 Source 10 Rel-8 SU-MIMO (4 x 2.

075 0.44 1.42 1.1.34 2.12 2.36 1.01 2.1.052 0.1.26 2.3-4 Performance of UL CoMP 1 x 2 (C) (3GPP Case1.096 0.1.070 Source 3 Source 15 Source 18 Rel-8 SIMO (1 x 4) 3GPP target Table 10.0.Release 9 32 3GPP TR 36.070 0.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 1.047 0.1.1.3-2 Performance of UL SU-MIMO 2 x 4 (C) (3GPP Case1.3-6 show spectrum efficiency results of 2-by-4 CoMP with eNB antenna configuration (A) and (C).051 0.087 0.3-4 show the spectrum efficiency results of 1-by-2 CoMP with eNB antenna configuration (A) and (C).814 V9. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.37 2.1. Table 10.33 1.040 Source 18 Rel-8 SIMO (1 x 2) 3GPP target 3GPP .46 2.040 Source 18 Rel-8 SIMO (1 x 2) 3GPP target Table 10.0 (2010-03) Table 10.112 0.093 0.1.080 0.050 0.00 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.1.3-3 and 10.1. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.43 2.1.1.40 1.3-5 and 10.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 1.95 2.3-3 Performance of UL CoMP 1 x 2 (A) (3GPP Case1.1.086 0. Results show that multi point reception provides additional performance enhancement compared with the Rel-8 SIMO scheme.1.1.1.33 2.070 Source 1 Source 3 Source 4 Source 5 Source 15 Source 18 Rel-8 SIMO (1 x 4) 3GPP target Tables 10.075 0. Tables 10. respectively. Results show that the Rel-8 SIMO scheme satisfies the 3GPP target.3-1 Performance of UL SU-MIMO 2 x 4 (A) (3GPP Case1.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. respectively.096 0.118 0.

Release 9 33 3GPP TR 36.075 0.4-2 show the spectrum efficiency results of 2-by-4 SU-MIMO with eNB antenna configuration (A) and (C).107 0.1.4-1 and 10.23 2.00 2.4-1 Performance of UL SU-MIMO 2 x 4 (A) (3GPP Case1. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.070 0.080 0. Results show that the Rel-8 SIMO scheme satisfies the 3GPP target and multi point reception provides performance enhancement compared with the Rel-8 SIMO scheme.3-5 Performance of UL CoMP 2 x 4 (A) (3GPP Case1.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.33 2.092 0.1.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.1.092 0.1.4 TDD.070 Source 18 Rel-8 SIMO (1 x 4) 3GPP target 10.1.1.1. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.080 0.4-2 Performance of UL SU-MIMO 2 x 4 (C) (3GPP Case1.1.3-6 Performance of UL CoMP 2 x 4 (C) (3GPP Case1.1.070 Source 3 Source 5 Rel-8 MU-MIMO (1 x 4) 3GPP target Table 10.814 V9.15 2.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.4-3 shows the spectrum efficiency results of 1-by-2 CoMP with eNB antenna configuration (C). FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.1.1.064 0.09 2.83 2. Uplink Tables 10.082 0.0.1.13 1.1.070 Source 18 Rel-8 SIMO (1 x 4) 3GPP target Table 10. Results show that the 2-by-4 SU-MIMO schemes satisfy the 3GPP target and provide gain compared with the Rel-8 SIMO and Rel-8 MU-MIMO schemes. 3GPP .1.95 2.1.075 0.070 Source 1 Source 3 Source 4 Rel-8 SIMO (1 x 4) 3GPP target Table 10.101 0.1. Table 10.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.18 2.0 (2010-03) Table 10. respectively.43 1.16 2.

4-4 shows the spectrum efficiency results of 2-by-4 MU-MIMO with eNB antenna configuration (C). and uplink SU-MIMO.2-1.814 V9. 3. downlink TDD. i. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 1.1.83 2.4-3 Performance of UL CoMP 1 x 2 (C) (3GPP Case1. The tables in the following subsections show the simulation results focusing on LTE-Advanced configurations.1.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.1.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.24 1.51 1.070 Source 2 Rel-8 SIMO (1 x 4) 3GPP target 10.2 Cell spectral efficiency and cell-edge spectral efficiency for ITU-R requirements Cell spectral efficiency and cell-edge spectral efficiency are evaluated through extensive simulations conducted by a number of companies. and 4 show the source specific simulation parameters that characterize the performances of downlink FDD.4 and [5]. downlink and uplink CoMP. Table 10.045 0.Release 9 34 3GPP TR 36.064 0. More detailed assumption on each component is shown in [5]. however.0. downlink and uplink MU-MIMO. Tables 10..1.59 1. The simulation assumptions applied in the following evaluations are shown in Annex A. 3GPP . Detailed information covering a range of possible configurations is provided in Annex A. respectively. that the performance differences among sources could also be explained by other factors such as detailed signal processing algorithms at the transmitter and the receiver.0 (2010-03) Table 10. Note. uplink FDD.3. Result shows that MU-MIMO provides performance enhancement compared with the Rel-8 SIMO scheme.079 0.040 Source 10 Rel-8 SIMO (1 x 2) 3GPP target Table 10. and uplink TDD.1. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.e.4-4 Performance of UL MU-MIMO 2 x 4 (C) (3GPP Case1.051 0.1. 2.

2-1 Simulation parameters (DL.Release 9 35 3GPP TR 36.2-2 Simulation parameters (DL.814 V9. FDD) CCH duration (L symbols) Source 1 Source 2 Source 3 Source 4 Source 5 Source 7 Source 8 3 2 3 3 3 3 3 (Normal subframe) 2 (MBSFN subframe) 3 2 3 (Normal subframe) 2 (MBSFN subframe) 3 3 3 2 2 Num of MBSFN subframe 0 0 6 0 6 0 6 Channel estimation Real Real (Ideal for JPCoMP) Real Real Real Ideal Real Receiver type MMSE MMSE MMSE with IRC MMSE MMSE MMSE MMSE CSI assumption at eNB Short-term Short-term Long-term Short-term Short-term No Short-term Source 9 Source 10 Source 11 0 0 6 Real Real Real MMSE with IRC MMSE with IRC MMSE with IRC Short-term Short-term Short-term and Long-term Short-term Long-term Long-term No Long-term Source 12 Source 13 Source 15 Source 16 Source 18 6 0 6 0 6 Real Real Real Ideal Real MMSE with IRC MMSE with IRC MMSE with IRC MMSE MMSE with IRC Table 10.0. TDD) CCH duration (L symbols) Source 1 Source 2 Source 3 Source 4 Source 7 Source 8 2 2 3 2 3 3 (Normal subframe) 2 (MBSFN subframe) 2 2 3 3 Num of MBSFN subframe 0 0 2 0 0 2 Channel estimation Real Real (Ideal for JPCoMP) Real Real Real Real Receiver type MMSE MMSE MMSE with IRC MMSE MMSE MMSE CSI assumption at eNB Short-term Short-term Long-term Short-term No Short-term Source 9 Source 10 Source 12 Source 13 0 0 2 0 Real Real Real Real MMSE with IRC MMSE with IRC MMSE with IRC MMSE with IRC Short-term Short-term Short-term Long-term 3GPP .0 (2010-03) Table 10.

3GPP . Results show that the MU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8.1. FDD) PUCCH bandwidth (RB/10 MHz) 6 4 4 4 4 4 4 4 4 4 6 3 (InH) 4 (Others) 4 Channel estimation Real Real Real Real Real Ideal Real Real Real Real Real Real Real Receiver type MMSE MMSE MMSE with IRC MMSE MMSE MMSE MMSE MMSE with IRC MMSE with IRC MMSE with IRC MMSE with IRC MMSE with IRC and MMSE with SIC for rank 2 MMSE with IRC Source 1 Source 2 Source 3 Source 4 Source 5 Source 7 Source 8 Source 9 Source 10 Source 11 Source 12 Source 14 Source 18 Table 10.1. Downlink (InH) Tables 10.0 (2010-03) Table 10.Release 9 36 3GPP TR 36.2-3 Simulation parameters (UL. the ITU-R requirements and the averaged results of Rel-8 schemes (4-by-2 SU-MIMO with L=3 for downlink. respectively. 10. TDD) PUCCH bandwidth (RB/10 MHz) 4 4 4 4 4 4 4 3 Channel estimation Real Real Real Real Real Real Real Real Receiver type MMSE MMSE MMSE with IRC MMSE MMSE MMSE with IRC MMSE with IRC MMSE with IRC Source 1 Source 2 Source 3 Source 4 Source 8 Source 9 Source 10 Source 12 Tables in the following subsections capture the spectrum efficiency results from individual sources. As performance references.2.1-1 and 10.0.1. and 1-by-4 SIMO for uplink) assuming real channel estimation.2-4 Simulation parameters (UL.1 Indoor (InH channel model) FDD.2.1 10.2.1-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna configuration (C) and (A).2.814 V9.

FDD) Cell spectral efficiency (bit/sec/Hz/cell) Source 8 Rel-8 SU-MIMO (4 x 2.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.00 3.100 Source 2 Source 8 Source 9 Source 10 Rel-8 SU-MIMO (4 x 2.100 Table 10.186 0.2.1.200 0.2.2.80 5. Results show that increasing the number of eNB antennas contributes to additional performance enhancement in cell spectral efficiency. Downlink (InH) Table 10. Table 10.91 3. Table 10.2-1 Performance of DL MU-MIMO 4 x 2 (C) (InH.60 4.66 4.1-3 shows the spectrum efficiency results of 8-by-2 MU-MIMO scheme with the eNB antenna configuration (C/E).68 5.2.224 0. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8.14 3.232 0.39 4.1.2 TDD.1-1 Performance of DL MU-MIMO 4 x 2 (C) (InH.2-1 shows the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna configuration (C). L=3) ITU-R requirement 5.90 3.0.203 0.198 0. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 5.211 0.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.221 0.220 0.1-2 Performance of DL MU-MIMO 4 x 2 (A) (InH.1.03 5. L=3) ITU-R requirement 10.2.1-3 Performance of DL MU-MIMO 8 x 2 (C/E) (InH.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.200 0.180 0.100 Source 2 Source 8 Source 15 Rel-8 SU-MIMO (4 x 2.211 0.39 5.Release 9 37 3GPP TR 36.1. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 5.2.814 V9.210 0. L=3) ITU-R requirement 3GPP .63 5.1.2.1. L=3) ITU-R requirement Table 10.00 3.100 Source 8 Rel-8 SU-MIMO (4 x 2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 5.0 (2010-03) Table 10.1.

3-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration (A).2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.1.91 4. Results show that the SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 1-by-4 SU-MIMO scheme in Rel-8.2. Results show that multi point reception provides performance enhancement compared with the Rel-8 1-by-4 SIMO scheme. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 5.65 3.400 0. Table 10.226 0.239 0.198 0. Table 10.25 2.211 0.3 FDD.42 3.0 (2010-03) Table 10.3-1 Performance of UL SU-MIMO 2 x 4 (A) (InH.3-2 shows the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna configuration (A).32 3.2.2-2 Performance of DL MU-MIMO 8 x 2 (C/E) (InH. 3GPP .070 Source 1 Source 18 Rel-8 SIMO (1 x 4) ITU-R requirement Table 10.1.2-2 shows the spectrum efficiency results of 8-by-2 MU-MIMO scheme with the eNB antenna configuration (C/E).01 4.90 3.274 0.0.2.265 0.25 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2.271 0. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 4.39 5.31 3.3-2 Performance of UL CoMP 1 x 4 (A) (InH.1.Release 9 38 3GPP TR 36.18 3.276 0.1.139 0. Uplink (InH) Table 10.070 Source 1 Source 3 Source 5 Source 12 Source 14 Source 18 Rel-8 SIMO (1 x 4) ITU-R requirement Table 10. Results show that increasing the number of eNB antennas enhances cell spectral efficiency. Results show that additional transmit antenna provides further performance enhancement.1. Table 10.1.226 0.25 2.2.273 0.814 V9.2.1.00 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. L=3) ITU-R requirement 10.2.1.84 4.100 Source 8 Rel-8 SU-MIMO (4 x 2.25 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. Use of advanced receiver such as IRC and SIC provides additional gains.3-3 shows the spectrum efficiency results of 2-by-4 CoMP schemes with eNB antenna configuration (A).

12 4.4-2 shows the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna configuration (A).2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 4.1. Table 10.4 TDD.2. Results show that multi point reception provides performance enhancement compared with the Rel-8 1-by-4 SIMO scheme.2.2.25 2.1.1. Results show that the 2-by-4 SU-MIMO schemes satisfy the ITU-R requirements.43 3.25 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.0.4-3 Performance of UL CoMP 2 x 4 (A) (InH.1.3-3 Performance of UL CoMP 2 x 4 (A) (InH.78 3.98 4.04 2. Table 10. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.04 2.124 0.206 0.25 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2.Release 9 39 3GPP TR 36.226 0.2.04 2.070 Source 1 Rel-8 SIMO (1 x 4) ITU-R requirement Table 10.4-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration (A). Results show that additional transmit antenna provides further performance enhancement in cell spectral efficiency.274 0.25 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.1.1. Table 10.0 (2010-03) Table 10.260 0.240 0. Performance improvement from the Rel-8 1-by-4 SIMO scheme is also demonstrated.070 Source 1 Rel-8 SIMO (1 x 4) ITU-R requirement 3GPP .4-2 Performance of UL CoMP 1 x 4 (A) (InH.4-1 Performance of UL SU-MIMO 2 x 4 (A) (InH.2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.278 0.84 3.243 0.070 Source 1 Source 18 Rel-8 SIMO (1 x 4) ITU-R requirement 10.250 0.1.070 Source 1 Source 3 Source 12 Rel-8 SIMO (1 x 4) ITU-R requirement Table 10.1. Use of advanced receiver such as IRC and SIC provides additional gains. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3. Uplink (InH) Table 10.206 0.206 0.4-3 shows the spectrum efficiency results of 2-by-4 CoMP schemes with eNB antenna configuration (A).2.41 3.814 V9.74 3.25 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.

2.103 0.61 2.86 3.079 0.15 2.74 2.1-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna configuration (C) and (A).100 0.061 0.084 0.17 2.2 10.078 0.51 2.1-3 shows the spectrum efficiency results of 8-by-2 MU-MIMO schemes with the eNB antenna configuration (C/E).1-2 Performance of DL MU-MIMO 4 x 2 (A) (UMi. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8 that cannot meet the requirement.14 2.105 0.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.075 Source 5 Source 8 Source 12 Rel-8 SU-MIMO (4 x 2.1 Microcellular (UMi channel model) FDD. Use of MBSFN subframes (overhead reduction).1-1 and 10.2. Table 10.2. Obviously. increasing the number of eNB antennas contributes to further performance enhancement.97 2.2.1-1 Performance of DL MU-MIMO 4 x 2 (C) (UMi.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2.65 1. respectively.083 0.18 2.2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.063 0.103 0. Downlink (UMi) Tables 10.2. L=3) ITU requirement Table 10.081 0. 3GPP .14 2.95 2.076 0.96 2.and celledge spectrum efficiency performances.2.2.068 0.2.88 3.075 Source 1 Source 2 Source 4 Source 5 Source 7 Source 8 Source 11 Source 12 Source 15 Rel-8 SU-MIMO (4 x 2.2.0 (2010-03) 10.2. IRC receiver. and short-term feedback tends to improve the cell. L=3) ITU requirement Table 10.814 V9.2.0.Release 9 40 3GPP TR 36. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.093 0.

2.2.11 2. Table 10.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.1-5 show the spectrum efficiency results of 4-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C) and (A). L=3) ITU requirement Table 10.84 3.075 Source 1 Source 5 Source 8 Source 10 Rel-8 SU-MIMO (4 x 2. Table 10.068 0.62 3.2.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.1-4 Performance of DL CS/CB-CoMP 4 x 2 (C) (UMi.14 2.1-3 Performance of DL MU-MIMO 8 x 2 (C/E) (UMi.2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.814 V9.2.084 0. L=3) ITU requirement Table 10.2.14 2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) Source 12 Rel-8 SU-MIMO (4 x 2.96 2. respectively.075 Source 3 Source 11 Source 12 Source 13 Source 18 Rel-8 SU-MIMO (4 x 2.138 0.99 3. Results show that the CS/CB schemes satisfy the ITU requirements and achieve significant gain compared with the Rel-8 4-by-2 SU-MIMO scheme. L=3) ITU requirement 2.2.14 2.105 0.140 0.0 (2010-03) Table 10.1-4 and 10.088 0.2.075 Source 3 Source 4 Rel-8 SU-MIMO (4 x 2. Results show that the JP-CoMP schemes provide additional performance enhancement over MU-MIMO in Table 10.2.2.086 0.72 3.2.99 3.089 0.0.114 0.075 Table 10. L=3) ITU requirement Tables 10.1-5 Performance of DL CS/CB-CoMP 4 x 2 (A) (UMi.2.2.1-7 shows the spectrum efficiency results of 4-by-2 JP-CoMP schemes with eNB antenna configuration (C).130 0. 3GPP .05 2.1-6 shows the spectrum efficiency results of 8-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C/E).2.092 0.15 2.36 2.083 0.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2.068 0.061 0.03 3.71 1.2.2.103 0.1-6 Performance of DL CS/CB-CoMP 8 x 2 (C) (UMi. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.Release 9 41 3GPP TR 36.21 3.2-1-1 (compared with the same source). FDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.068 0.

140 0.65 1.96 2.2-1 and 10. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 3. IRC receiver.14 2.1-7 Performance of DL JP-CoMP 4 x 2 (C) (UMi.0 (2010-03) Table 10. Table 10.2.061 0.2. and short-term feedback tends to improve the cell. L=3) ITU requirement 10.2-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna configuration (C) and (A).2.2. L=3) ITU requirement Table 10. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.Release 9 42 3GPP TR 36.075 Source 1 Source 2 Rel-8 SU-MIMO (4 x 2.06 2.2.1-9 show the spectrum efficiency results of 8-by-2 SU-MIMO scheme with eNB antenna configuration (A) and (C).60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2.1-8 and 10.2.075 Source 16 Rel-8 SU-MIMO (4 x 2.2.068 0.153 0. 3GPP .14 2.2.45 3.2 TDD.74 2.2. Use of MBSFN subframes (overhead reduction). Downlink (UMi) Tables 10. resulting in enhanced performance.2. Results show that use of 8 transmit antenna gives additional degrees of freedom.2.075 Source 16 Rel-8 SU-MIMO (4 x 2.097 0.2.1-9 Performance of DL SU-MIMO 8 x 2 (C) (UMi.2. respectively.108 0.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8 that cannot meet the requirement.1-8 Performance of DL SU-MIMO 8 x 2 (A) (UMi.0. respectively.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.and celledge spectrum efficiency performances.2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.068 0. L=3) ITU requirement Tables 10.2.814 V9.

33 2.0 (2010-03) Table 10.103 0.65 1.084 0. Obviously.2-3 shows the spectrum efficiency results of 8-by-2 MU-MIMO schemes with the eNB antenna configuration (C/E).2.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2.33 3.07 3.2.0.2-5 show the spectrum efficiency results of 4-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C). Table 10.20 2.Release 9 43 3GPP TR 36. increasing the number of eNB antennas contributes to further performance enhancement.105 0.076 0.99 2.2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.090 0.092 0.2.2-1 Performance of DL MU-MIMO 4 x 2 (C) (UMi. Table 10. 3GPP .067 0.85 3.2.814 V9.084 0.47 2.2.20 2.099 0.46 3. respectively.087 0.2.22 2.1-6 shows the spectrum efficiency results of 4-by-2 JP-CoMP schemes with eNB antenna configuration (C).106 0.095 0.2-3 Performance of DL MU-MIMO 8 x 2 (C/E) (UMi.07 3.2.71 3.2.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.076 0. L=3) ITU-R requirement Table 10. Results show that the CoMP schemes satisfy the ITU requirements and achieve significant gain compared with the Rel8 4-by-2 SU-MIMO scheme.85 3. L=3) ITU-R requirement Table 10.096 0.075 Source 1 Source 2 Source 9 Source 10 Rel-8 SU-MIMO (4 x 2.94 2.076 0.2.2.2. L=3) ITU-R requirement Tables 10.075 Source 1 Source 2 Source 4 Source 7 Source 8 Source 9 Source 10 Source 12 Rel-8 SU-MIMO (4 x 2.2. and 8-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C).2-4 and 10.2-2 Performance of DL MU-MIMO 4 x 2 (A) (UMi.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.089 0.90 2.075 Source 12 Rel-8 SU-MIMO (4 x 2.

2.814 V9.0.082 0.20 2.096 0.088 0.2.0 (2010-03) Table 10.76 3.112 0.2.087 0.2. L=3) ITU-R requirement Table 10.060 Source 1 Source 2 Rel-8 SU-MIMO (4 x 2.20 2. Use of advanced receiver such as IRC and SIC provides additional gains.050 Source 1 Source 3 Source 5 Source 11 Source 12 Source 14 Source 18 Rel-8 SIMO (1 x 4) ITU-R requirement 3GPP .08 1.2.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.20 2.2-6 Performance of DL JP-CoMP 4 x 2 (C) (UMi.092 0.3-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration (A).075 Source 2 Source 4 Rel-8 SU-MIMO (4 x 2. L=3) ITU-R requirement Table 10. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.83 3.60 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2.45 2.076 0.93 1.096 0.23 2.2.051 0.98 2.075 0. L=3) ITU-R requirement 10.28 2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.075 Source 2 Source 3 Source 13 Rel-8 SU-MIMO (4 x 2.2-4 Performance of DL CS/CB-CoMP 4 x 2 (C) (UMi.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.076 0.2. Results show that the SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 1-by-4 SU-MIMO scheme in Rel-8.2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.071 0.78 3. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3. Table 10.24 1.20 2.2.076 0.93 1.3 FDD.03 2. Uplink (UMi) Table 10.087 0.074 0.089 0.2-5 Performance of DL CS/CB-CoMP 8 x 2 (C) (UMi.3-1 Performance of UL SU-MIMO 2 x 4 (A) (UMi.Release 9 44 3GPP TR 36.80 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.096 0.20 2.2.52 2.2.083 0.

3-4 Performance of UL MU-MIMO 2 x 4 (A) (UMi.2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2. Results show that MU-MIMO provides performance enhancement compared with SU-MIMO shown in Table 10.80 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. 3GPP .3-4 shows the spectrum efficiency results of 2-by-4 MU-MIMO schemes with eNB antenna configuration (A).093 0.050 Source 1 Source 18 Rel-8 SIMO (1 x 4) ITU-R requirement Table 10. Table 10.22 2.2.2.2.086 0.3-1 (compared with the same source).050 Source 1 Source 18 Rel-8 SIMO (1 x 4) ITU-R requirement Table 10.93 1.Release 9 45 3GPP TR 36.93 1.2.80 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2.51 1.2.80 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.089 0.102 0.3-3 Performance of UL CoMP 2 x 4 (A) (UMi.93 1.2.18 1.3-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna configuration (A) and 2-by-4 CoMP schemes with eNB antenna configuration (A).074 0.2.2.2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.074 0. Uplink (UMi) Table 10.3-2 and 10.2.2.0 (2010-03) Tables 10. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.074 0.2. Results show that multi point reception provides performance enhancement compared with the Rel-8 1-by-4 SIMO scheme.2.102 0.0.4 TDD.4-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration (A).03 1.050 Source 12 Rel-8 SIMO (1 x 4) ITU-R requirement 10.3-2 Performance of UL CoMP 1 x 4 (A) (UMi. respectively. Table 10.2.2.45 2. Results show that the SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 1-by-4 SU-MIMO scheme in Rel-8.2.814 V9.

79 1.072 0. Results show that MU-MIMO provides performance enhancement compared with SU-MIMO shown in Table 10. Results show that additional eNB receive antennas contributes to additional performance enhancement. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.2.081 0.2.083 0.90 1.2.2.072 0.87 1. respectively.35 1.4-4 shows the spectrum efficiency results of 2-by-4 MU-MIMO schemes with eNB antenna configuration (A).2.2.87 1.2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) Source 1 Rel-8 SIMO (1 x 4) ITU-R requirement 2. Table 10.068 0. Table 10.15 1.12 2.050 Table 10.4-1 (compared with the same source).072 0.0 (2010-03) Table 10.4-2 and 10.050 Source 1 Source 3 Source 12 Rel-8 SIMO (1 x 4) ITU-R requirement Tables 10.2.4-4 Performance of UL MU-MIMO 2 x 4 (A) (UMi.058 0.80 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.0.4-2 Performance of UL CoMP 1 x 4 (A) (UMi.2.2.4-5 shows the spectrum efficiency results of 1-by-8 MU-MIMO schemes with eNB antenna configuration (C/E).100 0.87 1.80 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2.4-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna configuration (A) and 2-by-4 CoMP schemes with eNB antenna configuration (A).80 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) Source 12 Rel-8 SIMO (1 x 4) ITU-R requirement 2.050 Table 10.097 0. 3GPP .4-3 Performance of UL CoMP 2 x 4 (A) (UMi.072 0.2.4-1 Performance of UL SU-MIMO 2 x 4 (A) (UMi.80 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.050 Source 1 Rel-8 SIMO (1 x 4) ITU-R requirement Table 10.2.814 V9.Release 9 46 3GPP TR 36.2. Results show that multi point reception provides performance enhancement compared with the Rel-8 1-by-4 SIMO scheme.29 1. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.2.2.87 1.2.

Release 9

47

3GPP TR 36.814 V9.0.0 (2010-03)

Table 10.2.2.4-5 Performance of UL MU-MIMO 1 x 8 (C/E) (UMi, TDD)
Cell spectral efficiency (bit/sec/Hz/cell) Source 2 Rel-8 SIMO (1 x 4) ITU-R requirement 2.97 1.82 1.80 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.079 0.067 0.050

10.2.3
10.2.3.1

Base coverage urban (UMa channel model)
FDD, Downlink (UMa)

Tables 10.2.3.1-1 and 10.2.3.1-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna configuration (C) and (A), respectively. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8 that cannot meet the requirement. Use of MBSFN subframes (overhead reduction), IRC receiver, and short-term feedback tends to improve the cell- and celledge spectrum efficiency performances. Table 10.2.3.1-1 Performance of DL MU-MIMO 4 x 2 (C) (UMa, FDD)
Cell spectral efficiency (bit/sec/Hz/cell) 2.14 2.22 2.06 2.48 2.66 2.36 2.22 1.61 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.056 0.060 0.060 0.063 0.070 0.076 0.062 0.050 0.060

Source 1 Source 5 Source 7 Source 8 Source 11 Source 12 Source 15 Rel-8 SU-MIMO (4 x 2, L=3) ITU-R requirement

Table 10.2.3.1-2 Performance of DL MU-MIMO 4 x 2 (A) (UMa, FDD)
Cell spectral efficiency (bit/sec/Hz/cell) 1.75 2.33 1.45 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.047 0.064 0.043 0.060

Source 1 Source 5 Rel-8 SU-MIMO (4 x 2, L=3) ITU-R requirement

Table 10.2.3.1-3 shows the spectrum efficiency results of 8-by-2 MU-MIMO schemes with the eNB antenna configuration (C/E). Obviously, increasing the number of eNB antennas contributes to further performance enhancement.

3GPP

Release 9

48

3GPP TR 36.814 V9.0.0 (2010-03)

Table 10.2.3.1-3 Performance of DL MU-MIMO 8 x 2 (C/E) (UMa, FDD)
Cell spectral efficiency (bit/sec/Hz/cell) 2.77 2.80 2.55 1.61 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.091 0.104 0.070 0.050 0.060

Source 1 Source 5 Source 10 Rel-8 SU-MIMO (4 x 2, L=3) ITU-R requirement

Tables 10.2.3.1-4 and 10.2.3.1-5 show the spectrum efficiency results of 4-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C) and 8-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C), respectively. Results show that the CS/CB-CoMP schemes satisfy the ITU requirements and achieve significant gain compared with the Rel8 4-by-2 SU-MIMO scheme.

Table 10.2.3.1-4 Performance of DL CS/CB-CoMP 4 x 2 (C) (UMa, FDD)
Cell spectral efficiency (bit/sec/Hz/cell) 2.68 2.38 2.44 2.33 2.30 2.45 1.61 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.067 0.073 0.067 0.083 0.063 0.060 0.050 0.060

Source 2 Source 3 Source 11 Source 12 Source 13 Source 18 Rel-8 SU-MIMO (4 x 2, L=3) ITU-R requirement

Table 10.2.3.1-5 Performance of DL CS/CB-CoMP 8 x 2 (C) (UMa, FDD)
Cell spectral efficiency (bit/sec/Hz/cell) 3.34 2.70 3.22 1.61 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.088 0.085 0.073 0.050 0.060

Source 3 Source 4 Source 13 Rel-8 SU-MIMO (4 x 2, L=3) ITU-R requirement

Tables 10.2.3.1-6 and 10.2.3.1-7 show the spectrum efficiency results of 4-by-2 JP-CoMP schemes with eNB antenna configuration (C) and (A), respectively. Tables show that the JP CoMP schemes provide additional performance enhancement over MU-MIMO in Table 10.2.3.1-1 (compared with the same source).

3GPP

Release 9

49

3GPP TR 36.814 V9.0.0 (2010-03)

Table 10.2.3.1-6 Performance of DL JP-CoMP 4 x 2 (C) (UMa, FDD) 
Cell spectral efficiency (bit/sec/Hz/cell) 2.41 2.32 1.61 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.064 0.069 0.050 0.060

Source 1 Source 2 Rel-8 SU-MIMO (4 x 2, L=3) ITU-R requirement

Table10.2.3.1-7 Performance of DL JP-CoMP 4 x 2 (A) (UMa, FDD)
Cell spectral efficiency (bit/sec/Hz/cell) Source 5 Rel-8 SU-MIMO (4 x 2, L=3) ITU-R requirement 2.41 1.45 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.065 0.043 0.060

Tables 10.2.3.1-8 and 10.2.3.1-9 show the spectrum efficiency results of 8-by-2 SU-MIMO schemes with eNB antenna configuration (A) and (C), respectively. Results show that the use of 8 antenna elements gives additional degrees of freedom, which leads to improved performance. Table 10.2.3.1-8 Performance of DL SU-MIMO 8 x 2 (A) (UMa, FDD)
Cell spectral efficiency (bit/sec/Hz/cell) 3.53 1.45 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.150 0.043 0.060

Source 16 Rel-8 SU-MIMO (4 x 2, L=3) ITU-R requirement

Table 10.2.3.1-9 Performance of DL SU-MIMO 8 x 2 (C) (UMa, FDD)
Cell spectral efficiency (bit/sec/Hz/cell) 3.54 1.61 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.149 0.050 0.060

Source 16 Rel-8 SU-MIMO (4 x 2, L=3) ITU-R requirement

10.2.3.2

TDD, Downlink (UMa)

Tables 10.2.3.2-1 and 10.2.3.2-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna configuration (C) and (A), respectively. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8 that cannot meet the requirements. Use of MBSFN subframes (overhead reduction), IRC receiver, and short-term feedback tends to improve the cell- and celledge spectrum efficiency performances.

3GPP

062 0.2.040 0.89 1.814 V9.3.62 2.34 2.2.062 0.34 2.077 0. respectively. L=3) ITU-R requirement Tables 10.077 0.3.060 Source 1 Source 2 Source 7 Source 8 Source 9 Source 10 Source 12 Rel-8 SU-MIMO (4 x 2.3.2-3 Performance of DL MU-MIMO 8 x 2 (C/E) (UMa.2.069 0. L=3) ITU-R requirement Table 10.92 2. increasing the number of eNB antennas contributes to further performance enhancement. Obviously.2.42 3.2-3 shows the spectrum efficiency results of 8-by-2 MU-MIMO schemes with the eNB antenna configuration (C/E).Release 9 50 3GPP TR 36.0 (2010-03) Table 10.060 Source 1 Rel-8 SU-MIMO (4 x 2.3.43 2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.040 0.2-5 show the spectrum efficiency results of 4-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C) and 8-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C).3.2-1 Performance of DL MU-MIMO 4 x 2 (C) (UMa. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.036 0.32 1.27 2.075 0.2-4 and 10. Results show that the CS/CB-CoMP schemes satisfy the ITU requirements and achieve significant gain compared with the Rel8 4-by-2 SU-MIMO scheme.2-2 Performance of DL MU-MIMO 4 x 2 (A) (UMa.3.36 2.060 Source 1 Source 2 Source 9 Source 10 Rel-8 SU-MIMO (4 x 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.49 2.14 2.056 0. Table 10. 3GPP .2.080 0.85 1.2.069 0.0.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.071 0.63 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 1.062 0.34 2. L=3) ITU-R requirement Table 10.102 0.

20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2. Table 10.2.040 0. Uplink (UMa) Table 10.072 0.068 0.092 0.060 Source 2 Source 4 Source 13 Rel-8 SU-MIMO (4 x 2.040 0.34 1.23 1.060 Source 1 Source 2 Rel-8 SU-MIMO (4 x 2.24 1.2.95 2.2.34 2.093 0. 3GPP .Release 9 51 3GPP TR 36. L=3) ITU-R requirement Table 10.2-6 shows the spectrum efficiency results of 4-by-2 JP-CoMP schemes with eNB antenna configuration (C).74 2.040 0.2-6 Performance of DL JP-CoMP 4 x 2 (C) (UMa.3. L=3) ITU-R requirement Table 10.060 Source 2 Source 3 Source 12 Source 13 Rel-8 SU-MIMO (4 x 2.075 0.2-4 Performance of DL CS/CB-CoMP 4 x 2 (C) (UMa.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3. L=3) ITU-R requirement 10.0 (2010-03) Table 10.34 2.20 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.3.073 0.069 0. Results show that the 2-by-4 SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 1-by-4 SU-MIMO scheme in Rel-8.2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.77 3.34 2.30 2.3.065 0.21 2.3-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration (A). Results show that the JP-CoMP schemes satisfy the ITU requirements and achieve significant gain compared with the Rel-8 4-by-2 SU-MIMO scheme.34 2.2-5 Performance of DL CS/CB-CoMP 8 x 2 (C) (UMa.814 V9.076 0. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.0.3.3.3.3 FDD.2.

3.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.50 1.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.030 Source 1 Rel-8 SIMO (1 x 4) ITU-R requirement Tables 10.060 0.3.3.0.075 0.74 1.3-4 Performance of UL CoMP 2 x 4 (A) (UMa.3.62 1.067 0.2.087 0. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 1.086 0.067 0.046 0.62 2. Results show that additional transmit antenna provides further performance enhancement.814 V9.3-4 and 10.77 1. Table10.3-2 and 10. respectively.102 0.082 0.2.0 (2010-03) Table 10.095 0. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.95 1.3-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna configuration (A) and (C).2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 1.93 1.072 0.072 0.51 1.3-3 Performance of UL CoMP 1 x 4 (C) (UMa.3.2.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 1.3-5 show the spectrum efficiency results of 2-by-4 CoMP schemes with eNB antenna configuration (A) and (C).40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.067 0.Release 9 52 3GPP TR 36.55 1.55 1.3-2 Performance of UL CoMP 1 x 4 (A) (UMa. Table10.3-1 Performance of UL SU-MIMO 2 x 4 (A) (UMa.75 1.030 Source 1 Source 3 Source 5 Source 12 Source 14 Source 18 Rel-8 SIMO (1 x 4) ITU-R requirement Tables 10.2.89 1.096 0.3.2.3.03 1.030 Source 1 Source 18 Rel-8 SIMO (1 x 4) ITU-R requirement 3GPP .030 Source 1 Source 18 Rel-8 SIMO (1 x 4) ITU-R requirement Table10.3.081 0.11 1. Results show that the performance of CoMP schemes exceeds the ITU requirements. respectively.2.55 1.

2.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.76 1.2.2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 1.3.069 0. respectively.55 1. Table 10.2.54 1.076 0.51 1. Results show that the 2-by-4 SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 1-by-4 SU-MIMO scheme in Rel-8.3.4-1 Performance of UL SU-MIMO 2 x 4 (A) (UMa.3.2.030 Source 12 Rel-8 SIMO (1 x 4) ITU-R requirement 10.030 Source 1 Source 3 Source 12 Rel-8 SIMO (1 x 4) ITU-R requirement Tables 10.3.814 V9.3.2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 1.4-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration (A).030 Table 10.51 1.093 0. Table10.91 1.067 0.3-6 shows the spectrum efficiency results of 2-by-4 MU-MIMO schemes with eNB antenna configuration (A).099 0.3-1 (compared with the same source).73 1.4-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna configuration (A) and (C).2.0.89 1.060 0.3-5 Performance of UL CoMP 2 x 4 (C) (UMa.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.3.2.4 TDD. FDD) Cell spectral efficiency (bit/sec/Hz/cell) Source 1 Rel-8 SIMO (1 x 4) ITU-R requirement 2. Results show that multi point reception provides performance enhancement compared with the Rel-8 1-by-4 SIMO scheme.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. Table 10.3-6 Performance of UL MU-MIMO 2 x 4 (A) (UMa.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.50 1.05 1.3.030 Source 1 Rel-8 SIMO (1 x 4) ITU-R requirement 3GPP .3.4-2 and 10.2.4-2 Performance of UL CoMP 1 x 4 (A) (UMa.Release 9 53 3GPP TR 36. Uplink (UMa) Table 10.3.3.2.062 0.062 0. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 1.059 0.083 0. Results show that MU-MIMO provides performance enhancement compared with SU-MIMO shown in Table 10.0 (2010-03) Table 10.

4.2.030 Source 1 Rel-8 SIMO (1 x 4) ITU-R requirement Table 10. TDD) Cell spectral efficiency (bit/sec/Hz/cell) Source 1 Rel-8 SIMO (1 x 4) ITU-R requirement 1. and 8-by-2 MU-MIMO schemes with the eNB antenna configuration (C/E). Results show that additional transmit antenna provides further performance enhancement.097 0.2.1-1 and 10. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2. respectively.86 1.2.4-3 Performance of UL CoMP 1 x 4 (C) (UMa. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 4by-2 SU-MIMO scheme in Rel-8.3.74 1.030 Source 2 Rel-8 SIMO (1 x 4) ITU-R requirement 10.030 Source 1 Rel-8 SIMO (1 x 4) ITU-R requirement Table 10.2.45 1.2.090 0.45 1.3.3.1-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna configuration (C).3. respectively.4 10. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2. Downlink (RMa) Tables 10.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.3.06 1.2.2.2.51 1.4.2.076 0.1 High speed (RMa channel model) FDD.100 0.Release 9 54 3GPP TR 36.2.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.4-6 shows the spectrum efficiency results of 1-by-8 MU-MIMO schemes with eNB antenna configuration (C/E).059 0.062 0.00 1.0.4-5 show the spectrum efficiency results of 2-by-4 CoMP schemes with eNB antenna configuration (A) and (C).059 0.059 0. Table 10.3.4-6 Performance of UL MU-MIMO 1 x 8 (C/E) (UMa.030 Tables 10.3.2.814 V9.0 (2010-03) Table10.4.4-5 Performance of UL CoMP 2 x 4 (C) (UMa.4-4 and 10. Results show that additional eNB receive antennas contributes to further performance enhancement.40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. 3GPP .40 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. Table10.45 1.4-4 Performance of UL CoMP 2 x 4 (A) (UMa.

092 0.4.040 Source 2 Source 8 Source 9 Source 10 Rel-8 SU-MIMO (4 x 2.91 1.1-1 Performance of DL MU-MIMO 4 x 2 (C) (UMa.4.1-2 Performance of DL MU-MIMO 8 x 2 (C/E) (UMa. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 3. L=3) ITU requirement 10. L=3) ITU requirement Table 10.106 0.10 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.053 0.110 0.075 0.2-2 show the spectrum efficiency results of 4-by-2 MU-MIMO schemes with eNB antenna configuration (C).040 Source 3 Source 8 Source 15 Rel-8 SU-MIMO (4 x 2.4.60 1.4.34 2.10 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2-3 shows the spectrum efficiency results of 4-by-2 CS/CB-CoMP schemes with eNB antenna configuration (C).2.26 1.087 0.4.33 1.069 0.040 Source 2 Source 10 Rel-8 SU-MIMO (4 x 2. respectively.4.2 TDD. L=3) ITU requirement Table 10. Results show that the CS/CB-CoMP scheme satisfies the ITU-R requirements and provides significant gain compared with the 4-by-2 SU-MIMO scheme in Rel-8. Results show that the MU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 4by-2 SU-MIMO scheme in Rel-8.0 (2010-03) Table 10.91 1.10 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. 3GPP .94 3.112 0. Table 10.814 V9. L=3) ITU requirement Table 10.10 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.2. and 8-by-2 MU-MIMO schemes with the eNB antenna configuration (C/E). FDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.60 1.2-2 Performance of DL MU-MIMO 8 x 2 (C/E) (RMa.2.45 1.2-1 Performance of DL MU-MIMO 4 x 2 (C) (RMa.86 1.4.053 0.2.083 0.081 0.069 0.36 3.040 Source 10 Rel-8 SU-MIMO (4 x 2.2. Downlink (RMa) Tables 10.87 3.47 2.79 2.Release 9 55 3GPP TR 36.2.2.077 0.0.086 0.2.4.2-1 and 10.

4.4.040 Source 3 Rel-8 SU-MIMO (4 x 2.45 2.2.129 0.4.90 2.062 0.4.75 1.4.2-3 Performance of DL CS/CB-CoMP 4 x 2 (C) (RMa. Results show that the performance of the CoMP schemes far exceed the ITU requirements. Results show that the 2-by-4 SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 1-by-4 SU-MIMO scheme in Rel-8.129 0.80 0.3-1 Performance of UL SU-MIMO 2 x 4 (A) (RMa.2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.32 2. Table 10.46 1.109 0.4.053 0.3-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna configuration (A) and 2-by-4 CoMP schemes with eNB antenna configuration (A).2.2. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.126 0.129 0.60 1. L=3) ITU requirement 10. Table 10.015 Source 1 Source 3 Source 5 Source 12 Source 14 Source 18 Rel-8 SIMO (1 x 4) ITU requirement Tables 10.42 1.086 0.2.2.4.90 1.10 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0. respectively.089 0.015 Source 1 Source 18 Rel-8 SIMO (1 x 4) ITU requirement 3GPP .Release 9 56 3GPP TR 36.2.0 (2010-03) Table 10.814 V9.089 0.3 FDD.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.80 0.3-2 and 10. Uplink (RMa) Table 10.129 0.086 0.3-2 Performance of UL CoMP 1 x 4 (A) (RMa.3-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration (A).0.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.31 1. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 3.27 1.

4-1 shows the spectrum efficiency results of 2-by-4 SU-MIMO schemes with eNB antenna configuration (A).098 0.086 0.Release 9 57 3GPP TR 36.086 0. FDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.2. 3GPP .086 0.4 TDD.4-3 show the spectrum efficiency results of 1-by-4 CoMP schemes with eNB antenna configuration (A) and 2-by-4 CoMP schemes with eNB antenna configuration (A).4.2.3-3 Performance of UL CoMP 2 x 4 (A) (RMa.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.4.015 Table 10.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.089 0.4-4 shows the spectrum efficiency results of 1-by-8 MU-MIMO schemes with eNB antenna configuration (C/E).4.154 0.015 Source 1 Source 18 Rel-8 SIMO (1 x 4) ITU requirement 10. Uplink (RMa) Table 10.2.52 1.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.75 0.75 0.4.25 2.2.814 V9. Table 10.0 (2010-03) Table 10.153 0.80 0.015 Source 1 Source 3 Rel-8 SIMO (1 x 4) ITU requirement Tables 10. Table 10. respectively.2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) Source 1 Rel-8 SIMO (1 x 4) ITU requirement 2. TDD) Cell spectral efficiency (bit/sec/Hz/cell) 2.4. TDD) Cell spectral efficiency (bit/sec/Hz/cell) Source 1 Rel-8 SIMO (1 x 4) ITU requirement 2.4-3 Performance of UL CoMP 2 x 4 (A) (RMa.70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.02 1.23 1.2.4.015 Table 10.4.118 0.2.75 0.2.4.4-1 Performance of UL SU-MIMO 2 x 4 (A) (RMa.4.125 0. Results show that the 2-by-4 SU-MIMO schemes satisfy the ITU-R requirements and provide significant gain compared with the 1-by-4 SU-MIMO scheme in Rel-8.0.58 2.27 1. Results show that multi point reception provides performance enhancement compared with the Rel-8 1-by-4 SIMO scheme.4-2 and 10.2.4-2 Performance of UL CoMP 1 x 4 (A) (RMa. Results show that additional eNB receive antennas contributes to the further performance enhancement.127 0.

1.101 0.1.814 V9. A.Release 9 58 3GPP TR 36.1.1 Link simulation Scenarios Link simulation analyses with (8x8) DL MIMO configuration can be conducted in relation to studies concerning the peak data rate requirements set forth in [36.1 A. operating bandwidth (BW). some of which are reflected in Table A. Inter-site distance (ISD). with the modifications given in Table A.4.2.1.2 A.2.0 (2010-03) Table 10. A.913].1-1 along with additional assumptions related to carrier frequency (CF).70 Cell-edge user spectral efficiency (bit/sec/Hz/user) 0.73 0.2.1. simulation parameters should be aligned with the ITU guidelines in [IMT Eval].1 System simulation Scenarios System simulation assumptions Reference system deployments This section describes the reference system deployments to use for the different system evaluations.2.0. 3GPP .015 Annex A: Simulation model A. TDD) Cell spectral efficiency (bit/sec/Hz/cell) Source 2 Rel-8 SIMO (1 x 4) ITU requirement 2.1-2.64 1.2.2.1-1.EVAL] section-8 defines different antenna horizontal and vertical pattern from those defined in Table A.814].077 0. the system simulation baseline parameters for the macro-cell deployment model are as specified in [TR 25.2.1-2 which is for 3GPP case evaluation only.1 Homogeneous deployments The minimum set of simulation cases is given in Table A.2. For 3GPP cases only.1. penetration loss (PLoss) and UE speed.1.4-4 Performance of UL MU-MIMO 1 x 8 (C/E) (RMa. For the ITU cases. Note that [IMT.

814 Table A.0 (2010-03) Table A. 3GPP .2. Indoor hotspot scenario ITU Microcellular.1-2 and 25.1.1 Speed (km/h) 3 3 3 3 3 3 30 30 120 Additional Simulation Parameters Table A. See [IMT Eval] Annex. Urban macro-cell scenario ITU Base coverage urban extended.0.0 3.1-2 and 25. Rural macro-cell scenario CF (GHz) 2.0 2.0 0.1 See [IMT Eval] Annex.814 Table A.4 2.2.814 [IMT Eval] Section 8 [IMT Eval] Section 8 [IMT Eval] Section 8 [IMT Eval] Section 8 [IMT Eval] Section 8 [IMT Eval] Section 8 (*) Pending availability of applicable channel model.1.2. Urban micro-cell scenario ITU Base coverage urban.1.2.814 V9.Release 9 59 3GPP TR 36.1 See [IMT Eval] Annex. Urban macro-cell scenario ITU High speed.A.0 2.1 See [IMT Eval] Annex.1.5 2.A.0 2.4 3. N. Indoor hotspot scenario ITU Indoor extended.8 ISD (meters) 500 500 1732 60 60 200 500 500 1732 BW (MHz) FDD:10+10 TDD: 20 FDD:80+40* TDD: 80* FDD:10+10 TDD: 20 FDD:20+20 TDD: 40* FDD:80+40* TDD: 80* FDD:10+10 TDD: 20 FDD:10+10 TDD: 20 FDD:80+40* TDD: 80* FDD:10+10 TDD: 20 PLoss (dB) 20 20 20 N.1-2 and 25.1-1 – E-UTRA simulation case minimum set Simulation Cases 3GPP case 1 3GPP case 1 extended 3GPP case 3 ITU Indoor.

0 (2010-03) Table A. Am = 25 dB 2    . 46/49dBm – 10. Am } UE power class 3GPP Spatial Channel Model (SCM) [TR 25.2.1-2 – 3GPP Case 1 and 3 (Macro-cell) system simulation baseline parameters modifications as compared to TR 25. an optional orientation as shown can be used if needed in Coordinated Multipoint study (i.Release 9 60 3GPP TR 36.Case 1 . Table A. 23dBm (200mW) This corresponds to the sum of PA powers in multiple Tx antenna case In addition to the antenna bore-sight orientation in TR25. as well as for a potential additional mechanical tilt. but may be set to fit other RRM techniques used.1. For these evaluations [49 dBm] Total BS Tx power should be used. For calibration purpose. 20MHz carrier Some evaluations to exploit carrier aggregation techniques may use wider bandwidths e. The categorization of the low power nodes is as described in Table A.2-1. For calibration purposes. θ ) = − min{ − [ AH ( ϕ ) + AV (θ ) ]. A subset of the macro-cell layouts described in section A.0.1. 60 or 80 MHz (FDD).e. Am      = 70 degrees.Rural/high speed To assess the benefit of adding low-power nodes to become a heterogeneous network. The value for e this parameter. is not specified here.1. θ dB 3 = 10. the Typical Urban (TU) channel model may be used 43dBm – 1.996] For single transmit antenna evaluations.1. SLAv = 20 dB Combining method in 3D antenna pattern Channel model Total BS TX power (Ptotal) A( ϕ . the values θ etilt = 15 degrees for 3GPP case 1 and θ etilt = 6 degrees for 3GPP case 3 may be used.814 V9.2-1. performance comparison should be made to homogeneous macro-cell only deployment. the following cases should be used .25. Antenna height at the UE is set to 1.2.2.2.2 Heterogeneous deployments Heterogeneous deployments consist of deployments where low power nodes are placed throughout a macro-cell layout.1 could be used for heterogeneous network deployments evaluation.1.g.. Categorization of new nodes Backhaul Access Notes 3GPP .2. point to corners) for 3GPP internal evaluations A. 5MHz carrier.1. Antenna height at the base station is set to 32m.5m.814 Parameter Antenna pattern (horizontal) (For 3-sector cell sites with fixed antenna patterns) Assumption   ϕ AH (ϕ) = − min 12     ϕ3 dB  ϕdB 3 2    .1.1.814 (center direction points to the flat side).1. SLAv      Antenna pattern (vertical) (For 3-sector cell sites with fixed antenna patterns)   θ −θ etilt AV (θ ) = − min 12  θ   3dB  The parameter θ tilt is the electrical antenna downtilt.Case 3.

Table A.0 (2010-03) Several µs latency to macro X2 No X2 as baseline (*) Through air-interface with a macro-cell (for in-band RN case) Open to all UEs Open to all UEs Closed Subscriber Group (CSG) Open to all UEs Placed indoors or outdoors Placed indoors or outdoors.1. node for Femto cells) Relay nodes 61 3GPP TR 36.2.1 5.1) are studied separately.e. Evaluations with interference management for HeNBs (via X2 or other means) allowed to assess interference management benefits Table A. Placed indoors.2 Macro + Outdoor Environment Macro + Indoor Deployment Scenario Macro + femtocell Macro + indoor relay Macro + indoor RRH/Hotzone Macro + outdoor relay Macro + outdoor RRH/Hotzone Non-traditional node femtocell Indoor relay e.2. Indoor HeNB clusters 2.2. Heterogeneous network deployment scenario Case 5.Release 9 Remote radio head (RRH) Pico eNB (i.1. (*): Baseline is in accordance to Rel-8/9 assumption.2 5.g.e.g.1.2. More detailed modelling of new nodes propagation and channel model based on IMT. indoor pico Outdoor relay e. Typically planned deployment.0.1. Consumer deployed. outdoor pico Note 1: Priorities are as follows: 1.2.2-4) 3. 3GPP . Other scenarios can be studied with lower priority Note 2: Relay deployment scenario (5. node for Hotzone cells) HeNB (i.2-2 presents the baseline deployment scenario for Heterogeneous network.1.1..2-2.1. Table A.1. 6. Outdoor Hotzone cells with configuration #1 and #4 (in Table A.814 V9.3 6.1 6. Indoor Hotzone scenario (RAN4 femto or pico models could be used) 4. Placed indoors or outdoors Note: The reference to new nodes in this TR and its corresponding characteristics are applicable to evaluations in this TR only.EVAL should be considered for performance evaluation at a later stage.2-3 presents the baseline parameters for initial evaluations in heterogeneous networks.

1.0 (2010-03) Table A.2-3.2.814 V9. Heterogeneous system simulation baseline parameters 3GPP .1.Release 9 62 3GPP TR 36.0.

1+42.2log10(R) Case 1: Prob(R)=min(0.0) 1 0 R 3GPP .4+24.018/R.4+24.7 +36 . R in km. R in km. Pico to UE: L =2 + 0 lo 17 3 g Other links L= 140 .7log for 2GHz.01)/1.2) Case 3 (Rural/ Suburban): Prob(R)=exp(-(R-0. and outdoor relay: 1. 2. R in km 1 0 R L= 128.063) Case 3 (Suburban): Prob(R)=exp(-(R-0.6log10(R) for 2GHz.063))+exp(-R/0.Release 9 Parameter RRH / Hotzone Nodes per macro-cell 63 Assumption Femto 3GPP TR 36.1+37.0 (2010-03) Relay Distance-dependent path loss from new nodes to UE*1 RRH/Hotzone.01)/0. the Model 2: Macro to UE: PLLOS(R)= 103.814 V9. 4 or 10 (nodes) Femto and indoor relay: 1 (cluster) Note: The number of HeNB and indoor relay nodes in each cluster is FFS. For outdoor RRH/Hotzone Model 1: 5x5 Grid Macro to UE: PLLOS(R)= 103.6log10(R) For 2GHz.8log10(R) inside the same Macro to UE: cluster L= 128.0.1)*(1-exp(R/0.1+37.2log10(R) Femto to UEs Model 1: PLNLOS(R)= 131.

1.095)) For indoor RRH/Hotzone. R in km Case 1: Prob(R)=min(0.2-8 3GPP TR 36.2.5-1/2 number of floors in the path is assumed to be 0. 5exp(-R/0.3exp(-0.2.4+24.Release 9 64 PLNLOS(R)= 131.2.1+42.1. 3exp(-R/0. the links from other macros to optimized relay site could be LOS or NLOS. Penetration loss 20dB Case 1: Prob(R)=min(0.1.1.5min(0.5log10(R) For 2GHz.2.0) Note 4: Higher probability of LOS shall be reflected in consideration of the height of RN antenna 3GPP .0 (2010-03) Macro to relay: Relay with outdoor donor antenna: PLLOS(R)=100.1)*(1-exp(R/0.5exp(-0.2) Case 3 (Rural/ Suburban): Prob(R)=exp(-(R-0. R in km Case 1: Prob(R)=0.1.2-8 Model 2 See Table A.0) Pico to UE: PLLOS(R)=103. see Table 2.01)/1.1.1.01)/0.018/R.0.2.9log10(R) PLNLOS(R)=145.2log10(R) PLNLOS(R)= 131.156/R))+min(0.1.3/R))+min(0. R in km.5.23) Case 3 (Rural/ Suburban) Prob(R)=exp(-(R-0.4].2. Model 1: Dual strip Model See Table A.5log10(R) PLNLOS(R)= 125.5.1.01)/1.018/R.814 V9.072) Case 3 (Suburban): Prob(R)=exp(-(R-0.8+20. Note3: If link from donor Macro to optimized relay site is LOS.15) Note 1: Bonus for donor macro (from each of its sectors) to relay for optimized deployment by site planning optimization methodology described in [A.8log10(R) For 2GHz.063) Case 3: Prob(R)=exp(-(R0.4+37.5.1+42. described in [A.4]. Prob(R) based on ITU models: Case 1: Prob(R)=min(0.1)*(1-exp(R/0.018/R.1.1.1.3log10(R) For 2GHz.7+23. Note 2: Higher probability of LOS shall be reflected in consideration of the height of RN antenna and site planning optimization.063))+exp(-R/0.2-7 and A.1.03)) Case 3: Prob(R)=0.063) Case 3 (Suburban): Prob(R)=exp(-(R-0.8log10(R) For 2GHz.5min(0.01)/0.2-7 and A.1)*(1-exp(R/0. else all interference links from other macros are NLOS.1. Relay with indoor donor antenna: PLLOS(R)= 103.5.2+36. R in km.063))+exp(-R/0.01)/1.072))+exp(-R/0.

R in km Case 1: Prob(R)=0.1.4+37.Release 9 65 3GPP TR 36. 5exp(-R/0.095)) Note 1: this path loss model assumes in-band relay.1.1.2.5min(0.3exp(-0. HeNB UE.5.1.2-8 Similar to UMTS 30.e.. the number of floors in the path is assumed to be 0.03)) Case 3: Prob(R)=0.8+20.2135 (i. 8 dB Lognormal Shadowing Shadowing standard deviation*2 3GPP .5log10(R) For 2GHz.1.1.1.5exp(-0.1.9log10(R) PLNLOS(R)=145.5. 3exp(-R/0.41.5min(0. R in km. Model 1: Dual strip Model Use the HeNB to UE model in Table A.0.2.0 (2010-03) Relay to UE: Relay with outdoor coverage antenna: PLLOS(R)=103. Simulations for out-of-band relay should re-examine this assumption.2.156/R))+min(0.2. according to LOS.5. Relay with indoor coverage antenna: Model 1: 5x5 Grid Relay to UEs inside the same cluster L =2 + 0 lo 17 3 g 1 0 R Relay to UEs in different clusters L= 128.2-7 and A.03.4 [ETSI TR 101 112] For outdoor RRH/Hotzone Model 1: Macro to relay for pathloss model 1: 5x5 Grid Relay with outdoor donor 10 dB 10dB for Link antenna: 6 dB for pathloss model 2 : ITU-R between HeNB and Relay with indoor donor antenna: M.1+37. B 1.6log10(R) for 2GHz.3/R))+min(0.5.2-8 Model 2 Use the HeNB to UE model in Table A.814 V9.2-7 and A.

Indoor relay Model 1: 5x5 Grid: 20 dB.27/8. 8dB for other links.3.1. Model 2: 3dB for LOS Link between relay and relay UE.2.2-7/8.1.2. 8dB for NLOS other links. FFS for LOS other links.1. Dual-strip: 20 dB if the UE is indoors. Model 2: see the HeNB to UE model in Table A.1. see Table 2.2. see Table 2. 0 dB if the UE is outdoors. 8dB for NLOS other links. Model 2: 20 dB if the UE is indoors.5 Relay with indoor coverage antenna: 0 N/A Macro to UE Outdoor relay: 20 dB. Macro to relay: Relay with outdoor donor antenna: 0 dB Relay with indoor donor antenna: 5 dB Relay with outdoor coverage antenna to UE: 20 dB for Case 1. 8dB for other links. Dual-strip: see Table A.1. Model 2: see Table A.3.1.0 (2010-03) Relay to UE: Relay with outdoor coverage antenna: 10 dB Relay with indoor coverage antenna: Model 1: 5x5 Grid 10 dB for link between relay and relay UE.2.Eval for ITU Rural Relay with indoor coverage antenna to UE: Model 1: 5x5 Grid Relay to UEs inside the same cluster:0 dB Relay to UEs in different clusters: 20 dB. 0 dB if the UE is outdoors.5 For indoor RRH/Hotzone 0 N/A For outdoor RRH/Hotzone 20 dB for Case 1.814 V9.1.1.2-7/8.5-1/2 66 8dB for other links Dual strip: 4dB for Link between HeNB and HeNB UE. 3GPP TR 36.1.1. See ITU. 4dB for NLOS Link between relay and relay UE. 8dB for other links Dual strip: 4dB for link between relay and relay UE.Release 9 NLOS) For indoor RRH/Hotzone. FFS for LOS other links. See ITU.Eval for ITU Rural 0 Between sectors Penetration Loss N/A Model 1: 5x5 Grid Femto to UEs inside the same cluster: 0 dB All other links: 20 dB. Dual-strip: see the HeNB to UE model in Table A. Model 2 : 3dB for LOS Link between HeNB and HeNB UE. 4dB for NLOS Link between HeNB and HeNB UE.0. Relay with outdoor coverage antenna:0. Shadowing correlation Between cells*3 For outdoor RRH/Hotzone 0.2-7/8 above for both "Macro to UE" and "RRH/Pico to UE" For indoor RRH/Hotzone.5-1/2 3GPP .1.1.

the impairment of frequency-selective fading channels shall be captured in the physical layer abstraction.1. Relay) Minimum distance among new nodes CF= 2GHz for case 1 and case 3 CF = 0.4-3 Case3: 24.2.1.1.2-4 See Table A.4-3 5dBi 5dBi See Table 2.1.1. outdoor relay and outdoor RRH/Hotzone model 2 LOS and NLOS path loss models are based on field measurements and LOS/NLOS probability functions for macro to UE and macro to relay are based on ITU models and probability functions for relay or pico to UE are based on field measurements.2.8GHz for high speed rural If fast fading modelling is disabled in system level simulations for relative evaluations.2-4 40 m cluster radious *1 Outdoor RRH/Hotzone model 1 is based on TR 25. For SIMO. 4 rx antenna ports See Table 2. Fast fading may be modelled using any of the following: No fast fading as in current TR Fast fading with TU and fixed correlation matrix 3GPP .814 V9.1. Path loss models for case 3 with 10m antenna height are FFS.26 >=75m >= 35m For outdoor RRH/Hotzone > 10m For indoor RRH/Hotzone >= 3m 40 m >= 3m Outdoor relay:> 10m Indoor relay: >= 3m 40 m See Table A. *2 Shadowing fading value for relays is applicable to NLOS component of the path loss.4-3 See Table A.1. or 4 tx .2. the physical layer abstraction is FFS.1. RRH. hotzone or relay nodes. Outdoor relay antenna height for both access and backhaul link is 5m. 3 km/h for UEs served by femto cells. Case1: 24.4-3 antenna ports See Table 2. 2 rx antenna See Table 2. Femto. 30.814 and IMT.4-3 2 tx .1.1.0 (2010-03) A(θ ) = 0 dB (omnidirectional) A(θ ) = 0 dB (omnidirectional) See Table 2.1. femto and indoor relay path loss is based on ITU-R M1225 single floor indoor office model.1. or 4 tx . 37 dBm – 10MHz carrier carrier See Table 2.4-3 Carrier Frequency Channel model UE speeds of interest Doppler of relay-macro link Total BS TX power (Ptotal) UE power class Inter-cell Interference Modelling Antenna configuration Antenna gain + connector loss [Motorola: reference for these values?] Placing of new nodes and Ues Minimum distance between new node and regular nodes Minimum distance between UE and regular node Minimum distance between UE and new node (RRH/Hotzone.1.1.4-3 (37dBm is outdoor only) 23dBm (200mW) This corresponds to the sum of PA powers in multiple Tx antenna case UL: Explicit modelling (all cells occupied by UEs).1.Release 9 Antenna pattern (horizontal) 67 3GPP TR 36.1.1.EVAL UMi NLOS model.1.1. 2 rx antenna ports.0. 4 rx ports. 30 dBm – 10MHz carrier 20 dBm – 10MHz See Table 2. the physical layer abstraction is based on TU link curves. DL: Explicit modelling else cell power = Ptotal 2 tx . N/A N/A Jakes spectrum with [5]Hz for NLOS component.4-3 See Table 2.1. LOS component [K=10dB]. For MIMO. Case 1 and Case 3: 3 km/h Rural high speed: 120 km/h for UEs served by macro.1.1. *3 Cells including macro cells of the overlay network and new nodes. Value for LOS component is FFS.

Nusers .1. Randomly and uniformly drop Nusers_lpn users within a 40 m radius of each low power node. 3GPP . N users_lpn = P hotspot ⋅ N users /N  with Photspot defined in Table A. e.1. *** Clustered UE Placement for Hotzone cells: Fix the total number of users.2-5.0 (2010-03) - Fast fading with ITU/SCM models or possible simplifications [ref.1.Nusers_lpn*N.Release 9 68 3GPP TR 36.4. see section A2. within each macro geographical area (the same number N for every macro geographical area. The baseline fast fading model to be used for final evaluation results should be discussed in future (or in WI). UE density is defined as the number of UEs in the geographic area of a macro cell.1. 4b UE density across macro cells* Uniform 25/macro cell Non-uniform [10 – 100]/macro cell Non-uniform [10 – 100]/macro cell Non-uniform*** UE distribution within a macro cell Uniform Uniform Uniform Clusters New node distribution within a macro cell Uncorrelated Uncorrelated Correlated** Correlated** Comments Capacity enhancement Sensitivity to non-uniform UE density across macro cells Cell edge enhancement Hotspot capacity enhancement * New node density is proportional to the UE density in each macro cell.) For preliminary simulation. R4-091103] could also be used.814 V9. (Detailed proposals to be discussed. Configuration #4a and #4b parameters for clustered user dropping Configuration Configuration #4a* Nusers 30 or 60 N 1 2 4 10 1 2* 4* Photspot 1/15 2/15 4/15 2/3 2/3 2/3* 2/3* Configuration #4b 30 or 60 * Baseline for Configurations #4a and #4b. relevant propagation model to use with these. **Relay and hotzone nodes.. N. Randomly and uniformly drop the configured number of low power nodes. Nusers. Randomly and uniformly drop the remaining users. where hotspot users over the total number of users in the network. to the entire macro geographical area of the given macro cell (including the low power node user dropping area).g.1. where Photspot is the fraction of all Table A. where Nusers is 30 or 60 in fading scenarios and 60 in non-fading scenarios.2-5.1. 4.2. 10}). often deployed by planning. where N may take values from {1. dropped within each macro geographical area.1.2. all fast fading channel models can be applied.2-4. Table A.1. Placing of new nodes and UEs Configuratio n 1 2 3 4a.2. 2.0.

Release 9 69 3GPP TR 36.1.814 V9.2.0 (2010-03) Table A.1. Placing of femto cells and UEs 3GPP .2-6.0.

0 (2010-03) Placing of UEs Random placing of UEs within X meters of the femto cell 3GPP .Release 9 Configuratio n 1 Macro-femto Deployment Independent channel 70 Placing of nodes Clustered 3GPP TR 36.814 V9.0.

0.Release 9 2 Co-channel 71 Clustered 3GPP TR 36.814 V9.0 (2010-03) Random placing of UEs within X meters of the femto cell 3GPP .

0.1.2-7. (Note that this does not preclude that one femto cluster may contain no UEs) * Fraction of Macro UEs in Femto Cluster: 35% and/or 80% Table A. Indoor femto Channel models (dual strip model): Suburban deployment 3GPP .Release 9 72 3GPP TR 36.814 V9.2.1.0 (2010-03) * Femto cell with 5x5 grid or dual-strip apartment blocks * Non-uniform macro-UE drop in a femto cluster.

8+24. the last term is not needed InH.0.46) R and d2D. UE to macro BS Prob(R)=exp(-(R-10)/1000) Model1: PL (dB) =15.Release 9 73 3GPP TR 36. R in m.7+42. R in m. Prob(R)=exp(-(R-10)/1000) UE to HeNB (3) Dualstripe model: UE is inside the same house as HeNB PL (dB) = 38.indoor are in m n is the number of penetrated floors In case of a single-floor house.46 + 20 log10R + 0.2log10(R) + Low (2) UE is inside a house PLNLOS(R)=2. 3GPP .2log10(R) PLNLOS(R)= 2.3 n ((n+2)/ (n+1)-0.3 + 37.0 (2010-03) Fast Fading (when fast fading in both frequency and spatial domains is modeled) Cases Path Loss (dB) Model 1: PL (dB) =15.7d2D.6log10R + Low. Model2: (1) UE is outside: PL(R) PLLOS(R)= 30.6log10R For 2GHz.8log10(R) + Low RMa For 2GHz. R in m.8log10(R) RMa For 2GHz.indoor+ 18.8+24.3 + 37. R in m Model2: PLLOS(R)= 30.7+42.814 V9. LOS or NLOS depends on whether line-of sight from UE to HeNB.

46) + Low.indoor is the total distance inside the two houses.7d2D. equal to 0.indoor + 18.3 + 37.8 log10 R.indoor are in m InH (NLOS) Model 2: PL(dB) = max(2.7dB/m. which is 20dB. 38. the path loss modeling takes account of case (2) and case (3).3 + 37.indoor are in m Model 1: PL(dB) = max(15.indoor + 18.46) + Low (4) Dualstripe model: UE is outside R and d2D.1 + Low. 38.0.indoor is the distance inside the house. In Case (3). the path loss is modeled by free space loss.7+42.3 n ((n+2)/(n+1)-0.0 (2010-03) Model 1: PL (dB) = max(15.indoor are in m R is the Tx-Rx separation Low is the penetration loss of an outdoor wall. In Case (5).2 (5) Dualstripe model: UE is inside a different house R and d2D.1 + Low.7+42.indoor are in m InH (NLOS) Model 2: PL (dB) = max(2.46) + Low R and d2D. d2D.46 + 20log10R) + 0.46) + Low.7d2D.3 n ((n+2)/(n+1)-0.2 are the penetration losses of outdoor walls for the two houses.8 log10 R.1 and Low.7d2D.3 n ((n+2)/(n+1)-0.46 + 20log10R) + 0. penetration loss due to internal walls and floors.indoor + 18.7d2D.6log10R.46 + 20log10R) + 0. In Case (4). 38. The loss due to internal walls is modeled as a log-linear value.3 n ((n+2)/(n+1)-0. Low. d2D. 38.814 V9. 3GPP .46 + 20log10R) + 0.2 R and d2D.6log10R.indoor + 18.Release 9 74 3GPP TR 36.

1.1.814 V9.0.0 (2010-03) Table A. Indoor femto Channel models (dual strip model): Urban deployment 3GPP .2.Release 9 75 3GPP TR 36.2-8.

1)*(1-exp(-R/63))+exp(-R/63) UE to macro BS Model1: PL (dB) =15.3 + 37.6log10R. the last term is not needed InH.46) PL (dB) = 38.7+42.6log10R + Low. R in m Prob(R)=min(18/R.0 (2010-03) Fast Fading(when fast fading in both frequency and spatial domains is modeled) Cases Path Loss (dB) Model1: PL (dB) =15.46 + 20 log10R + 0.0.814 V9.Release 9 76 3GPP TR 36. LOS or NLOS depends on whether line-of sight from UE to HeNB.1)*(1-exp(-R/63))+exp(-R/63) UE to HeNB (3) Dualstripe model: UE is inside the same apt stripe as HeNB 0.2log10(R) + Low (2) UE is inside an apt PLNLOS(R)= 2. R in m Model2: PLLOS(R)= 30. R in m Model2: PLLOS(R)= 30.indoor are in m n is the number of penetrated floors q is the number of walls separating apartments between UE and HeNB In case of a single-floor apt.indoor+ 18.8log10(R) UMa For 2GHz.3 + 37.7d2D.2log10(R) (1) UE is outside PL(R) PLNLOS(R)= 2. Prob(R)=min(18/R. 3GPP .8log10(R) + Low UMa For 2GHz.7+42.3 n ((n+2)/(n+1)+ q*Liw R and d2D. R in m.8+24.8+24.

814 V9.0.8 log10 R.3 n ((n+2)/(n+1)-0.46) + q*Liw + Low InH (NLOS) R and d2D.1.1 + Low.2 InH (NLOS) R and d2D. Low.3 + 37. 38. The term 0. Low is the penetration loss of an outdoor wall.7+42.indoor + 18.2.3 + 37.7d2D.6log10R.0 (2010-03) Model 1: PL (dB) = max(15.3 n ((n+2)/(n+1)-0. which is 20dB. which is 5dB. A.46) + q*Liw + Low (4) Dualstripe model: UE is outside the apt stripe Model 2: PL (dB) = max(2.7d2D.2 are the penetration losses of outdoor walls for the two houses.7+42.7d2D.3 Assumptions for Coordinated Multi point Transmission and Reception Evaluations Performance evaluations should at least provide details related to: Cooperating scheduler 3GPP .indoor are in m q is the number of walls separating apartments between UE and HeNB Model 1: PL(dB) = max(15.7d2D.46 + 20log10R) + 0.7d2D.2 (5) Dualstripe model: UE is inside a different apt stripe Model 2: PL(dB) = max(2. 38.46 + 20log10R) + 0.indoor + 18.6log10R.indoor are in m q is the number of walls separating apartments between UE and HeNB Liw is the penetration loss of the wall separating apartments.indoor takes account of penetration loss due to walls inside an apartment.8 log10 R.3 n ((n+2)/(n+1)-0.Release 9 77 3GPP TR 36.1. 38.46) + q*Liw + Low.indoor + 18.3 n ((n+2)/(n+1)-0.46 + 20log10R) + 0.1 and Low.46) + q*Liw + Low.46 + 20log10R) + 0.1 + Low. 38.indoor + 18.

4-1.1. 3GPP .2. The proponents should describe and justify the model assumed in particular studies. A.814 V9.Release 9 78 3GPP TR 36.Relay 3GPP case 3. Positions. Evaluation scenarios for relay CASE 3GPP case 1. compared to the geometry cdf for a non-CoMP UE. the placement of relay also regarded as site planning could be taken by two major steps.1.0 2.4 Assumptions for Relay Evaluations The evaluation scenarios for relay is summarized as two basic scenarios as follows.0 For 3GPP case 1&3. [TBD] 2) Relay Site planning: Finding an optimal place among N candidate relay sites around the virtual relay which offers optimization of shadow fading. Relays were placed on positions they are most needed in terms of SINR (geometry).Relay scenarios. The performance of downlink/uplink multi-point transmission and reception.1.0. were found in an exhaustive search within N candidate relay sites around the virtual relay.Indoor relay Scenarios Urban Macro Rural Area Urban Macro Rural Area ISD[m] 500 1732 500 1732 Carrier[GHz] 2.0 (2010-03) - CoMP category Feedback assumption and feedback impairment modelling Backhaul assumptions Time/frequency synchronization assumptions Transmission modes: MU-MIMO and/or SU-MIMO operation in conjunction with CoMP Selection of transmission mode (assumptions on how dynamic or semi-static the transmission mode can be selected) - Creation and maintenance of CoMP sets: Assumptions on CoMP sets definition and creation fixed vs.0 2. 1) Virtual Relay Placement: A virtual relay is placed trying to enhance the cell edge throughput or overall cell throughput. or more generally latency between new nodes.Indoor relay 3GPP case 3. where a relay placement would result in the highest geometry gain. the backhaul latency could be classified into the following categories Minimal latency (in the order of μs) for eNB to RRH links Low latency (<1 ms) associated with co-located cells or cells connected with fibre links and only limited number of routers in between Typical inter-cell latency associated with X2 interfaces. LOS probability and etc. The X2 backhaul latency. size of cluster… Geometry cdf for the CoMP UE should be provided where appropriate.1. or new nodes and eNBs. and advanced ICIC techniques is sensitive to the backhaul capacity and latency. or between eNBs.2. In general. adaptive clusters. is highly deployment dependent such as whether there is a dedicated X2 fibre network or a generic IP network.Relay 3GPP case 1.0 2. Table A.

1.814 V9.1. for donor macro (from each of its sectors) to relay.1. Table A. This process offers optimization of shadow fading.4-2. For simplicity.Indoor relay scenarios.The best relay site is selected based on SINR criteria on the backhaul link. For 3GPP case 1&3. LOS probability and shadowing standard deviation are listed in Table 2. 1-(1. Macro-relay Path Loss For LOS: PLLOS(R) For NLOS: PLNLOS(R) Macro-relay LOS probability Prob(R) Alternative 2: Initialized in a system-level simulation by selecting best N relays according to a proposed site planning optimization approach. It is a process of finding an optimal place among N candidate relay sites around the virtual relay which offers benefit to the performance. 3GPP .Release 9 79 3GPP TR 36. the candidate relays are randomly placed in the searching area. otherwise.2-3. if the relay node is made up of a donor module and a coverage module. TWO alternatives to show this benefit in simulation are considered with respect to Alternative 1: Adding bonus to path loss formula. no planning is performed. This optimization is reflected by a penetration loss of 5 dB in Table A. The coverage module has to be placed in the center of the house to provide maximum coverage.0. the donor module has to be placed near a window to optimize the backhaul link quality. for non-donor cell and non optimized deployment B=0dB.1. for donor macro (from each of its sectors) to relay. The site planning optimization should be taken into consideration in a relay placement procedure in step 2. The site planning optimization approach are described as follows. The corrections of site planning with respect to macro-relay pathloss.2. For a typical system level simulation.1. both being placed indoors. for non-donor cell and non optimized deployment N=1. depending on the relay node configuration: if the relay donor antenna is outdoors. the following configurations are taken into consideration. otherwise. . Corrections of site planning (alternative 1) No site planning Correction after site planning For LOS: PLLOS(R) For NLOS: PLNLOS(R)-B Where B=5dB.2.4-2.1.Prob(R))^N Where N=3. LOS probability and etc.0 (2010-03) The site planning procedure provides benefit on backhaul SINR (geometry). N=5 candidate relay sites are considered within a searching area of 50m radius around the virtual relay. 2 cases are distinguished.

2    θ3dB   2 tx Am rx antenna ports. from the donor cell Omni Macro-Relay link A(θ ) = 0 dB 2 donor antenna:   θ Indoor 2 tx .2. Am = 20dB. 2 rx antenna ports.1.  . Am = 20 dB. 2    . 10 m Indoor antenna: N/A N/A HRN RS antenna height 5m 5m. Relay-UE link: Relay-UE link: 5dBi antenna gain.. 4 rx antenna ports or 4 tx . directional θ3dB = 70 degrees.Release 9 80 3GPP TR 36.0.       Macro-Relay link 7dBi. 2 tx . 5dBi antenna gain. Am = 20 dB. 4 rx antenna ports One antenna set Use of antenna downtilt Use of antenna downtilt and vertical antenna FFS and vertical antenna FFS Antenna Configuratio n Two antenna sets θ3dB   θ A(θ ) = − min 12     θ3dB  = 70 degrees.0 (2010-03) Table A. 4 rx antenna ports 5dBi antenna gain. or 4 tx . 4 rx antenna ports  . Am  directional A(θ ) = − min 12    Outdoor donor antenna:   θ3dB     Macro-Relay link 7dBi. 2 rx antenna ports. Am  A(θ ) = − min 12  or 4 tx . 2  antenna ports. or 4 tx . 5dBi. or 4 tx . 2 rx antenna ports. 4 rx antenna ports Use of antenna downtilt and vertical antenna FFS 5 dB 0 dB 5 dB Outdoor donor antenna: 2 dB Indoor donor antenna: 0 dB 3GPP . directional 7dBi. 4 rx antenna ports θ  Use of antenna downtilt and vertical antenna FFS θ3dB    3 dB    = 70 degrees. 2 tx . Omni 5dBi. 2 rx antenna ports. Am  A(θ ) = − min 12      . 2 rx   θ 2tx . 2 rx antenna ports. directional θ3 dB = 70 degrees. Omni Omni A(θ ) = 0 dB A(θ ) = 0 dB or 2 tx .814 V9.4-3. 10m 5dBi antenna gain.1. Omni 2 tx . 2 rx antenna ports. Typical configuration for simulation Parameter Description Case 1 Case 3 Case 1/3 Indoor Downlink: 20 dBm @ 10 MHz bandwidth Uplink: PRN Max Tx power 30 dBm @ 10 MHz bandwidth 30 or 37 dBm @ 10 MHz Indoor donor antenna: bandwidth 23 dBm @ 10 MHz bandwidth Outdoor donor antenna: 30 dBm @ 10 MHz bandwidth Outdoor antenna: case 1: 5 m case 3: 5 m. NFRN HWRN Noise figure Hardware loss/cable loss 5 dB 0 dB 2 tx . Relay-UE link: directional pointing away or 4 tx . Am   θ 2 = 20 dB.

5-1.1. R in km. R in km 40dB ITU InH (NLOS) UE to RRH/Hotz one 20dB PLLOS(R)= 89.1+42.018/R.018) / 0. The height of the floor is 6 m. Sketch of indoor hotspot environment The following two channel models are for evaluation purpose. R in km Prob(R)= 1 R ≤ 0.9log10(R) PLNLOS(R)= 147.01)/1.063))+exp(-R/0.1)*(1-exp(R/0.2.Release 9 81 3GPP TR 36.1. The floor contains 16 rooms of 15 m x 15 m and a long hall of 120 m x 20m. 147. 3GPP .037  LOS: 3dB NLOS: 4dB 0dB ITU InH * No Fast Fading orTU and fixed correlation matrix can also be used.8log10(R).1.1.4+43.5 Assumptions for indoor RRH/Hotzone Evaluations The indoor hotspot scenario consists of single floor of a building as Figure. Figure.0 (2010-03) A.063) Case 3: Prob(R)=exp(-(R-0.5-1.5-1 which is same with ITU sketch.3log10(R) For 2GHz.814 V9. Table A.1.018   exp(−( R − 0.018 < R < 0.1.027) . Two sites are placed in the middle of the hall at 30m and 90m with respect to the left side of the building.4+24.0) 0dB ITU UMa 20dB PL(dB) =Max(131.8log10(R) For 2GHz. 10dB Case 1: Prob(R)=min(0.5 R ≥ 0.4+43. 0.2log10(R) PLNLOS(R)= 131.0. Channel model 1 of indoor RRH/Hotzone Shadowing standard deviation Cases Path Loss (dB) Penetration Loss Fast Fading (when fast fading in both frequency and spatial domains is modelled)* (1) UE is outside UE to macro BS (2) UE is inside (1) UE is inside a different building as the indoor hotzone (2) UE is outside (3) UE is inside the same building as the indoor hotzone PLLOS(R)= 103.1+42.1. Channel model 2 is a simplified channel model which assumes all users are allocated within buildings.2.5 + 16.1.3log10(R)) 10dB For 2GHz.2.037  0.2.

063))+exp(-R/0. Channel model 2 of indoor RRH/Hotzone Shadowing standard deviation Cases Path Loss (dB) Penetration Loss Fast Fading (when fast fading in both frequency and spatial domains is modelled)* PLLOS(R)= 103.018   exp(−( R − 0.4+43.1.1+42.5 + 16.8log10(R) For 2GHz. evaluations with time-varying interference shall be carried out using bursty traffic models.3log10(R)) 10dB For 2GHz. 3GPP . which are specified in the IMT.5-2.1+42.1. A.3-1 proposes FTP traffic models to exercise system performance studies in bursty traffic.3 Traffic models Traffic models for system performance evaluations are given in Table A.1. System throughput studies shall be assessed using full-buffer traffic model capturing continuous traffic and non-varying interference.Release 9 82 3GPP TR 36.4+24. * * For UE is outside the same building as the indoor hotzone.2. Table A. R in km Prob(R)= 1 R ≤ 0. R in km.8log10(R).01)/1.0.3-1. A. R in km 20dB ITU InH (NLOS) PLLOS(R)= 89.037  LOS: 3dB NLOS: 4dB 0dB ITU InH * No Fast Fading orTU and fixed correlation matrix can also be used. UE to macro BS Case 1: Prob(R)=min(0.3log10(R) For 2GHz.2 Channel models Annex B describes the IMT-Advanced Channel Models. 0. Additionally.1.0 (2010-03) Table A.2.2.063) Case 3: Prob(R)=exp(-(R-0.0) (1) UE is outside the same building as the indoor hotzone* * UE to RRH/Hotz one (2) UE is inside the same building as the indoor hotzone 10dB 20dB ITU UMa PL(dB) =Max(131.9log10(R) PLNLOS(R)= 147.2.2.1)*(1-exp(R/0.5 R ≥ 0.814 V9.018) / 0.037  0.2log10(R) PLNLOS(R)= 131.1. 147.018 < R < 0.4+43.EVAL of ITU-R[15].027) .018/R. UE are allocated within buildings which either indoor hotzone nodes are deployed or not.1.

5.5 Mbytes can be chosen to speed-up the simulation.5.3. FTP Traffic Model 1 Parameter File size.2.1. S User arrival rate λ Statistical Characterization 2 Mbytes (0. Tables A.Release 9 83 3GPP TR 36.2. The reading time D is the time interval between end of download of previous file and the user request for the next file.1. DL and UL Real time services A.5 Mbytes can be also evaluated.3.1.1.1.1. Bursty traffic.1 FTP traffic models Two FTP traffic models are considered as non-full buffer traffic models. Table A. Possible range of λ: [0. The same traffic should be simulated for CoMP and non-CoMP schemes. Range of K can further be adjusted.1.2. 0.3. D Number of users.1.1.12. K Statistical Characterization 0.3-3 show the parameters for FTP traffic model 1 and model 2.2. FTP Traffic Model 2 Parameter File Size.5.5 Mbytes and Model 2 with file size of 0. 3GPP .2.3. DL and UL. 0. 10.2.37.1-2 illustrate the user arrival of traffic model 1 and 2.2) in non-CoMP SU-MIMO. respectively.0 (2010-03) Table A. 2.2) that leads to [10%. 50%] of RU in non-CoMP SU-MIMO. 5.1-1.3. Simulations are run for various λ to find performance metrics covering at least the range of HM-NCT (See A.25.625] for 2 Mbytes (See A. S Reading Time. 2.3.2.814 V9. 0.1-1: Traffic generation of FTP Model 1 Table A.1.1.3. D ≥ 0 λ = 0.2 - Simulations are run for various K to find performance metrics covering at least the range of HM-NCT that leads to [10%.4 for more details).1.2.3. 0.5 Mbytes. respectively.0. Continuous traffic.5 Mbytes Exponential Distribution. Possible range of K: [2. however Model 1 with file size of 0.1.4 for more details).2. 14] (See A. [0. 50%] of RU (See A. Figure A. Baseline model is Model 1 with file size of 2 Mbytes. Range of λ can further be adjusted. Traffic Models Traffic Models Full buffer Non-full buffer FTP models VoIP Model Applies to DL and UL. 8.1-2.3. 1.5 Mbytes optional) (one user downloads a single file) Poisson distributed with arrival rate λ - Small file size of 0.5] for 0.3. 1.1-1 and A. The above range of λ will cover RU from 10% to 50% for non-CoMP SU-MIMO Figure A.2. Mean= 5 seconds PDF: Fixed f D = λe −λD .3-1.2.3-2 and A.2.2.

1. X value is derived from the SU-MIMO case.3.3. SU-MIMO and MU-MIMO techniques: R1: X% HM-NCT value that leads to 50% RU for SU-MIMO (i. one RB allocated to N users within a cell is counted as used N times A.2.2 Performance metrics The following performance metrics are considered for non-full buffer traffic models.1 Reference points Possible reference points to compare CoMP. 50. The above range of K will cover RU from 10% to 50% for non-CoMP SU-MIMO. Figure A.1.0 (2010-03) - The same traffic should be simulated for evaluating CoMP and non-CoMP schemes.Release 9 84 3GPP TR 36. non-CoMP.2. and ends when the last bit of the packet is correctly delivered to the receiver Served cell throughput Served cell throughput = total amount of data for all users / total amount of observation time / number of cells - Harmonic mean normalized cell throughput (HM-NCT) Harmonic mean normalized cell throughput = served cell throughput / harmonic mean of the full buffer cell throughput Harmonic mean of the full buffer cell throughput = average number of full buffer users per cell / mean( 1 / full buffer user throughputs) The 2% worst and 2% best users are excluded from the harmonic mean Full buffer simulations are run with 10 UEs / cell - Normalized cell throughput (NCT) Normalized cell throughput = served cell throughput / full buffer served cell throughput - Resource utilization (RU) Resource utilization = Number of RB per cell used by traffic during observation time / Total number of RB per cell available for traffic over observation time In case of MU-MIMO.814 V9.e.2.0. 95 % user throughput User throughput = amount of data (file size) / time needed to download data time needed to download data starts when the packet is received in the transmit buffer.1-2: Traffic generation of FTP Model 2 A.2.3. 5. and is then the same for all techniques) R2: Traffic load that gives 50% RU for SU-MIMO R3: 45% NCT Per-user traffic S 3GPP .1. Mean.

4 For FTP Model 1 Estimation of range ofλ and K offered traffic =λ* S. 2. 2.1.3. - Method to restrict the impact of users in very poor radio conditions: if RLC is modelled: Remove from the system the users who experiment a MAC error rate higher than 3% a MAC error is defined as a transport block being not correctly received after the maximum number of retransmissions. or because the radio conditions are insufficient for reliable communication. and should be then described. Further refinement of adequate reference points can be considered. 10]/4 = [0. However. For a file size of 500kbyte = 4 Mbit. 1. λvalues for file size of 2Mbytes can be computed in the same way. A.2. 50%] range of resource utilization could be covered by the range of offered traffic [2 4 6 8 10] Mbps. 8. Method for ensuring queue stability in high load conditions: drop a file if its transfer is not completed within a maximum transfer time T_drop T_drop = 8s T_drop = 32s for S = 0.5 Mbyte for S = 2 Mbyte Methods for handling HARQ retransmission failure: model RLC ACK if RLC ARQ is not modeled. the results should be obtained at the other reference points to gain understanding about how these new metrics relate to each other. 1. For LTE 2x2 10MHz. - Dropping a file means: the file is given zero user throughput the data in the dropped file is not included in the served cell throughput - The amount of data dropped or not transmitted shall be recorded separately according to whether it arises from users removed due to high MAC error rate or other causes.5] users/s.0.0 (2010-03) Reference point R1 may be a baseline point. if RLC is not modeled: user outage is implicitly approximated by file dropping when the maximum number of HARQ retransmissions is reached. the inter-arrival time is given by Tt + D with Tt file transfer time. R2 captures the user throughput gain at a given offered load.3. For model2.Release 9 85 3GPP TR 36. the [10%. A.5. Alternatives methods can be considered. The total offered traffic per cell is then given by 3GPP . In this caseλ= 1/(Tt+D). drop the file after the maximum number of HARQ retransmissions is reached. Files may be dropped from the simulation either to preserve stability at high loads.3 File dropping criteria A guideline of modelling file dropping for non-full buffer traffic simulation is given as follows.5. 4. R1 and R3 capture both the user throughput and cell capacity gains in a reasonably well utilized network.814 V9.1. do not include removed users' throughput nor contribution to the served cell throughput in the simulation results. this corresponds to: λ= offered_traffic/S = [2. and the corresponding file dropping rates shall be indicated in the simulation results.2. 6.

32 ~= [2 5 8 10 14] users A. the following performance metrics need to be considered: Mean user throughput Throughput CDF Median and 5% worst user throughput For evaluations with bursty traffic model.e.0 (2010-03) offered traffic = λ* K * S = K * S / (Tt + D) Assuming download data rate of 12. System capacity is defined as the number of users in the cell when more than [95%] of the users are satisfied. i.Macro cell area throughput .32 sec for file size of 500 Kbytes = 4 Mbits.2.Release 9 86 3GPP TR 36. the corresponding range of K is K = [2 4 6 8 10]/4*5.1. A VoIP user is in outage (not satisfied) if [98%] radio interface tail latency of the user is greater than [50 ms].32+5).Existing full buffer and bursty traffic performance metrics . For VoIP capacity evaluations. the following performance metrics are the highest priority: .. defined as the size of a burst divided by the time between the arrival of the first packet of a burst and the reception of the last packet of the burst Average perceived throughput of a user defined as the average from all perceived throughput for all bursts intended for this user. Percentage of users with [1]% or more dropped packets.Macro and low power node serving UE throughput ratio The following table should be included along with the simulation assumptions accompanying all results: 3GPP . offered traffic is K*4/(0.4 - System performance metrics For evaluations with full-buffer traffic model. the following performance metrics need to be considered: VoIP system capacity in form of the maximum number of satisfied users supported per cell in downlink and uplink. For heterogeneous network performance evaluation. Overall average user throughput defined as average over all users perceived throughput. the file transfer (download) time Tt is 0. This assumes an end-to-end delay below [200 ms] for mobile-to-mobile communications.0. When the range of offered loads is [2 4 6 8 10]. Tail perceived throughput defined as the worst 5% perceived throughput among all bursts intended for a user User perceived throughput CDF (average and/or tail user perceived throughput).5 Mbps.814 V9.Fraction of throughput over low power nodes .Throughput CDFs are for all UEs. By using this rough estimate. the following performance metrics need to be considered: User perceived throughput (during active time). Median and 5% worst user perceived throughput (average and/or tail user perceived throughput). macro UEs and HeNB/pico UEs .

where ζ is the polarization slant angle1 and A(φ.2.0 (2010-03) Are Throughput Values based solely on an assumption of a number of trials of independent placing of UEs? Yes/No Comments If “Yes. Evaluations should include fairness.6.6.” either state the methodology by which confidence interval is achieved as well as confidence interval and confidence level. including relevant parameter values. ˆ and horizontal field direction are defined in terms of the spherical basis vectors. θ .1.θ) and FV (φ. θ) ).e.1.1.θ) = Fφ (φ. the polarization is modelled as angle-independent in both azimuth and elevation.θ ) sin (ζ ) .e. as described in section A.4 above.5 Scheduling and resource allocation Different scheduling approaches have impacts on performance and signalling requirements.θ ) = Fφ (φ. A.2.θ ) cos (ζ ) Fθ (φ.814 V9. in the antenna local coordinate system. 1 Hence. θ) = Fθ (φ.1. ζ ) = A(φ.θ ) = Fθ (φ.0. FH (φ. and feedback error assumptions should also be indicated. i. respectively. an antenna with ζ = 0 degrees corresponds to a purely horizontally polarized antenna. θ and φrespectively (i.1-1.. in the horizontal polarization and in the vertical polarization. delay. are given by Fφ (φ. the antenna element field pattern.θ.2. φ and elevation angle. ζ ) = and A(φ) cos (ζ ) Fθ (φ ) = Fθ (φ. ζ ) = and A(φ. and multipoint transmission schemes. For 2D element antenna gain pattern: Fφ (φ) = Fφ (φ.θ. similar indications should be given for any buffer status feedback.Release 9 87 3GPP TR 36.θ) is the 3D element antenna gain pattern as a function of azimuth angle.2. For uplink queue-related scheduling. which are defined in Figure A. 3GPP . or justify the method of user placing in different trials. A. ζ ) = A(φ ) sin (ζ ) .6 Antenna gain for a given bearing and downtilt angle A.” then state the number of trials. If “No.1 Polarized antenna modelling In case of polarized antennas. For frequency or carrier specific scheduling..2. any feedback approach. Evaluations should include a high-level description of the scheduling and resource allocation schemes simulated. placing of a group of UEs in cells used. For a linearly polarized antenna.1. Note that the vertical ˆ.

where n is the given direction. A. for sector antennas with cylindrical shape.0 (2010-03) z θ ˆ n ˆ φ ˆ θ φ y x Figure A.6. The global coordinate system is oriented with its z-axis along a vertical direction.1.y’. and all directions in space (angles of arrival and departure) are defined in the global coordinate system angular coordinates θ and φ . φ) in the global coordinate system to spherical angles ( θ . the antenna radiation pattern defined in the antenna-fixed coordinate system is measured with the antenna cylinder axis installed to coincide with the z'-axis of the antenna-fixed coordinate system. φ' ) in the local system.6.814 V9. the x-axis of the global coordinate system. a transformation must be performed to allow evaluation of the tilted antenna pattern as a function of coordinates in the global coordinate system. (1) (2) φ' = arg( cos φ sin θ cos β − cos θ sin β + j sin φ sin θ ). Having only the total gain pattern A' (θ' .1. does not in general affect the choice of antenna-fixed coordinate system. φ) -system is simply given by the relation 2 Note that the global coordinate system should not be interpreted as a system wide global coordinate system.z’) are defined with a common origin. should be aligned with the pointing direction of the sector2.2. but a coordinate system centered at the site antenna. Assume that an antenna is installed with the antenna aperture normal direction (and antenna main beam peak for a conventional sector antenna) in the xz-plane and that the antenna radiation pattern is defined in terms of angles θ and ' φ in the local coordinate system. φ) in the local (antenna-fixed) coordinate system and ' is defined as follows: θ' = arccos( cos φ sin θ sin β + cos θ cos β ).z) and a local (“primed”) coordinate system with Cartesian coordinates (x’. the total gain in the global (θ.6.0.Release 9 88 3GPP TR 36.2. thus having its xy-plane coinciding with a horizontal plane. Since the antenna pattern is defined in the local coordinate system. which has been rotated with respect to the global coordinate system.1-1.y. where β is the mechanical tilt angle around the y-axis as defined in Figure A. ' Mechanical tilt is modeled as a rotation of the antenna-fixed coordinate system around the y-axis. and hence beam pointing direction. that is.2. with the local coordinate system antenna-fixed. with the xy-plane parallel to horizontal plane.2 Antenna gain for a given direction and mechanical tilt angle A global coordinate system with Cartesian coordinates (x. By rotating the local coordinate system with respect to the global coordinate system the same effect as mechanical tilting is attained.1-1 Definition of spherical angles and spherical unit vectors in a Cartesian coordinate ˆ are ˆ ˆ system.1.2. Conventionally.2-1.1.6. with the y’-axis being parallel to the y-axis. θ and φ the spherical basis vectors. 3GPP . Note that electric tilt. This transformation relates the spherical angles ( θ ' . For zero mechanical tilt the antenna-fixed coordinate system coincides with the global coordinate system. see Figure A.

As an example.Release 9 89 3GPP TR 36. φ) = Fθ' (θ' .0 (2010-03) A(θ. used below. φ' ) sin ψ + Fφ' (θ' . that is. φ) = Fθ (θ. φ) and FH (θ. the field components F ' and F ' are related to the co-polarized and crossφ θ polarized field components in the local (antenna-fixed) coordinate system for the − 45  polarized antenna as 3GPP . equations (1). For a mechanical tilt angle β.1. φ' ) with θ ' and φgiven by (1) and (2). The notation FH (φ) and FV (φ) . φ) = A' (θ' .814 V9. If ± 45  polarized antennas are used. for θ = 90  .e. φ) . φ) ) are calculated from the field components Fθ' (θ' . in addition to the coordinate transformation. (2) and (5) become θ' = arccos( cos φ sin β ). (defined as FV (θ. φ) = Fφ (θ. For polarized fields a transformation of the field components is needed. φ) = Fθ (θ. φ) = Fφ (θ.6. φ' ) cos ψ where (3) (4) θ ' and φ are defined as in (1) and (2). φ' = arg( cos φ cos β + j sin φ ).2-1 Definition of angles and unit vectors when the local coordinate system has been rotated an angle β around the y-axis of the global coordinate system. φ) = Fθ' (θ' . φ' ) sin ψ FH (θ. should be interpreted as the field components for the specific θ that is given by the direction of the propagation. in the horizontal cut. φ) and FH (θ. the global coordinate system field components FV (θ. the angle between the global z-axis and the line connecting the site antenna and the UE. φ ) of the radiation pattern in the local (antenna-fixed) coordinate system as FV (θ. ψ = arg ( cos β + j sin φ sin β ) . ' z z' β θ' ˆ ˆ φ' φ ˆ n ˆ θ' ˆ n ˆ φ' φˆ ˆ θ y' y ψ ˆ θ' φ' ˆ θ x β x' Figure A. φ' ) ' and Fφ' (θ' . and ψ ' is defined as: (5) ψ = arg ( sin θ cos β − cos φ cos θ sin β + j sin φ sin β ) ..2. i. φ' ) cos ψ − Fφ' (θ' .0.

and nm is the thermal noise at antenna m. an MRC or MMSE.0. This section contains a simple effective IoT measure.2. to the power of the non-desired part of the signal. with M antennas is assumed. The receiver processing constitutes of estimating the transmitted symbol x by appropriately weighting the received signal at each antenna with a complex-valued weight wmH. Turbo SIC) are non-linear receivers whose performance improves with the reliability of the interference reconstruction as the number of iterations increases.1. x is the transmitted symbol. where bold variables are column vectors spanning the antenna domain: y = w H ( hx + e + n ) The effective SINR at the receiver output is given by the ratio of the power of the desired part of the signal.1.2.Release 9 90 3GPP TR 36. The sum-weighted output signal is given by H H y = ∑wm rm = ∑wm ( hm x + em + nm ) m =1 m =1 M M Alternatively. taking the interference suppression capabilities of the receiver into account.1 Advanced receivers modeling Iterative soft interference cancellation receivers Advanced receivers based on iterative soft interference cancellation receivers (e. A linear receiver.g.g.0 (2010-03) Fθ' = 1 ( Fco −45 + Fcross 2 −45 ) . Modeling this interference reconstruction reliability or not in system-level simulations has a significant impact on the accuracy of the performance evaluation.814 V9.1. the interference reconstruction reliability should be modeled. e. In order to ease the comparison of results from different sources. A. for system-level simulations employing a link-to-system abstraction of iterative soft interference cancellation receivers. (6) Fφ' = 1 ( −Fco −45 + Fcross 2 −45 ). the number of iterations should be indicated. wHhx.g. and then summing the weighted signals. Therefore. as this parameter impacts the receiver performance as well as its complexity.7 A.7. the simulation conditions should briefly describe the used modeling method.8 Effective IoT According to the ITU evaluation guidelines [4] an IoT (Interference over Thermal noise) measure reflecting the ‘effective’ interference received by the base station should be used. In addition. wH(e+n) which under the assumption that noise and interference is independent can be written as SINR eff = w H hh H w : w H ( Q + N)w 3GPP . aimed at other users). The received signal at antenna m is rm = hm x + em + nm where hm is the channel between the (single) transmitter and receive antenna m. using vector and matrix notation. A. especially if only a small number of iterations is performed in order to limit the receiver complexity. em is the sum of the interfering signals (e.2. (7) and similarly for the + 45  polarized antenna. No definition of this measure is provided however.

In a first step (1a).0.2. The parameters used are listed in Table A.e. w H Nw Since receive weights typically vary over frequency and time. non-used resources are excluded. In a second step (1c) downlink and uplink spectral efficiencies. user throughput distributions. The denominator in the effective SINR expression now contains the effective power of the interference and thermal noise. the effective IoT should be calculated per frequency and time instant. and SINR distributions for a basic LTE configuration have been evaluated and compared. The average is taken over allocated resources. The effective IoT can now be defined as the ratio between the total effective noise and interference power to the effective thermal noise power IoTeff = PINeff PNeff = w H (Q + N)w . where σ2 is the noise variance per antenna. Hence: PINeff = w H ( Q + N ) w .2-1.0 (2010-03) where Q = E ee H { } is the covariance matrix of the interference. A. and then averaged (i. and I is the identity matrix of size MxM.2 System level simulator calibration To facilitate LTE-A evaluations simulators have been calibrated to ensure that they produce comparable results.814 V9. downlink wideband SINR (also denoted ‘geometry’) and coupling loss distributions for the ITU scenarios and 3GPP case 1 have been evaluated and compared.2. Note that typically N =σ2I. do not average PINeff and PNeff separately). and N = E {nn H } is the covariance matrix of the noise. 3GPP .Release 9 91 3GPP TR 36.

5 wavelength separation at UE. BS antenna downtilt Feeder loss Channel model Intercell interference modeling For step 1a.2135 and 3GPP specifications FDD Synchronized 1dB 1x2 SIMO Round robin with full bandwidth allocation Wideband CQI. Rural macro-cell scenario (Rma): 6 deg Case 1 3GPP 3D: 15 deg Case 1 3GPP 2D: N/A 0dB.2.2. a summary of the results are presented in Figure A. MRC over antennas (no intercell interference rejection) Vertically polarized antennas 0. 10 wavelength separation at basestation Ideal. urban micro-cell scenario (Umi): 12deg ITU Base coverage urban.0 (2010-03) Table A.2-1.814 V9. no PMI on PUCCH (mode 1-0) 5ms periodicity. 6ms delay total (measurement in subframe n is used in subframe n+6) CQI measurement error: None MCSs based on LTE transport formats [5] Maximum four transmissions MRC 1x2 SIMO Frequency Domain Multiplexing – non-channel dependent.0. the results from different simulators are typically within 0. Ideal channel estimate from UL transmission in subframe n can be used for rate adaptation in subframe n+7 MCSs based on LTE transport formats [5] Maximum four transmissions Proponent to specify IR or CC MMSE in frequency domain. Urban macro-cell scenario (Uma): 12deg ITU High speed.Release 9 92 3GPP TR 36. 3GPP . ITU Indoor. With M users and Nrb PRBs available. Parameters for calibration of system level simulators Parameter General Duplex method Network synchronization Handover margin Downlink transmission scheme Downlink scheduler Downlink link adaptation Value Parameters and assumptions not explicitly stated here according to ITU guidelines M. Acknowledgements etc. indoor hotspot scenario (InH): N/A ITU Microcellular.M) users get floor(Nrb/M)+1 PRBs whereas Ml=M-Mh users get floor(Nrb/M) PRBs P0 = -106dBm.0 Based on delayed measurements. both demodulation and sounding LTE: L=3 symbols for DL CCHs.2-1 and are based on averaging independent results from 17 different simulators.5dB of the average SINR. overhead for demodulation reference signals. except for the ITU scenarios in step 1a where a feeder loss of 2dB is used. share available bandwidth between users connected to the cell. all users get resources in every uplink subframe. Mh=mod(Nrb. alpha = 1. M=4 resource blocks for UL CCH. According to ITU for ITU scenarios SCM urban macro high spread for 3GPP case 1 Explicit Downlink HARQ Downlink receiver type Uplink transmission scheme Uplink scheduler Uplink Power control Uplink Link adaptation Uplink HARQ Uplink receiver type Antenna configuration Channel estimation Control Channel overhead. For the downlink wideband SINR.

Distributions of downlink and uplink SINR after antenna combination.Release 9 93 3GPP TR 36.2-2.F. Downlink Calibration 100 90 80 70 100 90 80 70 Uplink Calibration C.2-2.2. A summary is given in Figure A. [%] 60 50 40 30 20 10 0 -140 -120 -100 -80 Coupling gain (Prx-Ptx) [dB] -60 InH UM i UM a RM a Case 13D Case 12D Calibration 100 90 80 70 60 50 40 30 20 10 0 -40 -10 0 10 20 30 Downlink wideband SINR (geometry) [dB] 40 InH UM i UM a RM a Case 13D Case 12D Figure A.2-1. where the results have been obtained by averaging the results of the different simulators.2-2. Distributions of coupling gain and downlink wideband SINR (geometry). Figure A.F. 3GPP . cell-edge user spectral efficiency. normalized user throughput distributions.2.D. [%] C. Results for step 1c in terms of cell spectral efficiency.2. [%] 60 50 40 30 20 10 0 -10 0 10 20 30 40 InH UMi UMa RMa Case 13D Case 12D 60 50 40 30 20 10 0 -5 0 5 10 15 InH UMi UMa RMa Case 13D Case 12D DL Per UE Average SINR [dB] 50 UL Per UE Average SINR [dB] 20 Figure A.D.0 (2010-03) Calibration 100 90 80 70 C.F.F.2.0.2.D. and post antenna combination signal-to-noise-and-interference ratio distributions (with linear averaging over time and frequency) from 16 different simulators were available.2-3 and Table A.814 V9. [%] C.D.

814 V9.2 0.91 0.2.99 0. The intra-site CoMP with up to 3 coordinated co-located cells is the baseline scenario for performance evaluation. The magnitude of this is summarized in Table A.022 0.0.4 0.D. e.3 UMi UMa RMa Case 1 3D Case 1 2D Normalized User Throughput [bps/Hz] 0. For full buffer traffic model.0 0. and clarify the need of enhanced feedback or SRS. directions and metrics.86 0. Spectral efficiencies for calibration in the different environments Direction Downlink Uplink Metric Cell spectral efficiency Cell-edge user spectral efficiency Cell spectral efficiency Cell-edge user spectral efficiency InH 2.F.1 0.026 0.2 0.0 0. [%] 50 40 InH C.0 (2010-03) Downlink Calibration 100 90 80 70 100 90 80 70 Uplink Calibration C.2 0.2-3.035 0.027 0.2-3.F.034 Case 1 3D 1.3 Downlink CoMP evaluation assumptions for intra-NodeB CoMP The objectives of downlink CoMP evaluation for intra-eNodeB CoMP is to clarify the performance gain of CoMP schemes over single-cell schemes. Direction Downlink Uplink Metric Cell spectral efficiency Cell-edge user spectral efficiency Cell spectral efficiency Cell-edge user spectral efficiency InH 3% 9% 5% 13% UMi 9% 19% 7% 16% UMa 8% 17% 5% 15% RMa 6% 14% 5% 11% Case 1 3D 5% 15% 5% 7% Case 1 2D 5% 15% 4% 9% A.3-1. as shown in Figure A.74 0.028 0.2. with both 2 transmit antennas per cell and 4 transmit antenna per cell.D. in terms of coefficients of variation (standard deviation divided by average) for the spectral efficiencies.3 0.1 0.026 RMa 1.5 0.2 0.77 0.5 0.031 A certain spread of the results exists. [%] 60 60 50 40 InH 30 20 10 0 0 0.2.1 0.g.3 0.2.2. 3GPP . Intra-eNodeB CoMP with distributed Radio Remote Heads.Release 9 94 3GPP TR 36. Table A.036 Case 1 2D 1.5 Normalized User Throughput [bps/Hz] Figure A.082 1.084 UMi 1. Table A.68 0. Distributions of downlink and uplink normalized user throughput. Coefficients of variation for the different environments.033 UMa 1. the evaluations include 2 users per cell and 10 users per cell.2.4 UMi UMa RMa Case 1 3D Case 1 2D 30 20 10 0 0.2-2.2-3. Alternative scenario (without X2 interface) can be evaluated.

814 V9.1 eNB: Cross-polarized (0.1.Channel estimation error based on DMRS Should be clarified for each transmission scheme (e.0 (2010-03) Figure A.Release 9 95 3GPP TR 36.0.Feedback scheme and/or UL sounding scheme .3 Evaluation assumptions for IMT-A The discussions during the study item phase have raised a lot of different choices.2.1.Model the impairments of JP-CoMP .5 l spacing) UE: Cross-polarized antennas Config. Clarify in detail the following on CoMP evaluation: .g. a common baseline of the key L1 parameters was proposed.1-2 Follow Annex A 2.1 Table A.1-2 3D as baseline 2D as additional Non-ideal.2..CSI knowledge of eNB .5 l within group. follow ITU evaluation) Uniform distribution Full buffer and Non-full-buffer Possible transmission schemes Network synchronization Antenna configuration Antenna pattern eNB Antenna tilt Channel estimation DL overhead assumption Placing of UEs Traffic model A. In order to have a meaningful set of evaluation results for ITU-R submission.2 eNB: Grouped co-polarized (0.1.1. Channel estimation error based on CRS/CSI-RS . Quantization error .Collision between CRS and PDSCH .Accuracy of CSI .Describe the way to handle the impairments Synchronized Following priority: Config.3 eNB: co-polarized (0.3-1: Baseline evaluation scenario: 3 cell co-located Intra-site CoMP Table A.2. in order that the 3GPP . 20MHz (TDD)  SU-MIMO  MU-MIMO  SU-MIMO with intra-eNB CS/CB  MU-MIMO with intra-eNB CS/CB  SU-MIMO with intra-eNB JP-CoMP  MU-MIMO with intra-eNB JP-CoMP .5 l spacing) UE: co-polarized antennas Follow Annex A 2. 10 l between group) UE: co-polarized antennas Config.2.1 Table A.1.1.Different control regions .3-1-system simulation parameters for downlink intra-site CoMP Evaluation Parameter Deployment scenarios Simulation case System bandwidth Values used for evaluation Homogeneous deployments 3GPP-case1 SCM-UMa (high spread) 10 MHz (FDD).

• The short-term and/or narrowband channel state information (PMI.Release 9 96 3GPP TR 36. SRS).g.0 dB Baseline transmission scheme (LTE Rel.) For baseline transmission scheme (LTE Rel.. etc. FDD: 10+10 MHz except Indoor hotspot with 20+20 MHz TDD: 20 MHz Baseline asymmetry during 5 subframes period: 2 full DL subframes. Baseline (LTE Rel. which can range from a single base station. MLD based receiver Each company should report a description on interference rejection and cancellation capabilities.213]): Baseline: 4x2 MIMO • MIMO single stream beamforming (transmission mode 7 [36.213]) Advanced scheme (LTE-A) • MU-MIMO without coordination • MU-MIMO with intercell coordination • Joint processing CoMP (SU-MIMO is possible for all cases. wideband.1) per PRB Advanced transmission scheme (LTE-A) Following scheme(s) is assumed for the base station(s) to access the channel state information.814 V9.0 (2010-03) results from different parties would not deviate too much so that forming any common view of the expected benefits of LTE-A is less difficult.8) • MIMO closed loop precoded spatial multiplexing (transmission mode 4 [36.) can be obtained. transmit covariance matrix. subband CQI with measurement error: N(0. to all UEs served by the CoMP cooperating set. and/or by feedback with the average feedback overhead to be described in the evaluation results. Duplex method and bandwidths Network synchronization Handover margin Downlink transmission scheme Downlink scheduler Downlink link adaptation Downlink HARQ scheme Downlink receiver type 3GPP ..2135].0.8): Proportional fair in time and frequency For advanced transmission scheme (LTE-A) Aligned with transmission scheme Non-ideal based on non-ideal CQI/PMI/RI reports and/or non-ideal sounding transmission. Incremental redundancy or Chase combining Baseline scheme MMSE Advanced scheme MMSE-SIC. Synchronized 1.3-1 Simulation assumption for Cell/cell-edge spectrum efficiency Parameter Deployment scenario Values used for evaluation • Indoor hotspot • Urban micro-cell • Urban macro-cell • Rural macro-cell Parameters and assumptions not shown here for each scenario are shown in ITU guidelines [ITU-R Report M. These parameters are not exclusive or restrictive to additional results. 2 full UL subframes Alternative special subframe configurations may be used if stated. Table A. frequency selective CQI report with 5ms periodicity. subband CQI with measurement error: N(0. spatial channel information such as transmit channel covariance matrices or angular information can be obtained. reporting mode: and period selected according to scheduler and MIMO transmission schemes. UpPTS 2 symbol. from uplink measurements (based on e. for example. channel transfer function.1) per PRB B) Sounding-based precoding. Special subframe: DwPTS 11symbol.8): A) Non-frequency selective PMI and frequency selective CQI report with 5ms periodicity. reporting delay and MCS based on LTE transport formats according to [36. to multiple base stations.213]. GP 1 symbol. SRS) or by feedback with an average feedback overhead similar to that in Rel-8. from uplink measurements (based on e. • The long-term.g. for example.

0 (2010-03) Uplink scheduler Uplink Power control Uplink link adaptation Uplink HARQ scheme Uplink receiver type Antenna configuration base station Baseline: Fractional power control. (IoT reported with simulation results.5 wavelengths between antennas (illustration for 4 Tx: |||| ) D) Uncorrelated cross-polarized: Columns with +-45deg linearly polarized antennas Columns separated 4 wavelengths (illustration for 4 Tx: X X) E) Correlated cross-polarized Columns with +-45deg linearly polarized antennas Columns separated 0. Incremental redundancy or Chase combining MMSE or MMSE-SIC (MU-MIMO) Baseline: 4 or 8 Tx antennas with the following configurations: A) Uncorrelated co-polarized: Co-polarized antennas separated 4 wavelengths (illustration for 4 Tx: | | | |) B) Grouped co-polarized: Two groups of co-polarized antennas.Release 9 Uplink transmission scheme 97 Baseline transmission scheme (LTE Rel.) Non-ideal based on delayed SRS-based measurements: MCS based on LTE transport formats and SRS period and bandwidths according to [36.8) • Urban Micro: C or E (MU-MIMO).8) • Urban Macro: C or E (MU-MIMO). Alternative: Other Rel. 10 wavelengths between center of each group. 0.213].5 wavelengths separation at UE Alternative: Columns with linearly polarized orthogonal antennas with 0. C or E (CoMP) • Rural Macro: C or E (MU-MIMO. Rel.5 wavelengths (illustration for 8Tx: XXXX) Baseline mappings between deployment scenario and antenna configurations: For downlink: • Indoor Hotspot: A (Rel.8) • SIMO with and without MU-MIMO Baseline: 1 x 4 SIMO Advanced transmission scheme (LTE-A) • SU-MIMO • UL CoMP Channel dependent 3GPP TR 36.814 V9.8) Note: MU-MIMO = MU-MIMO with / without coordination CoMP = Joint processing CoMP For uplink: • Indoor Hotspot: A • Urban Micro: A or B or C or E • Urban Macro: A or C or E • Rural Macro: A or C or E Baseline: Vertically polarized antennas with 0.5 wavelengths spacing between columns Recommended: Non-ideal (For non-ideal case. Rel.8 specified Power control parameters (P0 and alpha) are chosen according to the deployment scenario.0. B or E (CoMP.5 wavelength separation within each group (illustration for 4 Tx: || || ) C) Correlated: co-polarized: 0. consider both estimation errors both for demodulation reference signals and sounding reference signals) Antenna configuration UE Channel estimation (Uplink and downlink) 3GPP .

Antenna geometries and radiation patterns can be defined properly by the user of the model. 3GPP . NEC. CATT. Ericsson/ST-Ericsson.4 Detailed simulation results Detailed simulation results of cell spectral efficiency and cell edge spectral efficiency for full-buffer traffic in Section 10. Semi-persistent Included.0 (2010-03) For baseline transmission (LTE) schemes: • Overhead for CRS and antenna port 5 according to DL transmission schemes • SRS overhead according to UL (and DL) scheduler and transmission scheme • Overhead for UL CCH according to CQI/PMI reporting mode and periodicity used for DL simulation for the same scenario • Overhead for DL CCH of [1-3] OFDM symbols For advanced transmission schemes (LTE-A) • UL overhead for CSI feedback and SRS transmission according to transmission scheme factored into the uplink results for the same scenarios. Acknowledgements etc. excluded Baseline value 4 Both Included Configuration 0 and 1. LGE. Channel realizations are generated through the application of the geometrical principle by summing contributions of rays (plane waves) with specific small-scale parameters like delay.8). like the well-known spatial channel model (SCM) [B1]. Superposition results to correlation between antenna elements and temporal fading with geometry dependent Doppler spectrum. VoIP parameters and recommended baseline values Parameter Number of base station antennas Scheduling PDDCH limitation TDD configuration Range 2-8 Dynamic. The following 18 corporate entities (listed below alphabetically) participated in these simulations: Alcatel-Lucent/Alcatel-Lucent Shanghai Bell. Nokia/Nokia Siemens Networks. UpPTS 1 A. Texas Instruments. • DL overhead for additional DRS if used.2. RITT. P0 fitted to environment 4-5 RB Table A. CMCC. Geometry-based modelling of the radio channel enables separation of propagation parameters and antennas.3-3. but rather the directions of the rays. The channel parameters for individual snapshots are determined stochastically based on statistical distributions extracted from channel measurements. power. GP 1. Qualcomm. Mobility parameters and recommended baseline values Parameter Power control Scheduling bandwidth Baseline value Fractional (α = 0.Release 9 Control channel and reference signal overhead. It does not explicitly specify the locations of the scatters.0. DwPTS 12. It can also be called double directional channel model. Huawei. Hitachi. angle-of-arrival (AoA) and angle-of-departure (AoD). ZTE Annex B: Channel models (Informative) The ITU-R IMT-Advanced channel model is a geometry-based stochastic model.814 V9. Fujitsu. Samsung. VoIP capacity and mobility are contained in [6]. 98 3GPP TR 36. None Feedback and control channel errors Table A. Panasonic. Motorola.3-2. NTT DOCOMO.

. which is called generic model.m ⋅ rtx . φ m is the AoD unit vector.τ ) = ∑H n ( t . indoor.m ) n ( ) ( ) (5) where Frx. s .m ⋅ rrx .m )  αn .V and Frx.m. Elements of the MIMO channel. either or both in delay and angle domains. The IMT-Advanced channel model for the evaluation of IMT-Advanced candidate RITs consists of a Primary Module and an Extension Module.m is the Doppler frequency vector.1 rtx . only the primary module is included. i. antenna arrays at both link ends and propagation paths. s and rrx .u . Figure B-1 The MIMO channel Array 2 Array 1 (S Tx elements) rrx .0. e.m t )δ (τ −τ n .τ ) = ∫∫Frx (ϕ) h n ( t .g. and the dual-polarized propagation channel response matrix hn for cluster n as follows T H n ( t . m )   n .g.s ) 0 0 M T × exp ( j 2π υ . H (ϕn .n ( t .m . which is mandatory for IMT-Advanced evaluation.H are the antenna element u field patterns for vertical and horizontal polarizations respectively.u .s . If the radio channel is modelled as dynamic. The generic MIMO channel model is applicable for all scenarios. are illustrated in Figure B-1. ϕ)Ftx (φ) dφdϕ .τ ) = ∑   α a n .m. H (φn . HH Ftx . e. In this part.u. The channel from Tx antenna element s to Rx element u for cluster n is expressed as (4)  Frx . and ν n.m . they are functions of t [B8].m .m ) m =1 Frx . urban and rural. ϕ m is the AoA unit n.V (ϕn . component of ray n.τ ) . n =1 N (3) where t is time.u ) exp j 2π λ (φn .1  rtx.m respectively.S  rrx. rtx .0 (2010-03) A number of rays constitute a cluster.V (φn .τ. and n is path index.u. τ is delay. all the above mentioned small-scale parameters are time variant.VV αn .u are the location vectors of element s and u respectively. λ 0 is the wave length of the carrier frequency.VH  Ftx . The Primary Module covers the mathematical framework. a set of parameters as well as path loss models.m .814 V9..m. N is the number of paths.U Path 1 (U Rx elements) Path N The time variant impulse response matrix of the U x S MIMO channel is given by H ( t .VH are the complex gains of vertical-to-vertical and horizontal-to-vertical polarizations of ray n. 3GPP . In the terminology of this document we equate the cluster with a propagation path diffused in space.m )  H u .Release 9 99 3GPP TR 36.VV and α n. s . n. α n. φ. HV   −1 −1 × exp j 2π λ (ϕn .e. It is composed of the antenna array response matrices Ftx and Frx for the transmitter (Tx) and the receiver (Rx) respectively.

such as the spectrum. It is a system level 3 model in the sense that is employed. Then. the propagation effects are divided into three distinct types: These are the path loss.1. At this stage the geometric setup is fixed and the only free variables are the random initial phases of the scatters. The generic model allows the user to simulate over several drops to get statistically representative results.1. there is no further randomness left. 3 The term system-level means here that the model is able to cover multiple links. powers. the slow variation due to shadowing and scattering. In the third stage. the powers. Scenario selection -urban macro -urban micro -indoor -out2in -etc. A drop (or snapshot or channel segment) is a simulation entity where the random properties of the channel remain constant except for the fast fading caused by the changing phases of the rays.Release 9 100 3GPP TR 36.. First. and the rapid variation in the signal due to multipath effects.1 Channel models This section provides the reference channel model for each test environment. the simulation of the system behaviour is carried out as a sequence of “drops”. When the initial phases are also fixed. When using the generic model. with one mathematical framework by different parameter sets.814 V9. K -XPR -shadowing -path loss ChIR generation Network Layout Multi-path parameters -power. including the path loss models. large-scale and small-scale parameters are defined. B. where a “drop” is defined as one simulation run over a certain time period. the propagation scenario is selected. Next. In the second stage. It can describe an unlimited number of propagation environment realizations for single or multiple radio links for all the defined scenarios and for arbitrary antenna configurations. Figure B-2 Channel model creation process User defined parameters Propagation parameter generation Large scale parameters -DS. By picking (randomly) different initial phases. 3GPP . Consecutive drops are independent. These test environments are intended to cover the range of IMT-Advanced operating environments. AS.g. and directions of the rays. The generic model is based on the drop concept. in the SCM model [B1]. an infinite number of different realizations of the model can be generated. delay. coverage and power efficiencies.2 Void Primary module The following sections provide the details of the channel models. e.g. cells and terminals. The test operating environments are considered as a basic factor in the evaluation process of the candidate RITs. The generic channel model is a stochastic model with two (or three) levels of randomness. and directions of arrival and departure are drawn randomly according to tabulated distribution functions and random LS parameters.. delay. small-scale (SS) parameters like delays. For terrestrial environments. Figure B-2 shows the overview of the channel model creation. for the terrestrial component. The constant properties during a single drop are. the network layout and the antenna configuration are determined. The first stage consists of two steps. First.1 B.0 (2010-03) The generic channel model is a double-directional geometry-based stochastic model. AoA. channel impulse responses (ChIRs) are calculated. large-scale (LS) parameters like shadow fading. In a simulation the number and the length of drops have to be selected properly by the evaluation requirements and the deployed scenario. e. and angular spreads are drawn randomly from tabulated distribution functions. etc.0. AoD. delays. -BS & MS locations -velocities Channel coefficient generation ChIR Antennas -# elements -orientations -field patterns B. The reference models are used to estimate the critical aspects.

The rural path-loss formula can be applied to the desired frequency range from 450 MHz to 6 GHz.7 + 26log10(fc) σ =4 10 m < d < 2 000 m hBS = 10 m hUT =1-2. d l ) = PL LOS (d k ) + 17. Table B.l ∈ {1.1 Path loss models Path loss models for the various propagation scenarios have been developed based on measurement results carried out in [B2.5 m 3GPP .d2 ) 2) w = 20 m (street width) hBS = 10 m.0log10(d) + 28. The path loss models have been summarized in Table B.0. d1 ) ) where σ =4 10 m < d1 + d2 < 5 000 m.5 m 10 m < d < 150 m hBS = 3-6 m hUT = 1-2. and its standard deviation for each scenario is given in the table. w/2 < min(d1.5 m When 0 < min(d1. PL ( d k . 1. the channel model can be used for lower frequencies down to 450 MHz.1. The channel models also cover MIMO aspects as all desired dimensions (delay.1.1-1 Summary table of the primary module path loss models Scenario Path loss [dB] Note: fc is given in GHz and distance in meters! Urban Micro (UMi) Indoor Hotspot (InH) PL = 16. Note that the distribution of the shadow fading is log-normal.8 − 0.1.2. B7. antenna fading height default values std [dB] σ =3 3 m < d < 100 m hBS =3-6 m hUT =1-2.9log10(d) + 32.5n j + 10 n j log 10 (d l ) + 3 log 10 ( f c ) and n j = max ( 2.1-1. AoD and polarisation) are considered. B9-B14].3log10(d) + 11. hUT = 1.84 ) and PLLOS is the path loss of scenario UMi LOS and k.7log10(d) + 22.5 m1) σ =4 PL = 22. The models can be applied in the frequency range of 2 – 6 GHz and for different antenna heights.5 m 10 m < d1 < d’BP 1) d’BP < d1 < 5000 m1) hBS = 10 m1). PL (d 2 .0024 d k . For the rural macro-cell scenario (RMa). d 2 ). as well as results from the literature. AoA.2}.0 (2010-03) The channel models are specified in the frequency range from 2 GHz to 6 GHz.2.8 – 18log10(h’BS) –18log10(h’UT) + 2log10(fc) NLOS Manhattan grid layout: PL = min ( PL ( d1 .Release 9 101 3GPP TR 36.9 − 12.814 V9.5 + 20log10(fc) NLOS Shadow Applicability range. The channel models are targeted for up to 100 MHz RF bandwidth.0 + 20log10(fc) σ =3 σ =3 LOS PL = 40log10(d1) + 7.8 + 20log10(fc) LOS PL = 43. hUT = 1. B. Hexagonal cell layout: PL = 36.d2 ) < w/2 .2. the LOS PL is applied.

77) + 0.4. Explanations: see 3) σ =4 10 m < d < d’BP 1) LOS PL = 40 .10)log10(d) – min(0. hBS=10m. nFl=1.814 V9.0 + 20log10(fc)  P L = P L1 (d o u t+ d in ) b B  2  P L = 1 4+ 1 5(1 − c o sθ () ) tw  P L = 0. W = 20 m.14.03h1.0log10(d) + 28.1 log10 (W) + 7.0 log 10 (d1 ) + 7.5 m1) Urban Macro (UMa) PL = 161.7(h/hBS)2) log10 (hBS) + (43.5d in  in 10 m < dout+din< 1 000 m.002log10(h)d LOS PL2 = PL1 (dBP) + 40 log10(d/dBP) σ=4 10 m < d < dBP 4) σ=6 NLOS PL = 161.1 log10 (W) + 7.0 log 10 (h' BS ) σ = 4 −18 .0 log 10 ( f c ) d’BP < d < 5 000 m1) hBS = 25 m1). hUT = 1.97) σ =8 dBP < d < 5 000 m hBS = 35 m. hUT = 1. hUT = 1.2 (log10 (11.0 (2010-03) PL = PL b + PL tw + PL in Manhattan grid layout (θ known): σ =7 O-to-I For hexagonal layout (θ unknown): PLtw = 20. hBS. W. hUT are same as in UMa NLOS) 10 m < d < 5 000 m hBS = 35 m.4. W = 20 m.5m. W. PL = 22.0 log 10 (h'UT ) + 2.75 hUT)) 2 . 0 m < din< 25 m. other values remain the same. hUT=3(nFl -1)+1. h = 20 m The applicability ranges: 5 m < h < 50 m 5 m < W < 50 m 10 m < hBS < 150 m 1 m < hUT < 10 m Suburban Macro (SMa. optional) PL1 = 20log10(40π dfc /3) + min(0.97) NLOS σ =6 10 m < d < 5 000 m h = avg.42 – 3. W = 20 m.2 (log10 (11.7(h/hBS)2) log10 (hBS) + (43. h = 10 m (Applicability ranges of h.044h1.1 log10 (hBS)) (log10 (d)-3) + 20 log10(fc) – (3.5 m.5 m.37 – 3.72. building height W = street width hBS = 25 m. h = 10 m (The applicability ranges of h.5 log10 (h) – (24.8 −18 . hBS.04 – 7.37 – 3.72.Release 9 102 3GPP TR 36.75 hUT)) 2 .5 log10 (h) – (24.1 log10 (hBS)) (log10 (d)-3) + 20 log10(fc) – (3.5 m. hUT = 1.0.42 – 3. hUT are same as in UMa NLOS) 3GPP .04 – 7.

where fc is the centre frequency in Hz.0× 108 m/s is the propagation velocity in free space. 4) Break point distance dBP = 2π hBS hUT fc/c. 3) PLb = basic path-loss. d1 is the distance 3GPP .2. dout = distance from BS to the wall next to UT location.0. hUT are same as UMa NLOS) PL = 161. where fc is the centre frequency in Hz.1-1. The effective antenna heights h’BS and h’UT are computed as follows: h’BS = hBS – 1. PLtw = Loss through wall.1) ⋅ (1 − exp(− d / 63) ) + exp(− d / 63) PLOS 1.044h1. 2) The distances d1 and d2 are defined below in Figure B. hBS = 35 m. hUT = 1.2.1) ⋅ (1 − exp(− d / 36) ) + exp(− d / 36) (for outdoor users only) UMa PLOS = min(18 / d . W.7(h/hBS)2) log10 (hBS) + (43.5 log10 (h) – (24.04 – 7.1. respectively. h = 5 m (The applicability ranges of h.1-2 Scenario LOS probability as a function of distance d [m] InH PLOS d ≤ 18 1.  = exp( − ( d − 18) 27 ) . din = perpendicular distance from wall to UT (assumed evenly distributed between 0 and 25 m). as illustrated in Figure B. 4) PL2 = PL1 (dBP) + 40 log10(d/dBP) σ =6 dBP < d < 10 000 m. PL B1 = Loss of UMi outdoor scenario.0 m.77) + 0. and the effective environment height in urban environments is assumed to be equal to 1.42 – 3.72. hBS.1.2. c = 3.Release 9 103 3GPP TR 36.37 – 3. hBS. These distances are defined with respect to a rectangular street grid. where the UT is shown moving along a street perpendicular to the street on which the BS is located (the LOS street). h’UT = hUT – 1. where hBS and hUT are the actual antenna heights.2 (log10 (11.14.4.2.0× 108 m/s is the propagation velocity in free space. c = 3. W = 20 m. PLin = Loss inside. θ = angle between LOS to the wall and a unit vector normal to the wall. Table B. respectively. hUT are same as UMa NLOS) 1) Break point distance d’BP = 4 h’BS h’UT fc/c.5 m.0 (2010-03) PL1 = 20log10(40π dfc /3) + min(0. W.814 V9.1 log10 (hBS)) (log10 (d)-3) + NLOS 20 log10(fc) – (3.10)log10(d) – min(0.1. d ≥ 37  UMi PLOS = min(18 / d . W = 20 m. d > 10  1000  RMa The NLOS path loss model for scenario UMi is dependent on two distances.1 log10 (W) + 7. d1 and d2 in the case of the Manhattan grid. h = 5 m (Applicability ranges of h. The line-of-sight (LOS) probabilities are given in Table B.1-1.1. hUT = 1. d ≤ 10 =  d − 10  exp − .97) σ =8 10 m < d < 5 000 m.1-2.0 m. and h’BS and h’UT are the effective antenna heights at the BS and the UT.72. Note that probabilities are used only for system level simulations.5. hBS = 35 m.75 hUT)) 2 . and hBS and hUT are the antenna heights at the BS and the UT.002log10(h)d LOS Rural Macro (RMa) σ =4 10 m < d < dBP.5 m. 18 < d < 37 0.0 m.03h1.

Release 9

104

3GPP TR 36.814 V9.0.0 (2010-03)

from the BS to the centre of the perpendicular street, and d2 is the distance of the UT along the perpendicular street, measured from the centre of the LOS street.

Figure B.1.2.1-1 Geometry for d1 - d2 path-loss model
+

MS

d2

d1 d2

BS

-

B.1.2.1.1

Autocorrelation of shadow fading

The long-term (log-normal) fading in the logarithmic scale around the mean path loss PL (dB) is characterized by a Gaussian distribution with zero mean and standard deviation. Due to the slow fading process versus distance ∆ x, adjacent fading values are correlated. Its normalized autocorrelation function R(∆ x) can be described with sufficient accuracy by the exponential function [B6]

R ( ∆ )= x e

∆ x − d cor

(6)

with the correlation length dcor being dependent on the environment, see the correlation parameters for shadowing and other large scale parameters in Table B.1.2.2.1-4 (Channel model parameters).

B.1.2.2
B.1.2.2.1

Primary module channel model parameters
Generic model

The radio channels are created using the parameters listed in Table B.1.2.2.1-4. The channel realizations are obtained by a step-wise procedure [B2] illustrated in Figure B.1.2.2.1-1 and described below. It has to be noted that the geometric description covers arrival angles from the last bounce scatterers and respectively departure angles to the first scatterers interacted from the transmitting side. The propagation between the first and the last interaction is not defined. Thus, this approach can model also multiple interactions with the scattering media. This indicates also that e.g., the delay of a multipath component cannot be determined by the geometry. In the following steps, downlink is assumed. For uplink, arrival and departure parameters have to be swapped. Figure B.1.2.2.1-1 Channel coefficient generation procedure

3GPP

Release 9

105
General parameters :

3GPP TR 36.814 V9.0.0 (2010-03)

Set scenario, network layout and antenna parameters

Assign propagation condition (NLOS/ LOS)

Calculate path loss

Generate correlated large scale paramters (DS, AS, SF, K)

Small scale parameter:

Perform random coupling of rays

Generate arrival & departure angles

Generate cluster powers

Generate delays

Coefficient generation :
Draw random initial phases Generate channel coefficient Apply path loss & shadowing

General parameters: Step 1: Set environment, network layout, and antenna array parameters Choose one of the scenarios (InH, UMi, …) Give number of BS and UT Give locations of BS and UT, or equally distances of each BS and UT and relative directions and φ LOS and ϕ LOS of each BS and UT d. Give BS and UT antenna field patterns Frx and Ftx and array geometries e. Give BS and UT array orientations with respect to north (reference) direction f. Give speed and direction of motion of UT g. Give system centre frequency Large scale parameters: Step 2: Assign propagation condition (LOS/NLOS). Step 3: Calculate path loss with formulas of Table B.1.2.1-1for each BS-UT link to be modelled. Step 4: Generate correlated large scale parameters, i.e. delay spread, angular spreads, Ricean K factor and shadow fading term like explained in [B2, section 3.3.1] (Correlations between large scale parameters). Limit random rms arrival and departure azimuth spread values to 104 degrees, i.e., σ ϕ = min(σ ϕ ,104°). Small scale parameters: Step 5: Generate delays τ . Delays are drawn randomly from the delay distribution defined in Table B.1.2.2.1-4. With exponential delay distribution calculate a. b. c.

τ n ' = −rτ στ ln ( X n ) ,

(7)

where rτ is the delay distribution proportionality factor, Xn ~ Uni(0,1), and cluster index n = 1,…,N. With uniform delay distribution the delay values τ n’ are drawn from the corresponding range. Normalise the delays by subtracting the minimum delay and sort the normalised delays to descending order:

τ n = sort (τ n '− min (τ n ') ) .

(8)

In the case of LOS condition, additional scaling of delays is required to compensate for the effect of LOS peak addition to the delay spread. The heuristically determined Ricean K-factor dependent scaling constant is

D = 0.7705 − 0.0433 K + 0.0002 K 2 + 0.000017 K 3 ,

(9)

3GPP

Release 9

106

3GPP TR 36.814 V9.0.0 (2010-03)

where K [dB] is the Ricean K-factor defined in Table B.1.2.2.1-4.. The scaled delays
LOS τn =τn / D ,

(10)

are not to be used in cluster power generation. Step 6: Generate cluster powers P. Cluster powers are calculated assuming a single slope exponential power delay profile. Power assignment depends on the delay distribution defined in Table B.1.2.2.1-4. With exponential delay distribution the cluster powers are determined by
− Ζn  rτ − 1   ⋅ 10 10 P = exp − τ n  rτ σ τ    ' n

(11)

where Ζ n ~ N(0, ζ ) is the per cluster shadowing term in [dB]. Average the power so that the sum power of all cluster powers is equal to one, i.e.,

Pn =

Pn' P' n =1 n
N

(12)

Assign the power of each ray within a cluster as Pn / M, where M is the number of rays per cluster. Remove clusters with less than -25 dB power compared to the maximum cluster power. Step 7: Generate arrival angles ϕ and departure angles φ . As the composite PAS of all clusters is modelled as wrapped Gaussian (see Table B.1.2.2.1-4), except indoor hotspot scenario (InH) as Laplacian, the AoAs are determined by applying the inverse Gaussian function (13) or inverse Laplacian function (14) with input parameters Pn and RMS angle spread σ ϕ

ϕn ' =

2σ AoA − ln ( Pn max ( Pn ) ) . C

(13) (14)

ϕn ' = −

In the equation (13) σAoA =σϕ 1.4 is the standard deviation of the arrival angles (the factor 1.4 is the ratio of Gaussian std and the corresponding “RMS spread”). Constant C is a scaling factor related to total number of clusters and is given in the table below: Table B.1.2.2.1-1 # cluster s C 4 5 8 10 11 12 14 15 15 (InH) 1.434 16 19 19 (InH) 1.501 20

σ ϕ ln ( Pn max ( Pn ) ) C

0.77 9

0.86 0

1.01 8

1.09 0

1.12 3

1.14 6

1.19 0

1.21 1

1.22 6

1.27 3

1.289

In the LOS case, constant C is dependent also on the Ricean K-factor. Constant C in (13) and (14) is substituted by CLOS. Additional scaling of the angles is required to compensate for the effect of LOS peak addition to the angle spread. The heuristically determined Ricean K-factor dependent scaling constant is

C LOS = C ⋅ 1.1035 − 0.028 K − 0.002 K 2 + 0.0001 K 3 ,
As for indoor hotspot scenario, the scaling constant is:

(

)

(15)

3GPP

….0.2. LOS Step 10a: Generate channel coefficients for each cluster n and each receiver and transmitter element pair u. and uniform linear arrays (ULA).2.2.814 V9.2.2. substitute (17) by (18) to enforce the first cluster to the LOS direction ϕ ϕn = ( X nϕn '+Yn ) − ( X 1ϕ1 '+Y1 − ϕLOS ) .20 Basis vector of offset angles (19) α m ± 0.1-2).hv.8 9.N.8844 ± 1. Φhvm . where cAoA is the cluster-wise rms azimuth spread of arrival angles (cluster ASA) in Table B.5195 ± 2.1.m =ϕn +c AoA αm .Release 9 107 3GPP TR 36.s.2.0 (2010-03) C LOS = C ⋅ 0.4 5.9275 + 0. For the N – 2 weakest clusters.1.2. or within a sub-cluster in the case of two strongest clusters (see Step 10a and Table B. m vh . In the LOS case.1481 ± 1.vh.2.0439 K − 0.2.1-2 to the cluster angles ϕn . Φ hh } for both VV and HH polarisations.m to arrival ray angles ϕ n.3715 ± 0.1-2 Ray offset angles within a cluster.4. In the LOS case.0002 K 3 where K [dB] is the Ricean K-factor defined in Table B. ( ) (16) Assign positive or negative sign to the angles by multiplying with a random variable Xn with uniform distribution to the discrete set of {1.–1}.0071 K 2 + 0. draw also random initial phases {Φ vv LOS . m .0447 ± 0. given for 1 rms angle spread Ray number m 1.1. Finally add offset angles α m (18) from Table B. Couple randomly departure ray angles φ n.16 17.12 13. the channel coefficients are given by: 3GPP .1-4.18 19. n. Φn .2492 ± 0.5129 ± 0.1.1.10 11. π ). σ ϕ 7 to introduce random variation ( ) ϕn = X nϕ '+ n +ϕLOS .2.hh). } different polarisation combinations (vv. and add component Yn ~ N 0.6 7.m within a cluster n. say n = 3.1413 ± 0.1-4.1551 For departure angles φ n the procedure is similar. Step 8: Random coupling of rays within clusters. Table B. Φhhm for each ray m of each cluster n and for four n. see Step1c.14 15.2 3. Y n where ϕ LOS (17) LOS is the LOS direction defined in the network layout description. The distribution for initial phases is uniform within (-π .6797 ± 0. Coefficient generation: Step 9: Draw random initial phase {Φ vv n.

and λ 0 is the wavelength of the carrier frequency.m   Ftx .  −1   hv hh exp jΦn .3.u. The channel coefficients are given by: Hu. n and determine the channel coefficients by adding a single line-of-sight ray and scaling down the other channel coefficient generated by (20). with fixed delay offset {0. If polarisation is not considered.2. 3 =τn +10 ns (22) Twenty rays of a cluster are mapped to sub-clusters as presented in Table B.10 ns} (see Table B.2. s .2.2.6.s .u.1-2 with mapping of Table B.18 13.1.u .H (φn .n ( t ) = 1 H 'u .1-2. 2 =τn + 5 ns τn .2. H ( ϕ LOS )   0 exp jΦ LOS   Ftx . .1 =τn + 0 ns τn .V (φn . s . ( ) ( ) ( ) (23) If non-ULA arrays are used. (21) For the two strongest clusters.u .1.2.10.2. u . rays are spread in delay to three sub-clusters (per cluster).m m =1  κ exp jΦn .2.2.17.8.H are the antenna element u field patterns for vertical and horizontal polarisations respectively.2. s .12.s . κ is the cross polarisation power ratio in linear scale.15.5.1.m )  n .1. ( ) The Doppler frequency component is calculated from the angle of arrival (downlink). the distance term du in equations (23) and (20) is replaced by: 3GPP . say n = 1 and 2. n =H u .1-3 Sub-cluster information for intra cluster delay spread clusters sub-cluster # 1 2 3 mapping to rays 1. Table B. s .1-3 below.mt ) 0 0 n M T  Frx .0.m −θv ) λ0 . n ( t ) KR + 1 T vv   Ftx .1-3).m ) exp ( j 2π υ. define H 'u .H (ϕn .m )  ( ( ) ( ( ) ) ( ( ) ) ) (20) where Frx. The delays of the sub-clusters are τn .V ( φLOS )  0 K R  Frx .s.11.V ( ϕ LOS )  exp jΦ LOS + δ ( n − 1)  F    hh K R + 1  rx .V and Frx.14.0 (2010-03) H u .2.19.s .5.n ( t ) = Pn ∑  exp jΦvv κ −1 exp jΦvhm   Ftx .Release 9 108 3GPP TR 36.4. The corresponding offset angles are taken from Table B.m )  F   rx .1.m )    − − ⋅ exp jd s 2π λ1 sin (φn . see Figure B. s .m = v cos (ϕn . UT speed v and direction of travel θ v υn . H ( φLOS )  − − ⋅ exp jd s 2π λ1 sin ( φLOS ) exp jd u 2π λ1 sin ( ϕ LOS ) exp( j 2π υ t ) 0 0 LOS ( ) where δ ( ) is the Dirac’s delta function and KR is the Ricean K-factor defined in Table B.1-3.2.814 V9.m n.1.2.2 0 9.V (ϕn . the equations must be modified. ds and du are the uniform distances [m] between transmitter elements and receiver elements respectively.16 power 10/20 6/20 4/20 delay offset 0 ns 5 ns 10 ns In the LOS case.m and only vertically polarised n field patterns are applied.7. the 2x2 polarisation matrix can be replaced by the scalar exp jΦ . For arbitrary array configurations on the horizontal plane.1-4 converted to linear scale.2.u .1.m ) exp jd u 2π λ1 sin (ϕn .2.

m ) . y Au yu d u'dn .yu) are the co-ordinates of the uth element Au and A0 is the reference element.1-2 Modified distance of antenna element u with non-ULA array Step 10b: Generate channel coefficients for each cluster n and each receiver and transmitter element pair u. Alternatively to Step 10a channel coefficients can be generated by applying the so called correlation matrix based method. Step 11: Apply path loss and shadowing for the channel coefficients.’ . 3GPP . The MIMO covariance matrix can be composed of transmitter and receiver spatial correlation matrices.2. This is valid only for system level simulations. Spatial correlation matrices can be derived for each cluster directly from the ChIRs of Step 10a [B3.1. and by the polarization covariance matrices with the Kronecker product of matrices.1-4 [B5]. Alternatively. sin ϕn . m = xu + yu cos ( arctan ( yu x u ) − ϕn . m (24) where (xu.s. MIMO antenna correlations are introduced by performing a linear transformation to the temporally correlated sequences. m u A0 ϕn.1. n.0.2. Unique covariance matrices are determined for each drop and radio link based on the parameters defined in the previous steps. AoD) determined in Step7 with XPR parameter from Table B. spatial correlation matrices and the polarization covariance matrices can be derived from the antenna configuration and the model parameters (AoA.2.814 V9.2. B4].0 (2010-03) ' du .Release 9 2 2 109 3GPP TR 36.m xu x Figure B. Temporal correlations (Doppler effect) are introduced either by filtering independent and identically distributed complex Gaussian sequences with a proper Doppler spectrum shaping filter or by applying some equivalent method.

75 0.5 -0.39 1.25 1.5 0.44 0.3 8 7 12 20 20 20 5 2 11 15 3 3 30 18 15 37 12 40 50 50 50 N/A 7 N/A 4 N/A 0 -0.41 0.19 3 9 5 0.95 0.28 1.62 0.16 7 N/A N/A 0.76 0.24 0.4 0 -0.32 1.4 -0.6 -0.17 1. Scenarios Delay spread (DS) log10([s]) AoD spread (ASD) log10([°]) AoA spread (ASA) log10([°]) Shadow fading (SF) [dB] K-factor (K) [dB] InH µ σ µ σ µ σ σ LOS -7.84 0.7 0 N/A N/A N/A N/A Exp O-to-I -6.7 12 7 11 10 20 20 2 2 3 3 3 3 50 25 35 37 40 36 30 40 120 N/A Table B.4 0.48 0.41 0.2 9 12 20 3 17 3 7 8 8 10 15 3 10 19 20 5 11 3 5 3 3 6 N/A 3 8.5 0 N/A N/A N/A N/A Exp LOS -7.43 1.87 0.Release 9 110 3GPP TR 36.4 0 0.2 -0.11 6 RMa LOS NLOS -7.4 0.8 0.62 0.70 0.13 4 8 µ σ ASD vs DS ASA vs DS ASA vs SF ASD vs SF DS vs SF Cross-Correlations * ASD vs ASA ASD vs Κ ASA vs Κ DS vs Κ SF vs Κ Delay distribution AoD and AoA distribution Delay scaling parameter rτ XPR [dB] µ Number of clusters Number of rays per cluster Cluster ASD Cluster ASA Per cluster shadowing std ζ [dB] Correlation distance [m] DS ASD ASA SF Κ Laplacian 3.2.5 2.15 0.4 0 0 -0.3 -0.0.8 -0.16 4 N/A N/A 0.814 V9.2 N/A -0.8 -0.1-5 Expectation (median) output values for large scale parameters Scenario LOS InH NLOS LOS UMi NLOS O-to-I LOS UMa NLOS LOS RMa NLOS DS (ns) 20 39 65 129 49 93 365 32 37 AS at BS (º) 40 42 16 26 18 14 26 8 9 AS at UT (º) 42 59 56 69 58 65 74 33 33 3GPP .1-4 Channel model parameters • • In the table below DS = rms delay spread. The sign of the shadow fading is defined so that positive SF means more received power at UT than predicted by the path loss model.54 1.41 0.40 1.6 -0.38 0.4 -0.2.0 (2010-03) Table B.2.6 11 15 20 5 8 6 8 7 5 10 4 Wrapped Gaussian 3.52 1.5 0 -0.6 0.4 0.49 -7.4 -0.8 0.2 -0.4 0 N/A -0. ASA = rms azimuth spread of arrival angles. SF = shadow fading.52 0.77 0.4 0.4 0 -0.81 0.2 9 12 20 5 8 4 10 11 17 7 N/A 9 N/A 3.4 -0.89 0.1.4 -0.20 4 UMa NLOS -6.5 0 0 0 N/A 0 N/A 0 N/A 0 N/A Exp Exp Wrapped Gaussian 3.43 0.19 0.5 -0.14 1.5 Exp NLOS -7.4 N/A 0 N/A Exp Exp Wrapped Gaussian 2.2.66 1.20 0.03 0.4 0 -0.15 4 N/A N/A 0 0.55 0.0 19 20 10 22 3 10 10 9 13 N/A 2.28 1.18 1.22 3 7 4 0.90 0.5 Exp UMi NLOS -6. and K = Ricean K-factor.4 0 0 0 0 0 0.62 0.6 -0.25 0.1.8 1.5 N/A 0.7 0.18 1.45 1.42 1.5 -0.5 0.4 0 0.60 0. ASD = rms azimuth spread of departure angles.5 0 N/A N/A N/A N/A Exp LOS -7.5 -0.

Omote and T. Mogensen. L. Vehicular Technology Conference. [B4] P. Xinying Gao. Invited paper. 2924-2928. 2003.” Proc. Tech. "The impact of complexity in MIMO channel models.1816. 2007. Di Dong. Fujii.0 (2010-03) B. [B2] IST-WINNER II Deliverable 1. “The double-directional radio channel. Schumacher. Xinyin. January 2008.org/deliverables.Release 9 111 3GPP TR 36. The 2005 IEICE General Conference. Smith and M. Steinbauer. March 2005 (in Japanese). Japan. 30 2007-Oct. pp." IEEE Journal on Selected Areas in Communications..” Doc. 10. Lett. 2007 (http://www. 5. J. Geneva. 20. Xin Nie. Dublin. Ping. Zhang. Chen Huang and Guangyi Liu. pp. “WINNER II Channel Models”. 2264-2267.1. 2007. no. IST-WINNER2. Okamoto and S. [B14] Dong. “Propagation characteristics of wideband MIMO channel in urban micro. November.0. 2145-2146).Yu Zhang.814 V9.ist-winner. P.pdf.html). 2007 IEEE 66th.. Wu. S23-S24. 27. Gao. [B7] T. available at http://www. vol." in Proceedings of IEEE International Conference on Communications (ICC '04). June 2008.” IEEE PIMRC 2008. [B9] “Proposed Propagation Models for Evaluation Radio Transmission Technologies in IMT-Advanced. Correlation Model for Shadow Fading in Mobile Radio Systems. Yanping Liang. Paris. [B6] Rec.. 5D/88. and F.1. and E. K. Sept. Kermoal. Japan.and macrocell. E. “Cluster Identification and Properties of Outdoor Wideband MIMO Channel”. [B8] M. Switzerland. Zhang. ” The prediction of the time and the spatial profile for broadband land mobile services using UHF and SHF bands”. Electron. Frederiksen. 23. “Empirical arrival angle profile prediction formula for mobile communication systems” Proceeding of IEEE 2007 VTC Spring. M. Yufei. Shafi. Rep.2 v. Weihui. [B13] Jianhua Zhang.2. [B12] H. 3GPP . 3 2007 Page(s): 829 – 833. Fujii. Jianhua. 2002.2 References [B1] 3GPP TR25.0 Channel Model. WiMAX Forum Mobile Release 1. ITU-R P. A. August 2001. [B3] J. “Path-loss prediction formula in mobile communication -an expansion of “SAKAGAMI” path-loss prediction formula-. F.wimaxforum. 1991. Vol.1. June 2004. [B5] WiMAX Forum. [B11] H. I. Ichitsubo “Investigations of outdoor-to-indoor propagation loss on 800MHz-8GHz at urban area.996 V6. P. France. J86-B. Molisch. pp.org/technology/documents/wimax_forum_mobile_release_1_0_channel_model_v100 . 6.” Trans. vol. pp.0 (2003-09) “Spatial channel model for multiple input multiple output (MIMO) simulations” Release 6. Bonek. VTC-2007 Fall. "A stochastic MIMO radio channel model with experimental validation. IEICE. pp.” IEEE Antennas and Propagation Mag. 1211-1226. Pedersen. [B10] Gudmundson. 51-63.

8 and 9 are captured TP for simulation models are captured TP for winder bandwidth are captured TP for UL scheme and channel models are captured TP for UL/DL MIMO.0.814 ver.7.0 (2010-03) Annex C: Change history Table C.1. Relay are captured TR36. Subject/Comment Old New R1#54bis R1#55 R1#55bis R1#56 R1#56 18/03/10 RAN_47 18/03/10 RAN_47 R1-090510 R1-091009 R1-091125 RP-100357 - Draft skeleton TR TP for clause 5.0 2.0.0 0.4.0.0.0.3.6.01 9.2.1: Draft History Change history Date TSG # TSG Doc.2.0 0.814 V9.0. DL RS.1 0.0 3GPP .1 for approval by plenary TR is put under change control as per RAN_47 decision 2.0 1.0 0.Release 9 112 3GPP TR 36.0 2.

Sign up to vote on this title
UsefulNot useful