Professional Documents
Culture Documents
LTE Protocol Stack Physical Layer PDF
LTE Protocol Stack Physical Layer PDF
LTE
Physical Layer
LTE Protocol Stack
Author: Surya Patar Munda
3PCA-L1
If you need 3GNets LTE Physical Layer for Amateur Level (3PCA-L1), you need this course. This
knowledge and level is required for the next level – Professional Level (3PCP-L1) where you can
be trained for higher level with Hands on Projects and real implementation. Full Amateur level
courses are:
About Author:
Surya Patar Munda has been in Telecommunications Since 1987 and has gone through the life cycle
of Software Development, Software Testing, Network Deployments, Integration, Testing,
Troubleshooting, Handphone Testing with Specification etc.. a full round of the Telecom industry. He
has worked with Motorola, Nortel Networks, Spirent Communications, Sasken etc. companies with full
round cycle. The Software engineers midset and Testing engineers mindsets are different and so is
the mindset of an RF optimization engineer. This book will cater to all.
Author also conducted many trainings for Telecom industry and has a very good understanding of
what kind of requirement is there for engineers. The goal is not just what and how does it work, but
also the goal is how do I start implementing and how do I test.
First OFDM patent filed at Bell Labs in 1966, initially only as analog. In 1971, Discrete Fourier
Transform (DFT) was proposed. Later in 1980, application of the Winograd Fourier Transform (WFT)
or Fast Fourier Transform (FFT) was employed. OFDM then became modulation of choice for ADSL
and wireless systems.
OFDM tended to focus broadcast systems such as - Digital Video Broadcasting (DVB) and Digital
Audio Broadcasting (DAB), and WLANs. Main thing to control in OFDM was PAPR and thats why in
low power WLAN it was good. First cellular mobile based on OFDM was proposed in 1985 by IEEE
to LTE downlink. Other benefits of OFDM was to operate in different bandwidth according to spectrum
availability.
A guard period is created at the beginning of each OFDM symbol, to eliminate the remaining impact
of ISI. A Cyclic Prefix (CP) is added at the beginning of each symbol xk. The CP is generated by
duplicating the last G samples of the IFFT output and appending them at the beginning of xk. This
yields the time domain OFDM symbol [xk[N − G], . . . , xk[N − 1], xk[0], . . . , xk[N − 1]]T . CP length
G should be longer than the longest channel impulse response to be supported. The CP converts the
linear (i.e. aperiodic) convolution of the channel into a circular (i.e. periodic) one which is suitable for
DFT processing. The IFFT output is then Parallel-to-Serial (P/S) converted for transmission through
frequency-selective channel. Here is an example of OFDM LTE signal.
At the receiver, the reverse operations are performed to demodulate the OFDM signal, CP are
removed and ISI-free block of samples is passed to the DFT. If number of subcarriers N is designed
to be a power of 2, a highly efficient FFT implementation may be used to transform the signal back to
the frequency domain. Among the N parallel streams output from the FFT, the modulated subset of M
subcarriers are selected and further processed by the receiver.
Let x(t) be the signal symbol transmitted at time instant t . The received signal in a multipath
environment is then given by r(t) = x(t) * h(t) + z(t), where h(t) is the continuous-time impulse
response of the channel, represents the convolution operation and z(t) is the additive noise. Assuming
that x(t) is band-limited to [−1/2Ts ,1/2Ts], the continuous-time signal x(t) can be sampled at sampling
rate Ts such that the Nyquist criterion is satisfied. Due to multipath, several replicas of the transmitted
signals arrive at the receiver at different delays. The received discrete-time OFDM symbol k including
CP, under the assumption that the channel impulse response has a length smaller than or equal to G,
Receiver has to process equalization to recover xk[n] signals. CP of OFDM changes the linear
convolution into a circular one. The circular convolution is very efficiently transformed by an FFT into
a multiplicative operation in frequency domain. Hence, the transmitted signal over a frequency-
selective (multipath) channel is converted into a transmission over N parallel flat-fading channels in
Let x[n] be the signal after IFFT. PAPR of an OFDM symbol is defined as the square of the peak
2 2
amplitude divided by the mean power, i.e. PAPR = Max,n{|x[n]| } / E{|x[n]| }
It is observed that a high PAPR does not occur very often. However, when it does occur, degradation
due to PA non-linearities may be expected.
If T != 0, then the modulated data are in wrong positions, resulting in BER of 0.5 if the frequency offset
is not compensated at the receiver independently of the value of “e”. In case of T=0 and e<>0,
perfect orthogonality is lost, resulting in ICI with BER.
Even relative speed between transmitter and receiver also generates a frequency error due to
Doppler shift fd. ICI resulting from a mismatch fo between the transmitter and receiver oscillator
frequencies can be modelled as a Doppler shift.
The sensitivity of the BER depends on the modulation order. QPSK modulation can tolerate up to e max
= 0.05(5%), whereas 64-QAM requires e ≤ 0.01(1%).
This phase shift can be recovered as part of the channel estimation operation, with cyclic prefix but
not zero-padding.In the general case of a channel with delay spread, for a given CP length the
maximum tolerated timing offset without degrading the OFDM reception is reduced by an amount
equal to the length of the channel impulse response: To ≤ TCP − Td. For greater timing errors, ISI and
ICI occur. Timing synchronization becomes more critical in long-delay spread channels. Initial timing
is achieved by the cell-search and synchronization procedures. Thereafter, for continuous tracking of
timing-offset, either CP correlation or Reference Signals (RSs) is used.
If an OFDM system, CP is sufficiently designed of lengthG samples such that Channel impulse
Response L<G, to turn the linear convolution into a circular one to keep the subcarriers orthogonal.
The main propagation characteristics – (1) expected delay spread Td, (2) maximum Doppler
frequency fdmax , and, (3) targeted cell size, are considered for parameter dimensioning. This defines
the CP length and subcarrier spacing.
CP should be longer than channel impulse response for robustness against ISI. Large cells, longer
delay spreads need a longer CP. But longer CP means larger overhead (energy per transmitted bit).
Out of the N + G transmitted symbols, only N convey information, leading to a rate loss. Let TCP =
GTs and OFDM symbol period Tu = NTs , then overhead factor: βoverhead = TCP / (Tu + TCP).
To maximize spectral efficiency, Tu should be larger relative to CP, but small enough to ensure that
the channel does not vary within one OFDM symbol.
Further, Tu is related to subcarrier spacing by df = 1/Tu. Choosing a large Tu leads to a smaller
subcarrier separation df perfection , which impact sensitivity to Doppler and frequency offset.
Thus, in summary, three design criteria are identified:
Note that, with normal CP, the CP for the first symbol in each 0.5 ms slot is slightly longer than the
next six symbols, to accommodate an integer (7) number of symbols in each slot, with assumed FFT
block-lengths of 2048. For 20 MHz, FFT order of 2048 is assumed for efficient implementation.
However, in practice the implementer is free to use other Discrete Fourier Transform sizes.
These parameterizations are designed to be compatible with a sampling frequency of 30.72 MHz,
which is 8*3.84Mhz(UMTS sampling rate), for backward compatibility. Thus, the basic unit of time in
LTE, is defined as Ts = 1/30.72 μs. Lower sampling frequencies (and proportionally lower FFT orders)
are always possible to reduce RF and baseband processing complexity for narrower BW: Example,
for 5 MHz, FFT order and sampling frequency could be 512 and fs = 7.68 MHz respectively, while
only 300 subcarriers are actually modulated with data.
For simple implementation, direct current (d.c.) subcarrier is left unused, to avoid d.c. offset errors.
Smallest unit of resource is the Resource Element (RE) - one subcarrier for a duration of one OFDM
symbol. A RB comprised of 84 REs in normal CP (72 RE in extended CP). Within certain RBs, some
REs are reserved for synchronization signals (PSS/SSS), reference signals (RS), control signalling
and critical broadcast system information (CFICH,PHICH,PDCCH). Remaining REs are used for data
transmission(PDSCH), and are usually allocated in pairs (in time domain) of RBs.
Synchronization procedure detects specially designed Primary Synchronization Signal (PSS) and
Secondary Synchronization Signal (SSS). This enables time and frequency synchronization, provides
the UE with physical cell identity (PCI) and CP, and informs UE whether cell uses FDD or TDD. Here
is figure explaining relative location of PSS and SSS in frame structure of FDD and TDD respectively.
In initial synchronization, UE proceeds to decode PBCH for critical system information (SI). For new
cell identification, UE does not need to decode PBCH; it makes quality-level measurements
(RSRP/RSRQ) and reports to the serving cell.
thus enabling UE to acquire the slot boundary timing independently of the CP. SSS is located
in the symbol immediately preceding PSS, for coherent detection of SSS relative to PSS.
In TDD cell, PSS is located in 2 symbol of the 2 and 12 slots, while the SSS is located 3
nd nd th
PSS in a given cell is same in every subframe, SSS may change & thus UE knows the position of the
10 ms radio frame boundary. PSS and SSS are transmitted in the central six Resource Blocks (RBs),
irrespective of the system BW (6 to 110 RBs), without knowing BW. The PSS and SSS are each
comprised of a sequence of length 62 symbols, mapped to the central 62 subcarriers around d.c.
subcarrier which is left unused. Five REs at each extremity of each sync sequence are not used. Thus
a UE can detect the PSS and SSS with size-64 FFT and a lower sampling rate if all 72 subcarriers
were used.
In case of MIMO at eNodeB, PSS and SSS are always transmitted from same antenna port in a
subframe, while between different subframes they may be transmitted from different antenna ports for
diversity.
PSS and SSS sequence indicate one of 504 unique PCI, grouped into 168 groups of three identities.
The three identities in a group are assigned to cells under same eNodeB. Three PSS sequences are
used to indicate the cell identity within the group, and 168 SSS sequences are used to indicate the
identity of the group.
PSS uses Zadoff–Chu sequences
The NID,1 is the PCI group ID - , NID,1 can have values 0 to 167 and
NID,2 is the PCI local (may be sector) ID - NID,2 can have values 0 to 2.
How does the PCI optimization affect my Network? Well, this is the main parameter, by which the
PSS, SSS and reference signals will be generated. Even your scrambling code with which every DL
and UL signal will be scrambled, will depend on this. So, every generated signals uniqueness
depends on this parameter. Lets understand how some of the signals are generated based on PCI.
If PSS, SSS, RS and other generated signals are not unique, then my every operation will be affected
and it may reflect as latency in Synchronization detection, Interference and lower SINR values for
signals, which will end up in low CQI.
j un( n 1)
e 63 n 0,1,...,30
d u (n) u ( n 1)( n 2)
e j 63 n 31,32,...,61
where the Zadoff-Chu root sequence index u is given by following table.
(2)
N ID Root index u
0 25
1 29
2 34
The mapping of PSS to resource elements depends on the frame structure, FDD or TDD. The
sequence d(u,n) is mapped to the resource elements according to
ak ,l d n , n 0,...,61
DL RB
N RB N sc
k n 31
2
DL RB
N RB N sc
k n 31
2
n 5,4,...,1,62,63,...66 are reserved and not used for transmission of the primary synchronization
signal. N=0,1,…61 are used with above formulae.
PSS is constructed from a freq-domain ZC sequence of length 63, with middle element punctured to
avoid transmitting on d.c. subcarrier.
This set of roots for ZC sequences was chosen for its good periodic autocorrelation and cross-
correlation properties. These sequences have a low-frequency offset sensitivity (maximum undesired
autocorrelation peak /desired correlation peak) at a certain frequency offset, giving best robustness.
Also the ZC sequences are robust against frequency drifts. Thus, PSS can be easily detected during
the initial synchronization with a frequency offset up to ±7.5 kHz.
The selected root combination satisfies time-domain root-symmetry, sequences 29 and 34 are
complex conjugates of each other and can be detected with a single correlator. UE must detect PSS
without any prior knowledge of the channel, so noncoherent correlation is required for PSS timing
detection.
30
Thus UE determines the 10 ms radio frame timing from a single observation of a SSS. SSC2 is
scrambled by a sequence that depends on the index of SSC1. Sequence is then scrambled by a code
that depends on the PSS. Scrambling code is mapped to the PCI within the group corresponding to
the target eNodeB.
The resource mapping is done as per the following:
SSS sequences are spectrally flat. PSS, the SSS can be detected with a frequency offset up to ±7.5
kHz. Channel is known based on the PSS sequence first and then SSS detection is done.
However, in the case of synchronized neighbouring eNodeBs, coherent detector performance can be
degraded. If an interfering eNodeB employs the same PSS, phase difference between them can have
adverse impact on estimation of the channel coefficients. If BW of the channel is less than the six RB
for SSS, impact may be bad, hence minimum 6 RB legth is chisen. M-sequence and Walsh–
Hadamard matrices are similar and index remapping is done. This reduces complexity of SSS
detector, as complexity = N log2 N with N=32, complexity= 32 log2 32 = 160.
Cell search performance is measured as 90-percentile(maximum time required to detect a target cell
90%of the time) identification delay. After detection of PSS-SSS, RSRP is measured. For initial
synchronization case the time taken to decode PBCH is adapted, and not just of reporting of
measurements on RS.For inter-frequency handover, performance can be derived from the intra-
frequency performance timing.
Coherent detection uses amplitude and phase information exchanged between eNodeBs and UEs.
This comes at a price of overhead of channel estimation by exploiting known signals which do not
carry any data, sacrificing spectral efficiency. Known reference signals are inserted into the
transmitted signal structure. Reference signals(known) are multiplexed with data symbols (unknown
at receiver) in either frequency, time or code domains. Time multiplexing, known preamble-based
training transmission also is another technique.
Orthogonal RS multiplexing is the most common technique. OFDM transmission is a two-dimensional
lattice in time and frequency, which helps multiplexing of RSs mapped to specific REs according to a
specific pattern. Since RS are sent only on particular OFDM REs (particular symbols and subcarriers),
channel estimates for non-RS REs have to be computed via interpolation.
In frequency domain, there is one RS every six subcarriers on each symbols including RS symbol, but
staggered so that within each RB there is one RS every 3 subcarriers. This spacing is goverened by
expected coherence BW of channel, governed by channel delay spread. The 90% and 50%
coherence BW are given respectively by Bc,90% = 1/50d=20kHz and Bc,50% = 1/5d=200kHz where
d is the r.m.s delay spread=1000ns. In LTE spacing between two RS in frequency, is 45 kHz (3
symbols), enough to resolve expected frequency domain variations of the channel.
RS patterns are designed to work with MIMO antennas defined for multiple „antenna ports‟ at
eNodeB. An antenna port may be either a single physical antenna, or a combination of multiple
physical antenna elements. The transmitted RS in a given antenna port defines the antenna port from
the point of view of the UE, and enables UE to derive channel estimate for that antenna port.
Up to eight cell-specific antenna ports may be used by eNodeB, requiring UE to derive up to eight
separate channel estimates. For each antenna port, a different RS pattern is designed, to minimize
intra-cell interference between multiple transmit antenna ports. Rp is used for RS Tx on antenna port
p. Also, when a RE is used for RS on one antenna port, corresponding RE on other antenna ports is
set to zero to limit interference. Mark that, density of RS for third and fourth antenna ports is half of
the first two, to reduce overhead. In cells with a high prevalence of high-speed users, use of four
antenna ports is unlikely, RSs with lower density can provide sufficient channel estimation accuracy.
All the RSs (cell-specific, UE-specific or MBSFN specific) are QPSK modulated to ensure low PAPR.
The signal can be written as r(l,ns,m) = 1/√2[1-2c(2m)] + j1/√2[1-2c(2m+1)] where m is RS index, ns
= slot number and „l‟ =symbol number within slot, c(i) is length-31 Gold sequence, with different
initialization values depending on type of RSs. RS sequence carries unambiguously one of the 504
different cell identities, Ncell ID. For the cell-specific RSs, a cell-specific frequency shift (Ncell ID mod 6)
is also applied. This shift avoids collisions between common RS from up to six adjacent cells.
Transmission power of RS is boosted, up to max 6 dB relative to surrounding data symbols, designed
to improve channel estimation. If adjacent cells also transmit high-power RS on same REs,
interference will prevent the gain.
Propagation conditions characterize the channel in three dimension (frequency, time and spatial)
domains. Each MIMO multipath channel component can experience different scattering conditions
across the three domains. LTE specifications do not mandate any specific channel estimation
technique, and there is therefore complete freedom in implementation provided that the performance
requirements are met and the complexity is affordable.
PBCH uses dual-antenna receive diversity enabling wider cell coverage with fewer cell sites. Transmit
antenna diversity may be also employed at eNodeB to further improve coverage.
The exact REs used by PBCH is independent of transmit antenna ports; REs used for RS are avoided
by PBCH. Number of transmit antenna ports used by eNodeB must be determined blindly by the UE.
Discovery of number of transmit antenna ports is helped by CRC on each MIB which is masked with a
codeword representing the number of transmit antenna ports.
Low latency and a low impact on UE battery life is also facilitated by low code rate with repetition.
Full set of coded bits are divided into four subsets, each is self-decodable, which are sent in one of
four different frames during the 40 ms. UE may decode the MIB correctly from the transmission in less
than four radio frames, then UE does not need to receive other parts of PBCH in the remainder of 40
ms. On the other hand, if SIR is low, UE can receive further parts of MIB, soft-combining each part,
until successful decoding is achieved.Timing of 40 ms interval is not indicated explicitly to UE; it is
determined by scrambling and bit positions. UE can initially do four separate decodings of the PBCH
and checking the CRC for each decoding to determine 40ms boundary.
A simple approach is to perform decoding using soft combination of the PBCH over four radio frames,
advancing 40 ms sliding window one radio frame at a time until the window aligns with 40 ms period
of the PBCH and the decoding succeeds.
Except Tx-mode 7, RS for demodulating PDSCH is given by CRS. Number of eNodeB antenna ports
for PDSCH is same as PBCH. In Tx-mode 7, UE-specific RSs provide phase reference for the
PDSCH. Tx-mode also affects DL control signalling, and CQI from UE.
After coding and mapping to spatial layers, coded data bits are mapped to modulation symbols
depending on radio channel conditions and data rate required. Modulation order may be between two
bits per symbol (QPSK) and six bits per symbol (64QAM).
The RE for PDSCH can be any which are not reserved for other purposes (i.e. RS, PSS, SSS, PBCH
and control signalling(PCFICH,PHICH,PDCCH)). When UE is given a pair of RB of PDSCH, in a
subframe, only the available RE within RB can carry PDSCH data. Allocation of RB to PDSCH for a
Example of distributed Mapping: In Voice-over-IP (VoIP) service, certain frequency resources may
be „persistently-scheduled‟, on a periodic basis to a specific UE by RRC signalling rather than
PDCCH. As data per UE for VoIP is small (one or two pairs of RB), degree of frequency diversity
obtainable via localized scheduling is very limited. When dynamic channel-dependent PDCCH
scheduling is not done, frequency diversity is achieved through distributed mapping. A frequency-hop
occurs at slot boundary in the middle of subframe, block of UE data transmitted on one RB in first half
of subframe and on a different RB in the second half.
The potential number of VoIP users which can be accommodated in a cell increases by distributed
mapping, compared to localised.
Some spare time samples usage is unspecified between the end of the last control signalling symbol
and the first PMCH symbol, PMCH remaining aligned with the end of the subframe; eNodeB may
transmit an undefined signal or alternatively switch off its transmitter –UE cannot assume anything
about transmitted signal during these samples.
A UE measuring a neighbouring cell does not need to know the allocation of MBSFN subframes,
since UE knows that the first two OFDM symbols in all subframes use the same CP and RS pattern.
The MBSFN subframes patterns in a cell is indicated in SI, which indicates if pattern of MBSFN
subframes in neighbouring cells is same or different from current cell. If different pattern, then UE can
only ascertain the pattern by reading SI of that cell.
PCFICH is transmitted on same ports as PBCH, with transmit diversity if more than one antenna port
is used. For frequency diversity, 16 REs are distributed across BW with a predefined pattern in the
first symbol in each DL subframe, so that UEs can always locate. This is prerequisite to decode rest
of the control signalling.
A cell-specific (PCI based) frequency offset is applied to the positions of PCFICH REs. In addition, a
cell-specific scrambling sequence (PCI based) is applied to the CFI codewords to uniquely be sent.
CCEs are numbered and used consecutively. A PDCCH with a format consisting of n CCEs may only
start with a CCE with a number equal to a multiple of n. Format is decided by eNB based on RF
conditions.
If a UE has good downlink RF (e.g. close to eNB), format 0 may be sufficient, but for a cell border UE,
format 3 may be required for robustness with good power level of a PDCCH.
- SI-RNTI / P-RNTI / RA-RNTI, use Common Search Space. UL/DL C-RNTI/ SPS C-RNTI, and DL
Temp. C-RNTI, use UE-Specific Search Space. TPC-PUCCH-RNTI / TPC-PUSCH-RNTI and UL
Temp. C-RNTI is not considered for default CCE management.
- For SI-RNTI PDCCH candidate CCEs between 0 and (CS_Agr-1) is used and reserved in FDD and left
vacant if no SI-RNTI is scheduled. For TDD the default UL/DL configuration type 1, this PDCCH
candidate is reserved forS I-RNTI in sf 0 & 5 (and UL grant for C-RNTI/SPS-RNTI is not scheduled).
- CCEs between CS_Agr and (2*CS_Agr-1) can be used either for P-RNTI or RA-RNTI.
For FDD:
- For DL C-RNTI/SPS-RNTI/Temp C-RNTI the lowest m =m' from CCEs between 2*CS_Agr and
(Max_CCE-1) shall be used.
- For UL C-RNTI/SPS-RNTI the lowest m =m">m' from CCEs between 2*CS_Agr and (Max_CCE-1)
shall be used.
For TDD:
- For DL C-RNTI/SPS-RNTI/Temp C-RNTI the lowest m =m' which has a PDCCH available from
CCEs between 2*CS_Agr and (Max_CCE-1) shall be used.
- For UL C-RNTI/SPS-RNTI the lowest m =m">m' which from CCEs between 2*CS_Agr and
(Max_CCE-1) shall be used.
CCE resources utilized are well defined for default values of common search space aggregation level
=4, UE-specific search space aggregation L=2 resulting in 6 PDCCH candidates m=0..5. For different
Each TDD subframe (take sf config 1 as example) having different PHICH group number, and for
5/10/15/20 MHz bandwidth, each subframe has, therefore, different number of MAX_CCE. SF0 and
SF5 cannot be used for UL grant. SF1 and SF6 are not used for DL assignment. SF2, SF3, SF7 and
SF8 are not applicable to PDCCH CCE allocation since they are uplink subframes.
Format 0. DCI Format 0 is used for resource grants for the PUSCH.
TDDIndex 2-bits For TDD config 0, this field is the Uplink Index.
For TDD Config 1-6, this field is the Dow nlink Assignment Index.
Not present for FDD
Format 1A. DCI Format 1A is used for compact resource assignments for single codeword PDSCH,
and allocating a dedicated preamble signature to a UE for contention-free random access:
Format 1B. DCI Format 1B is used for compact resource assignments for PDSCH using closed loop
precoding with rank-1 (transmission mode 6). Information is same as in Format 1A, but with addition
of precoding vector indicator applied for the PDSCH.
4-bits (TDD)
NewData 1-bit New data indicator
RV 2-bits Redundancy version
TPCPUCCH 2-bits PUCCH TPC command
Format 1C. DCI Format 1C is used for very compact PDSCH assignments. With 1C format, PDSCH
is uses QPSK. This is used for paging, and some SI:
Format 1D. DCI Format 1D is used for compact signalling of resource assignments for PDSCH using
multi-user MIMO (transmission mode 5). Information is similar as in Format 1B. Instead of one of
precoding vector indicators bits, there is a single bit for power offset indicators, to show if transmitted
power is shared between two UEs.
DCI Formats DCISTR Fields Size Description
Format 2. DCI Format 2 is used for resource assignments for PDSCH for closed-loop MIMO
(transmission mode 4):
Format 2A. DCI Format 2A is used for resource assignments for PDSCH for open-loop MIMO
(transmission mode 3). Info is the same as Format 2, except that if eNodeB has two antenna ports,
there is no precoding information, and for four antenna ports two bits are used to indicate the
transmission rank.
Format 2B
DCI Formats DCISTR Fields Size Description
DCIFormat - „Format2B‟
AllocationType 1-bit Resource allocation header: type 0, type 1
(only if downlink bandwidth is >10 PRBs)
Allocation variable Resource block assignment/allocation
TPCPUCCH 2-bits PUCCH TPC command
HARQNo 3-bits (FDD) HARQ process number
4-bits (TDD)
„Format2B‟
Formats 3 and 3A. DCI Formats 3 and 3A are used for power control for PUCCH and PUSCH with 2-
bit or 1-bit power adjustments respectively.
DCIFormat - „Format3A‟
CRC attachment. For UE to know whether it has received a PDCCH correctly, a 16-bit CRC is
appended to each PDCCH. CRC is scrambled with „UE identity‟ for this to be identified for a particular
UE. In UL MIMO, antenna may be indicated using Format 0 by antenna-specific mask to the CRC.
This way, no extra bit needed.
PDCCH construction. The PDCCH bits are encoded. The coded and rate-matched bits are then
scrambled with a cell-specific scrambling sequence to distinguish from neighbouring cells. The
scrambled bits are QPSK modulated and mapped to blocks of four REs (REGs). Interleaving is
applied for frequency diversity, followed by RE mapping to symbols indicated by PCFICH, excluding
PCFICH and PHICH. The PDCCHs are transmitted similar to PBCH, and diversity is applied if more
antenna ports are used.
PHICH index is implicitly associated with the index of the lowest uplink RB used for PUSCH. The
adjacent PUSCH RBs are associated with PHICHs in different PHICH groups, for load balancing.
However, for MU-MIMO, this is not sufficient to enable multiple UEs to be allocated the same RBs for
a PUSCH. In this case, different cyclic shifts of RS are configured for different UEs for the same
PUSCH resources in time-frequency, and same cyclic shift index is then used to shift PHICH so that
each UE will receive its ACK or NACK on a different PHICH.
Resource allocation Type 0. A bitmap indicates Resource Block Groups (RBGs) allocated to
scheduled UE, where a RBG is a set of consecutive PRBs. RBG size (P=1,2,3,4) depends on BW.
Total number of RBGs = NRBG = Total_RB /P.
Resource allocation Type 1. Here Individual PRBs can be addressed only within a subset of the
PRBs available. Bitmap is slightly smaller than Type 0, since some bits are used to indicate the
subset of addressed RBG, and a shift in position. Total bits is still the same as for Type 0. One bit is
used for subset selection and another bit to indicate the shift. The provides flexibility in spreading the
RBs across BW to exploit frequency diversity.
Usually Transmission mode 3 is used in MIMO test cases with 2 Transmit antenna, where UE is
expected to decode only DCI formats 2A and 1A. Similarly for TM4, UE is expected to decode only
DCI formats 2 and 1A.
General DL scheduling scheme
The bandwidth of 5/10/20 MHz makes 25/50/100 available physical resource blocks respectively.
These resource blocks are divided into three distinct sets. Exact set sizes and the elements contained
in the individual sets depend upon the DCI combination to be applied.
The first set is reserved for BCCH mapped to DL-SCH (SI-RNTI).
The second set is reserved for PCCH mapped to DL-SCH (P-RNTI).
The third set is used for one of mutually exclusive transmissions of:
a. 'Random Access Response' mapped to DL-SCH (RA-RNTI); or
b. UE-dedicated scheduling mapped to DL-SCH (C-RNTI/ SPS C-RNTI/ Temp C-RNTI).
For each subframe where data is scheduled, eNB shall select a Transport Block Size (TBS),
independently for each type of data scheduled, such that:
All the scheduled data is transmitted respecting the timing information.
Not more than MaxRbCnt resource blocks are used, for DCI format 1C, NPRB = MaxRbCnt.
Minimum MAC Padding is performed.
If all scheduled Data cannot be transmitted in the indicated subframe, for example due to TDD and half
duplex configuration, it shall be transmitted in the next available subframe.
RACH procedure: UE in idle mode, handed over to a new cell or connected mode but PUCCH is
unsynchronized (sometimes referred to as PUCCH is not configured) will trigger RACH procedure on
data ready for transmission in UL.
Scheduling Request: UE in connected mode, no grant configured, PUCCH is synchronized and
has data ready for transmission in UL, will transmit a scheduling request on PUCCH.
Buffer Status Reports: UE in connected mode, PUCCH synchronized, has a configured grant
for current TTI, but grant is not sufficient to transmit all the data will include MAC control element BSR
in the UL MAC PDU.
RACH and SR indicate on data availability and BSR provides an estimate of data available for
transmission. CQI/PMI/RI feedback from the UE which indicates the channel conditions and
recommended number of layers.
Hence to determine the exact need of the grant requirement of the UE a network needs to act on all
four of the above.
The NW disables aperiodic CQI/PMI/RI feedback from the UE by setting the „CQI request field‟ to 0 in
DCI format 0/RAR grant.
eNB, will periodically transmit automatically MAC PDUs containing the MAC control element 'Timing
Advance'. The period normally is set to 80 % of the 'Time Alignment Timer' default value (750 ms)
configured at UE.
With such small search spaces, eNodeB may not find CCE to send PDCCHs to all the UEs. To
minimize blocking persisting into next subframe, a UE-specific hopping sequence is applied to the
starting positions of the dedicated search spaces.
To keep blind decoding attempt computational load, UE is not required to search for all DCI formats
simultaneously. Typically, in dedicated search space, UE will always search for Formats 0 and 1A. In
addition, UE may receive a further format 1, 1B or 2 DCI depending on PDSCH transmission mode.
In the common search space, UE will search for Formats 1A and 1C. In addition UE may be
configured to search 3 or 3A DCI, and may be distinguished by CRC scrambled by a different
(common) identity, rather than a UE-specific one. This way not it is limited blind decoding required.
If the UE is configured with a carrier indicator field, then the UE shall monitor one or more EPDCCH
UE-specific search spaces at each of the aggregation levels on one or more activated serving cells.
On each serving cell c, UE shall monitor EPDCCH with CRC scrambled by C-RNTI and SPS C-RNTI
in the EPDCCH UE specific search space of serving cell c.
For a given serving cell, for each EPDCCH-PRB-pair set p , the UE is configured with
resourceBlockAssignment-r11 indicating a combinatorial index r .
EPDCCH formats
The EPDCCH carries scheduling assignments which is transmitted using an aggregation of one or
several consecutive enhanced control channel elements (ECCEs) where each ECCE consists of
multiple enhanced resource element groups (EREGs). The number of ECCEs used for one EPDCCH
depends on the EPDCCH format and the number of EREGs per ECCE is pre-defined. Both localized
and distributed transmission is supported.
An EPDCCH can use either localized or distributed transmission, differing in the mapping of ECCEs
EREG
to EREGs ( N ECCE = 4 for normal or 8 for extended CP per ECCE) and PRB pairs.
A UE shall monitor multiple EPDCCHs. One or two sets of PRB pairs which a UE shall monitor for
EPDCCH transmissions can be configured. All EPDCCH candidates in EPDCCH set S m use either
only localized or only distributed transmission as configured. Within EPDCCH set S m in subframe i ,
the ECCEs available for transmission of EPDCCHs are numbered from 0 to N ECCE,m,i 1 and ECCE
number n .
EREG
N ECCE ECCE
is the number of EREGs per ECCE, and N RB 16 N ECCE
EREG
is the number of ECCEs/RB pair.
The PRB pairs constituting EPDCCH set S m are assumed to be numbered in ascending order from 0
Sm
to N RB 1.
When DCI formats 2, 2A, 2B, 2C or 2D is used and N RB
- DL
25 , - 2,4,8,16 or 32 ECCEs may
be there
- any DCI format when nEPDCCH 104 and normal cyclic prefix is used in normal subframes or
special subframes with configuration 3, 4, 8(TDD) - 2,4,8,16 or 32 ECCEs may be there
- otherwise - 2,4,8,16 or 32 ECCEs may be there
The quantity nEPDCCH for a particular UE is defined as the number of downlink resource elements
(k , l ) in a PRB pair configured for possible EPDCCH transmission of EPDCCH set S 0 and fulfilling all
of the following criteria:
- they are part of any one of the 16 EREGs in the PRB pair, and
EPDCCH-Config
The IE EPDCCH-Config is used to configure the subframes and resource blocks for EPDCCH
monitoring.
-- ASN1START
-- ASN1STOP
EPDCCH
dmrs-ScramblingSequenceInt - The DMRS scrambling sequence initialization parameter nID,i .
epdcch-SetConfig - Provides EPDCCH configuration set. E-UTRAN configures at least one epdcch-
SetConfig when EPDCCH-Config is configured.
epdcch-SetIdentity - Indicates the indentity of the EPDCCH set.
epdcch-StartSymbol (1,2,3,4) - Indicates the OFDM starting symbol for any EPDCCH and PDSCH
scheduled by EPDCCH on the same cell, if the UE is not configured with tm10. If not present, the
configuration is released and the UE shall derive it from PCFICH. It is not configured for UEs
configured with tm10.
epdcch-SubframePatternConfig - Configures the subframes which the UE shall monitor the UE-
specific search space on EPDCCH. If it is not configured when EPDCCH is configured, the UE
monitors the UE-specific search space on EPDCCH in all subframes except for pre-defined rules.
epdcch-TransmissionType - Indicates whether distributed or localized EPDCCH transmission mode
is used.
numberPRBPairs - Indicates the number of PRB pairs used for the EPDCCH set. Value n2
corresponds to 2 PRB pairs; n4 corresponds to 4 PRB pairs and so on. n8 is not supported for dl-
Bandwidth having value n6.
pucch-ResourceStartOffset - PUCCH format 1a and 1b resource starting offset for the EPDCCH
set.
re-MappingQCLConfigListId - Indicates the starting OFDM symbol, the related rate matching
parameters and quasi-collocation assumption for EPDCCH when the UE is configured in tm10. This
The eNodeB selects modulation & code rate depending on a prediction of DL channel conditions,
based on Channel Quality Indicator (CQI) transmitted by the User Equipment (UE) in UL. CQI is an
indication of data rate which can be supported by the channel, based on SINR and UE‟s receiver. The
eNodeB can select different CQI feedback modes to trade off the improved DL link adaptation
enabled by CQI against UL overhead caused by the CQI itself.
Signalling necessary for interoperability between eNodeB and UEs are defined for link adaptation, but
exact methods used by eNodeB are left to the manufacturer‟s choice. In response to CQI, eNodeB
can select between QPSK, 16-QAM and 64-QAM schemes and a range of code rates.
a0 , a1 ,..., a A 1
Transport block
CRC attachment
b0 , b1 ,..., b B 1
Channel coding
Rate matching
Code block
concatenation
f 0 , f1 ,..., f G 1
Should all RBs allocated to one UE in a subframe use same MCS or should be frequency dependent?
In general only a small throughput improvement arises from a frequency-dependent MCS compared
to an RB-common MCS, hence frequency-dependent MCS is not justified, and they are constant over
the allocated RB for a given user, and time-domain scheduling and AMC is supported. If multiple TBs
are transmitted to one user in a subframe using MIMO, each TB can use an independent MCS.
CQIs assists the eNodeB in selecting MCS for DL. CQI is derived from DL RSRQ of Ref Signal,
Reported CQI indicates highest MCS it can decode with a TB error rate probability not exceeding
10%, and not the SINR. Hence a UE designed with better algorithms (like interference cancellation
techniques) can report higher CQI and can receive a higher data rate. CQI could be based on a set of
−1
Block Error Rate (BLER) thresholds. UE reports CQI value for MCS to ensures BLER ≤ 10 based on
measured received signal quality.
Trellis diagram represents a finite state machine including time dimension. Consider an input block
with L bits encoded with a rate 1/n (i.e. k = 1) convolutional encoder, resulting in a codeword of length
(L + m) × n bits, including m trellis termination bits (or tail bits) inserted at the end of the block to drive
the shift register contents back to all zeros at the end of the encoding process.Using tail bits is just
one possible way of terminating an input sequence.
Tail Biting Method: Method of simple truncation (no tail bits appended) is called tail-biting. In tail-
biting, initial and final states of convolutional encoder are required to be identical. Usually tail-biting for
feed-forward convolutional encoders is achieved by initializing the shift register contents with the last
m bits in the input block. Tailbiting encoding facilitates uniform protection of the information bits and
suffers no rate-loss owing to the tail bits. Tail- can be decoded using, Circular Viterbi Algorithm (CVA).
ck
D D D D D D
d k( 2) G2 = 165 (octal)
Let M(yi | xi) = Σ(j..n)( log P(yi,j | xi,j )) be branch metric at the ith trellis step. The VA, using trellis,
computes „best‟ partial path metric at each step by adding, comparing and selecting metrics. For each
state s, VA computes the possible partial path metrics corresponding to all the edges arriving in state
s, and selects the best partial metric. In this example, at time = 2, there are three possible paths
ending. The VA computes and selects the best path metrics as the survival edge and the other one is
discarded. This is carried on for each state and for = 0, . . . , L + m − 1. At the last stage l=L+m−1, VA
Convolutional codes are most widely used for good performance, decoding speed based on VA and
flexible codeword sizes adaptation. Followed by this, turbo code and Low-Density Parity Check
(LDPC) codes were discovered that provided near-Shannon limit performance.
Turbo Codes
Turbo codes is an iterative decoding algorithm to achieve near-Shannon limit performance. Encoder
has two convolutional encoders linked by an interleaver. Two identical convolutional codes have g0 =
[13] and g1 = [15]. Turbo code encodes the input block twice (with and without interleaving) to
generate two set of parity bits. Each encoder is terminated to all zero state by using tail bits. The
nominal code rate of turbo code is 1/3.
Number of states in the trellis of a turbo code is significantly larger due to the interleaver, making it
intractable (except for trivial block sizes). Therefore, iterative decoding is done based on separate
optimal decoder for each constituent convolutional coder, both iteratively exchanging bits via a
(de)interleaver.
The two decoders cooperate by iteratively exchanging bits via (de)interleaver. After a certain number
of iterations, the output can be used to obtain final hard decision estimates of the information bits.
xk
ck
D D D
Output
Input
Turbo code internal
interleaver 2nd constituent encoder zk
Output
ck
D D D
xk
It was enhanced by the ability to select different redundancy versions for HARQ retransmissions.
However the decoder shows the strain at 10Mbps. LTE effort began for data rates of 100 Mbps to
1Gbps in view. For LTE, turbo interleaver was replaced with a „contention-free‟ interleaver.
It requires that for each window, the memory banks accessed be unique between any two windows,
thus eliminating access contentions. Instead of using M separate memories, better to use single
physical memory and fetch/store M values on each cycle from a single address. This requires CF
interleaver to satisfy a vectorized decoding property where the intra-window permutation is the same
for each window.
A variety of possible parallelism factors provides freedom for each individual manufacturer to select
the degree of parallelism based on the target data rates for different UE categories. After
consideration of performance, available flexible classes of CF interleavers and complexity benefits, a
new contention-free interleaver was selected for LTE.
i K f1 f2 i K f1 f2 i K f1 f2 i K f1 f2
1 40 3 10 48 416 25 52 95 1120 67 140 142 3200 111 240
2 48 7 12 49 424 51 106 96 1152 35 72 143 3264 443 204
-- --- --- --- --- --- --- --- --- --- --- --- --- --- --- ---
46 400 151 40 93 1056 17 66 140 3072 47 96 187 6080 47 190
47 408 155 102 94 1088 171 204 141 3136 13 28 188 6144 263 480
Fig 2.4.5 – Interleaver position converter f1-f2 Table
A total of 188 interleavers are defined for LTE, of which 153 have quadratic inverses while the
remaining 35 have degree-3 and degree-4 inverses. Attractive feature of QPP interleavers is that they
are „maximum contention-free‟, supporting parallelism. For example, for K = 1024, supported
parallelism factors include {1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024}, although factors that result in a
window size less than 32 may not be required in practice.
QPP interleavers also have „even-even‟ property whereby even and odd indices in the input are
mapped to even and odd indices respectively in the output; this enables the encoder and decoder to
process two bits per clock cycle.
Following 188 byte-aligned interleaver sizes spaced in a semi-log manner are selected with
approximately 3% filler bits:
K=
40 + 8t if 0 ≤ t ≤ 59 (40–512 in steps of 8 bits)
512 + 16t if 0 < t ≤ 32 (528–1024 in steps of 16 bits)
1024 + 32t if 0 < t ≤ 32 (1056–2048 in steps of 32 bits)
2048 + 64t if 0 < t ≤ 64 (2112–6144 in steps of 64 bits)
Maximum turbo interleaver size is increased from 5114 in UMTS to 6144 in LTE, such that a 1500
byte TCP/IP packet would be segmented into only two segments rather than three, minimizing
potential segmentation penalty and (marginally) increasing turbo interleaver gain.
1.4.6. Rate-Matching
Rate-Matching (RM) algorithm selects bits for transmission by puncturing and/or repetition, based on
the available physical resources. RM should send as many new bits as possible in retransmissions to
maximize Incremental Redundancy (IR) HARQ gains.RV = 0 starts at an offset relative to the
beginning of the CB(Code Block) to enable systematic bit puncturing on the first transmission.Circular
buffer RM was selected for LTE as it generates puncturing patterns simply and flexibly for any
arbitrary code rate, with excellent performance.
Each of the three output streams of the turbo coder (systematic part, parity0, and parity1) is
rearranged with its own interleaver (Sub-block interleaver). The 12 tail bits are distributed equally into
the three streams, resulting in sub-block size Ks = K + 4, where K is the QPP interleaver size. Then,
an output buffer is formed by concatenating the rearranged systematic bits with the interlacing of the
two rearranged parity streams. For any desired code rate, the coded bits for transmission are simply
read out serially from a certain starting point in the buffer, wrapping around to the beginning of the
buffer if the end of the buffer is reached.
A Redundancy Version (RV) specifies a starting point in the circular buffer to start reading out bits.
Different RVs are specified with different starting points to enable HARQ operation. RV = 0 is selected
antenna
codewords layers
ports
c ( q ) (i) . The scrambling sequence generator shall be initialised at the start of each subframe, where
the initialisation value of cinit is
1.4.11. Modulation
For each codeword q of M scrambled bits, shall be modulated as using one of the modulation
schemes of QPSK, 16QAM or 64QAM, resulting in a block of complex-valued Msymb modulation
symbols.
1.4.13. Precoding
Precoder takes as input a block of vectors x(i) for v layers ( i 0,1,...,M symb
layer
1 ) from layer mapping and
(p)
generates a block of vectors y (i) for p antenna ports each ( i 0,1,...,M symb
ap
1 ) to be mapped onto
resources, where y ( p ) (i) represents the signal for antenna port p . For transmission on a single
antenna port, precoding is defined by y ( p) (i) x (0) (i) for each symbol i 0,1,...,M symb
ap
1 ,
ap
M symb M symb
layer
.
y (0) (i ) x (0) (i )
W (i )
y ( P 1) (i ) x ( 1) (i )
Values of W (i) shall be selected among the precoder elements in the codebook configured in the
eNodeB and the UE.
y (0) (i ) x (0) (i )
W (i ) D(i )U
y ( P 1) (i ) x ( 1) (i )
The diagonal size- matrix D(i) supporting CDD and the size- matrix U are given for
different numbers of layers . A different precoder is used every vectors, where denotes the
number of layers.
power allocation and be mapped in sequence starting with y ( p ) (0) to resource elements k, l , when:
- they are in RB assigned for transmission, and
Single-Input Single-Output (SISO) exploit time or frequency domain pre-processing and decoding of
the transmitted and received data respectively. The use of multiple antenna at either eNodeB or UE
(on DL or UL) requires signal precoding and detection. Here is an example how signals are precoded
in every path and decoded back.
In multi-antenna enabled base station with a single antenna UE, uplink is SIMO and downlink is
MISO. When multi-antenna UE is used, it is called MIMO, also SIMO and MISO also is used within
MIMO definition. A point-to-point multiple-antenna link between eNB and one UE is referred as
Single-User MIMO (SU-MIMO), Multi-User MIMO (MU-MIMO) features several UEs communicating
simultaneously with a common eNB using the same frequency- and time-domain resources.
In a multicell context, neighbouring eNBs sharing their antennas in virtual MIMO to communicate with
the same set of UEs in different cells will be termed multicell multi-user MIMO.
There are basically three advantages of MIMO over SISO – (a)Diversity gain, (b)Array gain and
(3)Spatial multiplexing gain.
Diversity gain corresponds to mitigation of multipath fading, by transmitting or receiving over multiple
antennas at which fading is sufficiently decorrelated. It is expressed in terms of an order of number of
effective independent diversity branches or to slope of BER curve (function of SNR or link budget
gain). Gain is related to improvement of statistics of instantaneous SNR, array gain and multiplexing
gain.
Array gain corresponds to a spatial version of matched-filter gain in time-domain receivers.
Multiplexing gain refers to the gain where data of multiple users is multiplexed and separated by
orthogonal spreading codes, timeslots or frequency assignments. MIMO multiplexing has no extra
cost of bandwidth expansion; but needs added antennas and signal processing complexity.
The ith column of H (hi) can be receive spatial signature of ith transmitting antenna
The jth row of H (hj) can be transmit spatial signature of jth receiving antenna.
The hij can be the signal transmitted by antenna i and received by j.
The optimal SU-MIMO multiplexing uses SVD-based transmit and receive beamforming to
decompose MIMO channel into a number of parallel non-interfering subchannels („eigen-channels‟),
each one with an SNR (singular value λi based) and chosen power level pi .
The philosophy of optimal power allocation across the eigen-channels is not to equalize the SNRs,
but to render them more unequal, by „pouring‟ more power into the better eigen-channels, while
allocating little power (or even none at all) to the weaker ones because they are seen as not
contributing enough to the total capacity. In practice this is done by selecting a suitable Modulation
and Coding Scheme (MCS) for each stream.
Beyond classical linear detection (ZF or MMSE) receivers, more advanced but nonlinear detectors
can be exploited at extra complexity like- Successive Interference Cancellation (SIC) detector and
Maximum LikelihoodDetector (MLD).
SIC treats individual channel-encoded streams, like layers which are peeled off one by one by a
processing sequence consisting of linear detection, decoding, remodulating, re-encoding and
subtraction from the total received signal R.
MLD selects the most likely set of all streams from R, by an exhaustive search procedure or a lower-
complexity equivalent such as sphere-decoding technique.
Multiplexing gain
Multiplexing gain multiplicative factor by which spectral efficiency is increased by a given scheme.
MIMO achieves gain for various antennas to experience a sufficiently different channel response, to
be sufficiently
decorrelated and linearly independent to allow for the channel matrix H to be invertible. There is a
limitation to MxN number of independent streams which may be multiplexed into the MIMO channel
(rank(H) streams). SU-MIMO between a four-antenna eNB and a dual antenna UE can, at best,
support multiplexing of two data streams, doubling UE‟s data rate compared with a single stream.
Diversity
A diversity-oriented design will feature some level of repetition between the entries of Y. For „full
diversity‟, each symbol xi must be assigned to each transmit antennas at least once during T symbol
durations. Diversity symbol-to-transmit signal mapping function is called Space-Time Block Code
(STBC). In addition to STBC, orthogonality of matrix Y improves performance and easy decoding at
receiver, realized by Alamouti space-time code. The diversity order is equal to MxN. For this
transmission, no knowledge of channel and feedback is necessary.
In single antenna case, no precoding can be applied and each UE simply transmits an independent
message. Thus, if K UEs are selected for transmission in the same time-frequency resource, each UE
k transmitting symbol s(k) , the received signal at eNB, over T = 1 symbol period is R = H.X + N where
X =x(1) ...x(k). The columns of H correspond to the receive spatial signatures of different UEs. eNB
can recover transmitted symbol by applying beamforming filters, using MMSE or ZF solutions. Note,
no more than N UEs can be served (K ≤ N). MU-MIMO in UL is sometimes referred as „Virtual MIMO‟,
UE has no knowledge of the simultaneous transmissions of the other UEs. On DL, eNB must resort to
transmit beamforming to separate data streams intended for the various UEs.
MU-MIMO in DL with single-antenna UEs: eNB transmits to K selected UEs simultaneously. Their
contributions are separated by multiple antenna precoding at eNB, based on channel knowledge.
Over T=1 symbol period, signal received by UEs 1 to K = R = r1 ...rK = HVP.X + N. The rows of H
correspond to the transmit spatial signatures of various UEs. V is the transmit beamforming matrix
and P is the (diagonal) power allocation matrix selected . To cancel out fully the inter-user
interference when K ≤ N, a transmit ZF beamforming solution is employed.
However, for low SNR, the SU-MISO and MU-MIMO schemes perform very similarly and both are
worse than in the idealized i.i.d. channels. So, MU-MIMO should be used in better SINR environment.
The full MIMO benefits (array gain, diversity gain and multiplexing gain) assume ideally
decorrelated antennas and full-rank MIMO channel matrices. In single-user case, antennas at
both eNB and UE are typically separated by between 1/2 wavelength to a few
wavelengths at most, which is very short in relation to eNB to UE. In a LOS situation, this will
cause a strong correlation between spatial signatures, limiting the use of multiplexing
schemes. But design itself may provide the necessary orthogonality properties even in LOS
situations. Two antennas (at both transmitter and receiver) that operate on orthogonal
polarizations (horizontal and vertical polarizations, +45◦ and −45◦ polarizations, which give a
twofold multiplexing capability even in LOS). Orthogonal polarizations at UEs may not always
be recommended as it results in non-omnidirectional beam patterns. Also, in SU-MIMO, the
condition of spatial signature independence can only be satisfied with rich random multipath
propagation.
Another source of discrepancy between theoretical MIMO gains and practically achieved
performance lies in the (in-)ability of the receiver to give right CSI frequently and perfectly,
but we know that they are limited and finite. This degrades the performance.In DL, MU-MIMO
relies on eNB to compute required transmit beamformer, which in turn requires CSI. If no
sufficient CSI available MU-MIMO gains disappear and SU-MIMO strategy becomes
optimal.SO, accurate Channel State Information (CSI) to be delivered by UE to eNB in a
resource-efficient manner is required. This requires use of appropriate codebooks for
quantization.
Another issue is the interaction between the physical layer and the scheduling protocol. In
both UL and DL cases, number of UEs served in MU-MIMO is limited to K = N, assuming
linear combining. Number of active users U will typically exceed K, right k users will have to
be scheduled for simultaneous transmission over a particular RB. This algorithm is not
specified in LTE and various approaches are possible. A combination of rate maximization
and QoS constraints may be considered. The choice of UEs that will maximize the sum-rate
is one that favours UEs exhibiting not only good instantaneous SNR but also spatial
separability among their signatures.
In closed-loop spatial multiplexing, UE feeds back to eNodeB most desirable entry from a predefined
codebook. Preferred precoder matrix would maximize capacity. Some important properties of the LTE
codebooks are as follows:
• Constant modulus property. Precoders in LTE are mostly pure phase corrections –no amplitude
changes, with exception of identity matrix as the precoder. Identity precoder may completely switch
off one antenna on one layer, as long as the net effect across the layers is still constant modulus to
the PA.
• Nested property. Codebooks of different ranks are arranged, so that lower rank codebook is subset
of the higher rank codebook vectors. It ensures that the precoded transmission for a lower rank is a
subset of the precoded transmission for a higher rank. For example, if a specific index in the
codebook corresponds to the columns 1, 2 and 3 in case of a rank 3 transmission, then the same
index in rank 2 transmission must consist of either columns 1 and 2 or columns 1 and 3.
• Minimal ‘complex’ multiplications. The 2-antenna codebook consists entirely of a QPSK alphabet,
all codebook multiplications use only ±1 and ±j . The 4-antenna codebook does contain some QPSK
entries which require a √2 magnitude scaling as well. Here are precoding examples with diversity and
spatial multiplexing respectively.
Adding a time delay = applying a phase shift in frequency. Same time delay is applied to all
subcarriers, phase shift will increase linearly across the subcarriers. Each subcarrier will experience a
different beamforming pattern as non-delayed subcarrier interferes constructively or destructively with
the delayed version from another antenna. The diversity effect of CDD therefore arises from the fact
that different subcarriers will pick out different spatial paths in the propagation channel, thus
increasing the frequency-selectivity of the channel. It helps to ensure that any destructive fading is
constrained to individual subcarriers rather than affecting a whole TB. It is beneficial if CSI is
unreliable.
Delay is added before the CP means, any delay value is used without increasing overall delay spread.
The time-delay/phase-shift equivalence means that the CDD operation can be implemented as a
frequency-domain precoder. The implementation designer can choose whether to implement CDD in
the time domain or the frequency domain.
CDD is discussed so far for rank-1 transmission, single layer. In practice, CDD is only applied in LTE
when PDSCH rank>1. Each layer benefits independently from CDD as for a single layer. For
multilayer CDD operation, the mapping of the layers to antenna ports is carried out using precoding
matrices selected from the spatial multiplexing codebooks.
In the case of 2 transmit antenna ports, the predetermined spatial multiplexing precoding matrix W is
always the same (identity matrix). In the case of 4 transmit antenna ports, ν different precoding
matrices are used from the 4 transmit antenna port codebook where ν is the transmission rank.
In the standard, included are: (1) format and procedure for feedback from UEs, which may or may not
include method of calculating quantities to feed back; (2) Codebooks for feedback calculation and
The performance degradation of zero-forcing beamforming may occur if the reported channel
feedback information is very inaccurate. Zero-forcing beamforming can be less robust against
inaccuracies in the channel representation compared to unitary precoding. This would be the case, if
the codebook size for the channel feedback is small or channels are spatially highly uncorrelated.
When it is not possible to provide sufficiently accurate channel information, the unitary precoding can
prove more robust and than a zero-forcing approach.
SC-FDMA advantage over UMTS is that it achieves intra-cell orthogonality even in frequency-
selective channels. However, a UE with code-multiplexed uplink also suffers increased CM/PAPR.
OFDMA could have been suitable for UL as well, except for low CM/PAPR. SC-FDMA combines
OFDM with the low CM/PAPR of single-carrier transmission schemes.
Like OFDM, SC-FDMA divides the bandwidth into multiple parallel subcarriers, with the orthogonality
between the subcarriers by Cyclic Prefix (CP) or guard period to prevent Inter-Symbol Interference
(ISI) between SC-FDMA information blocks. It transforms the linear convolution of the multipath
channel into a circular convolution to equalize the channel.
However, unlike OFDM, in SC-FDMA the signal modulated onto a given subcarrier is a linear
combination of all the data symbols transmitted at the same time instant. In each symbol period, all
the transmitted subcarriers of an SC-FDMA signal carry a component of each modulated data symbol.
This gives SC-FDMA its crucial single-carrier property, which results in low CM/PAPR than OFDM.
Usually N is larger than the maximum number of occupied subcarriers, thus providing for efficient over
sampling and „sinc‟ (sin(x)/x) pulse-shaping. The equivalence of DFTS-OFDM and a time-domain-
generated SC-FDMA transmission can readily be seen by considering the case of M = N, where the
DFT operation cancels the IFFT of the OFDM modulator resulting in the data symbols being
transmitted serially in the time domain. However, this simplistic construction would not provide any
oversampling or pulse-shape filtering.
• Localized transmission.
The subcarrier mapping allocates a group of M adjacent subcarriers to a user. M <N results in
zero being appended to the output of the DFT spreader resulting in an up-sampled/interpolated
version of the original M QAM data symbols at the IFFT output of the OFDM modulator. The
transmitted signal is thus similar to a narrowband single carrier with a CP (equivalent to time-domain
generation with repetition factor L = 1) and „sinc‟ pulse-shaping filtering (circular filtering).
• Distributed transmission.
The subcarrier mapping allocates M equally-spaced subcarriers (e.g. every Lth subcarrier). (L
− 1) zeros are inserted between the M DFT outputs, and additional zeros are appended to either side
of the DFT output prior to the IFFT (ML< N). As with the localized case, the zeros appended on either
side of the DFT output provide upsampling or sinc interpolation, while the zeros inserted between the
DFT outputs produce waveform repetition in the time domain. This results in a transmitted signal
similar to time-domain IFDMA with repetition factor L and „sinc‟ pulse-shaping filtering.
Subframe duration 1 ms
Slot duration 0.5 ms
Subcarrier spacing 15 kHz
SC-FDMA symbol duration 66.67 μs
CP duration 4.69 μs all other symbols Normal CP: 5.2 μs first symbol in each slot,
Extended CP: 16.67 μs all symbols
Number of symbols per slot 7 (Normal CP)
6 (Extended CP)
Number of subcarriers per RB 12
An important feature of the LTE SC-FDMA parameterization is that the numbers of subcarriers which
can be allocated to a UE for transmission are restricted such that the DFT size in LTE can be
constructed from multiples of 2, 3 and/or 5. This enables efficient, low complexity mixed-radix FFT
implementations.
The same basic transmission resource structure is used for the uplink as for the downlink: a 10 ms
radio frame is divided into ten 1 ms subframes each consisting of two 0.5 ms slots. It uses the same
In practice in LTE, all the uplink data transmissions are localized, using contiguous blocks of
subcarriers. Frequency-diversity is achieved by frequency hopping, which can occur both within one
subframe (at the boundary between the two slots) and between subframes. In the case of frequency
hopping within a subframe, the channel coding spans the two transmission frequencies, and therefore
the frequency diversity gain is maximized through the channel decoding process. The only instance of
distributed transmission in the LTE uplink (using an IFDMA-like structure) is for the „Sounding
Reference Signals‟ (SRSs) which are transmitted to enable the eNodeB to perform uplink frequency-
selective scheduling.
System bandwidth is scalable from approximately 1.4 MHz up to 20 MHz with the same subcarrier
spacing and symbol duration for all bandwidths. Sampling rates resulting from FFT sizes are designed
to be small rational multiples of UMTS 3.84 MHz chip rate, for ease of implementation in a multimode
UE.
Pulse Shaping
In SC-FDMA, there is no need for explicit pulse-shaping thanks to the implicit „sinc‟ pulse-shaping.
Nevertheless, an additional explicit pulse-shaping filter can further reduce the CM/PAPR, but at the
expense of spectral efficiency. As a result of this trade-off, additional pulse shaping is not specified in
LTE.
The important properties of the SC-FDMA transmission scheme used for the LTE uplink are derived
from its multicarrier OFDM-like structure with single-carrier characteristic. The multicarrier-based
structure gives the LTE uplink the same robustness against ISI as the LTE downlink, with low-
complexity frequency-domain equalization being facilitated by the CP. At the same time, the DFT-
based pre-coding ensures that the LTE uplink possesses the low CM required for efficient UE design.
Crucially, LTE uplink is designed to be orthogonal in the frequency domain between different UEs,
thus virtually eliminating the intra-cell interference associated with CDMA.
The parameters of the LTE uplink are designed to ensure maximum commonality with the downlink,
and to facilitate frequency-domain DFT-S-OFDM signal generation. The localized resource allocation
scheme of the LTE uplink allows both frequency selective scheduling and the exploitation of
frequency diversity, the latter being achieved by means of frequency hopping.
The baseband SC-FDMA transmit signal for SC-FDMA symbol is of the form,
for 0 ≤t <(NCP, + N)Ts, where NCP= #samples of CP in SC-FDMA symbol (see Section 15.3), N =
RB
2048 is IFFT size, df = 15 kHz subcarrier spacing, Ts = 1/(N·df ) is the sampling interval, N is
(−)
system BW in RBs, Nsc = 12 subcarriers per RB, k = k + [NRBNsc /2] and ak,l is the content of
subcarrier k on symbol l. For PUSCH data, ak,l (k=0,1,2….M-1) is obtained by DFT-spreading the
data QAM symbols, [d0,l, d1,l, . . . , dM−1,l] to be transmitted on data symbol l,
A Hybrid Automatic Repeat reQuest (HARQ) is used synchronously, using N-channel stop and wait
(retransmissions occur in specific periodically-occurring subframes (HARQ channels)).
Frequency-Selective Scheduling
Localized allocation is used in both slots of a subframe –no frequency hopping during a subframe. RB
allocation and MCS are chosen based on location and quality in UL channel response. Timely
1. Intra-subframe hopping - frequency hop occurs at the slot boundary in the middle of a
subframe; this provides frequency diversity within a codeword (within TB).
2. Inter-subframe hopping - frequency diversity between HARQ retransmissions of TB, as the
frequency allocation hops every allocated subframe.
Either (1) a pre-determined pseudo-random frequency hopping pattern, or (2) an explicit hopping
offset signalled in the UL resource grant on the PDCCH.
For NRB < 50RBs, size of hopping offset is approximately NRB/2, &
For NRB >50 RBs, possible hopping offsets are NRB /2, and +- NRB/4..
In UL, resource grant indicated frequency hopping. Semiselective scheduling is when the frequency
resource is assigned for the first slot and frequency diversity is also achieved by hopping to a different
frequency in the second slot.
RBs (in each slot) that can be used for PUCCH transmission within cell is NRB(PUCCH) („pusch-
HoppingOffset’), indicated by SIB. PUCCH RBs per slot = number of PUCCH regions per subframe.
In case of odd numbers of PUCCH RB pair used, the unused RB pair may be used for PUSCH or
PUCCH may be hopped in that slot.
Both 1 and 2-bit acknowledgements are supported using BPSK and QPSK respectively. The HARQ
ACK/NACK bits (which are not scrambled) are BPSK/QPSK modulated according to the modulation
mapping resulting in a single ACK/NACK modulation symbol. ACK is encoded as „1‟ and NACK as „0‟.
The modulation mapping is the same formats 2a/2b.
Just as CQI transmission, the ACK/NACK symbol (which is phase-rotated by 90 degrees in the
second slot) is transmitted on each SC-FDMA data symbol by modulating a cyclic time shift of the
base RS sequence of length-12 prior to OFDM modulation. Time-domain spreading with orthogonal
spreading codes is used to code-division-multiplex UEs. Thus, a large number of UEs (data and RSs)
can be multiplexed on the same PUCCH RB using frequency-domain and time-domain code
multiplexing. For cyclic time shift multiplexing, cell-specific dPUCCH shift ∈ {1, 2, 3}, indicating 12, 6, or 4
shifts respectively, and selected based on expected delay spread in the cell.
For time-domain spreading CDM, spreading codes are limited by number of RS symbols, as
multiplexing capacity of RS is smaller than that of the data symbols. For example, in the case of six
supportable cyclic time shifts and three (or two) orthogonal time spreading codes in normal (or
extended) CP with three (or two) RS symbols, acknowledgments from 6x3=18 (or 12) different UEs
can be multiplexed within one PUCCH RB. The length-2/4 orthogonal block spreading codes are
based on Walsh–Hadamard codes, and the length-3 spreading codes are based on DFT codes. A
subset of size-s orthogonal spreading codes of a particular length L (s ≤ L) is used depending on the
number of RS SC-FDMA symbols. For normal CP with four data SC-FDMA symbols and three
orthogonal time spreading codes (due to there RS symbols), indices 0, 1, 2 of the length-4 orthogonal
spreading codes are used for data time-domain block spreading.
The ACK/NACK and SRS may be configured in the same subframe configured by SI. One option is
for the ACK/NACK to take precedence over the SRS, only HARQ is transmitted in the relevant
subframe. The alternative is to configure UEs to use shortened PUCCH in such subframes, whereby
the last SC-FDMA symbol in the second slot of the subframe is not transmitted. This is known as a
shortened PUCCH format, length of the timedomain orthogonal block spreading code is reduced by
one (compared to the first slot). Hence, it uses the length-3 DFT spreading codes in place of length-4
Walsh–Hadamard codes. A UE may not simultaneously transmit on SRS and PUCCH or PUSCH to
avoid violating the single-carrier nature of the signal. Therefore, a PUCCH or PUSCH symbol may be
punctured if SRS is transmitted.
The number of HARQ resource-index NPUCCH,RB= c.P
where c=3(normal CP) or 2(extended CP)
P=12/(dPUCCHshift(1,2 or 3)).
As in CQI, cyclic time shift hopping is used for HARQ. In SPS, PDSCH without a DL grant on PDCCH,
PUCCH ACK/NACK resource index n(1)PUCCH should to be used by a UE for initial HARQ. For
dynamically DL scheduling, (including HARQ retransmissions for SPS) on PDSCH (indicated by DL
assignment on PDCCH), PUCCH HARQ resource index n(1)PUCCH is determined based on the index
of the first Control Channel Element (CCE) of DL control assignment. PUCCH region m used for
HARQ with format 1/1a/1b for the case with no mixed PUCCH region is given by
(2)
m = (nPUCCH/NPUCCH,RB) + N RB (RB for PUCCH format 2a/2b).
The PUCCH resource index n(1)(ns ), corresponding to a combination of a cyclic time shift and
orthogonal code (nPUCCH RS and noc), within the PUCCH region m in even slots is given by
n(1)(ns) = n(1)PUCCH mod N(1)PUCCH,RB for ns mod 2 = 0
The PUCCH resources are first indexed in cyclic time shift domain, followed by orthogonal time
spreading code domain. The cyclic time shifts used on adjacent orthogonal codes can also be
staggered, providing the opportunity to separate the channel estimates prior to de-spreading. As high
Doppler breaks down the orthogonality between the spread blocks, offsetting the cyclic time shift
values within each SC-FDMA symbol can restore orthogonality at moderate delay spreads. To
randomize intra-cell interference, PUCCH resource index remapping is used in the second slot. Index
remapping includes both cyclic shift remapping and orthogonal block spreading code remapping. The
PUCCH resource index remapping function in an odd slot is based on the PUCCH resource index in
the even slot of the subframe.
The RE used for each of CQI/PMI, ACK/NACK and RI is based on the MCS assigned for PUSCH and
an offset parameter, ∂CQIoffset, ∂HARQ-ACKoffset , or ∂RIoffset, configured by higher-layer, different code
rates can be used for control. PUSCH data and control are mapped to different RE‟s. Controls are
mapped to be present in both slots. eNodeB has prior knowledge of UL control signaling, it can easily
de-multiplex control and data. CQI/PMI is placed at the beginning of UL-SCH data RE and mapped
sequentially to all symbols on one subcarrier before continuing on next subcarrier. UL-SCH data is
rate-matched around CQI/PMI data with same MCS. For small CQI and/or PMI report sizes up to 11
bits, a (32,k) block code, similar to the one used for PUCCH, is used, with optional circular repetition
of encoded data; no CRC is applied. For large CQI reporting modes (> 11 bits), 8-bit CRC is attached
and channel coding and rate matching is performed using the tail-biting convolutional code.
ACK/NACK resources are mapped, by puncturing PUSCH data, positions next to the RS, to benefit
from the best possible channel estimation. The maximum RE for ACK/NACK is 4 symbols.
The coded RI are placed next to ACK/NACK symbol positions irrespective of whether ACK/NACK is
actually present in a given subframe. The modulation of ACK/NACK or RI is such that the distance of
symbols carrying ACK/NACK and RI is maximized. The outermost constellation points of the higher-
order 16/64-QAM PUSCH modulations are used, resulting in increased transmit power for
ACK/NACK/RI relative to the average PUSCH data power. The coding of RI and CQI/PMI are
separate, with UL-SCH data rate-matched around RI similarly to CQI/PMI. For 1-bit ACK/NACK or RI,
repetition coding is used. For 2-bit ACK/NACK/RI, a (3, 2) simplex code is used with optional circular
repetition of encoded data. The resulting code achieves the theoretical maximum values of the
minimum Hamming distance of the output codewords in an efficient way.
The main uplink physical channels are PUSCH and PUCCH. PUSCH supports both frequency-
selective scheduling and frequency-diverse(hopping) transmissions.
Control signalling (consisting of ACK/NACK, CQI/PMI and RI) is carried by PUCCH when no PUSCH
allocated. PUCCH is deliberately mapped to edge RB, to reduce out-of-band emissions caused by
RSs are time-multiplexed with data symbols. The DM RSs of a given UE occupy the same RBs as its
PUSCH/PUCCH data transmission, hence each UE RS‟s are orthogonal. to each other. The SRSs, if
configured, are transmitted on the last symbol in a subframe; SRS can occupy a bandwidth different
from that used for data transmission. UEs transmitting SRS in the same subframe can be multiplexed
via either FDMA or CDMA.
RS characteristics:
1. Constant amplitude in all allocated subcarriers for unbiased channel estimates.
2. Low Cubic Metric (CM) (at worst no higher than data transmissions).
3. Good autocorrelation for accurate channel estimation.
4. Good cross-correlation properties between different RSs in other (or same) cells.
RS sequence length, Np, = number of subcarriers, ( multiple of subcarriers per RB, NSC(/RB) = 12)
Np = MSC(/RS) = m · NSC(/RB) 1 ≤ m ≤ NRB(/UL)(uplink system bandwidth in terms of RBs).
The length-Np RS sequence is directly applied (without DFT spreading) to Np RS subcarriers at the
input of IFFT.
Cyclic extension preserves the constant amplitude and zero autocorrelation cyclic shift orthogonality.
Cyclic extension of ZC sequences is used rather than truncation, as in general it provides better CM
characteristics. For sequence lengths of three or more RBs, this provides at least 30 sequences with
CM smaller than or close to that of QPSK.
30 special RS sequences are defined in LTE for resource allocations of one or two RBs, as there are
very few ZC sequences within 2 RB. These special sequences are QPSK rather than ZC-based
jϕ(n)π/4
sequences. The QPSK RS sequences are given by r(n) = e , n= 0, 1, . . . ,MSC(/RS)− 1, where
MSC(/RS) =#subcarriers.
.
The base sequences for resource allocations larger than three RBs are selected such that they are
the sequences with high cross-correlation to the single 3 RB base sequence in the sequence-group.
Since cross-correlation between the 3 RB base sequences of different sequence-groups is low due to
the inherent properties of the ZC sequences, such a method for assigning the longer base sequences
to sequence-groups helps to ensure that the cross correlation between sequence-groups is kept low,
thus reducing inter-cell interference.
The v base RS sequences of length 3 RBs or larger (i.e.MRS sc ≥ 36) assigned to a sequence-group
u are given by,
ru,v(n) = aq (n mod NZC), n = 0, 1, . . . ,MRS sc − 1
where u ∈ {0, 1, . . . , 29} is the sequence-group number,
v is the index of the base sequence of length MRSsc within the sequence-group u, given by
v = 0, 1 forMRSsc ≥ 72
= 0 otherwise
NZC is the largest prime number smaller than MRSsc , and
q is the root ZC sequence index.
If RS SC-FDMA symbol duration is Tp and channel impulse response duration is less than Tcs, then
up to Tp/Tcs different transmitters can transmit in the same symbol, with different cyclic shift values,
with separable channel estimates at the receiver. For example, if Tp/Tcs = 4 and there are four
transmitters, then each transmitter t ∈ {1, . . . , 4} can use a cyclic time shift of (t − 1)Tp/4 of the same
base sequence. At the eNodeB receiver, by correlating the received composite signal from the
different transmitters occupying the same set of subcarriers with the base sequence, the channel
estimates from the different transmitters are separable in the time domain.
Since a cyclic time shift is equivalent to applying a phase ramp in the frequency domain, the
frequency-domain representation of a base sequence with cyclic shift, α, is given by
r(α) u,v (n) = ejαn .ru,v(n)
where .ru,v(n) is the base (or unshifted) sequence of sequence-group u, with base sequence index v
within the sequence-group, α = 2πnt/P with nt the cyclic time shift index for transmitter t, and P is the
number of equally spaced cyclic time shifts supported. In the example, P =4 and n = 0..3 for the four
transmitters respectively.
In LTE, 12 equally spaced cyclic time shifts are defined for DM RS on PUSCH and PUCCH. This
allows for delay spreads up to 5.55 μs (total 1 symb duration, 12*5.55=.66.6 μs).
Cyclic time shifts spaced the furthest apart experience the least cross-talk between the channel
estimates. Thus, when the number of UEs using different cyclic time shifts is less than the number of
cyclic time shifts supported (P), it is beneficial to assign cyclic time shifts with the largest possible
(circular) separation, and this is supported in LTE. Alternatively, cyclic time shift hopping can be
employed for UE. In LTE, hopping of the cyclic time shifts between the two slots in a subframe is
always enabled.
Sequence-Group Hopping
It is enabled in a cell by 1-bit broadcast parameter „groupHoppingEnabled‟. This mode is a
combination of hopping and shifting of the sequence-group according to one of 504 sequence-group
hopping/shifting patterns corresponding to the 504 unique cell-IDs. Since there are 30 base
sequence-groups, 17(= 504/30) unique sequence-group hopping patterns of length 20 are defined
(corresponding to 20 slots in a frame), each of which can be offset by one of 30 sequence-group shift
offsets. The sequence-group number u depends on the sequence-group hopping pattern fgh (0..16) and
the sequence-group shift offset fss. The sequence-group hopping pattern changes u from slot to slot in
a pseudo-random manner, while the shift offset fss is fixed in all slots. Both fgh and fss depend on the
cell-ID.
The pattern fgh is obtained from a length-31 Gold sequence generator, of which the second
constituent M-sequence is initialized at the beginning of each radio frame by fgh of the cell. Up to 30
cell-IDs can have the same fgh (planned coordinated cell cluster), with different fss. The same fgh is
used for PUSCH DM RS, SRS and PUCCH DM RS. The fss can be different for PUSCH and PUCCH.
For PUSCH, plan to assign cell-IDs such that the same fgh and fss, and hence same base sequences,
are used in adjacent cells. This can enable RSs from UEs in adjacent cells to be orthogonal to each
other by using different cyclic time shifts of the same base sequence. Therefore fss for PUSCH is
explicitly configured by 5bit, „groupAssignmentPUSCH‟(dss) such that fss = (cell-ID mod 30 + dss)
mod 30, where dss ∈ {0, . . . , 29}.
There are two base sequences per sequence-group for each RS sequence length greater than 60 (5
RBs), with possibility of sequence-hopping between two base sequences at the slot boundary in the
middle of each subframe. If sequence-group hopping is used, base sequence automatically changes
between each slot, and therefore additional sequence hopping within the sequence group is not
needed; therefore only the first base sequence in the sequence group is used if sequence-group
hopping is enabled.
Sequence-Group Planning
If sequence-group hopping is disabled, the same sequence-group number u, is used in all slots and u
= fss. Planned sequence-group assignment is possible for up to 30 cells in LTE. In this case,
neighbouring cells to be assigned sequence groups u with low cross-correlation.
The same sequence-group number (base sequences) are used in the three cells of each eNodeB,
with different cyclic time shifts assigned to each cell. With sequence-group planning, sequence
hopping within group between two slots of a subframe is enabled by 1-bit parameter,
„sequenceHoppingEnabled‟. The base sequence index for MRS sc ≥ 72 used in slot ns is then
obtained from length-31 Gold sequence generator. In order to enable the use of the same base RS
sequence (and hopping pattern) in adjacent cells for PUSCH, the pseudo-random sequence
generator is initialized at the beginning of each radio frame by the sequence-group hopping pattern
index (based on part of the cell-ID), offset by the PUSCH sequence-group shift index of the cell.
For PUCCH, cyclic time shift hopping (among P = 12 evenly spaced cyclic time shifts) is performed
per SC-FDMA symbol, with cyclic shift α for a given SC-FDMA symbol in a given slot being derived
from a combination of the assigned PUCCH resource index and the output of the length-31 Gold
sequence generator. In order to randomize interference on the PUCCH arising from the fact that the
same band-edge RBs are used for PUCCH transmissions in all cells, the pseudo-random sequence
generator is initialized at the beginning of each radio frame by the cell-ID.
In addition, to achieve intra-cell interference randomization for the PUCCH DM RS, the cyclic time
shift used in the second slot is hopped such that UEs which are assigned adjacent cyclic time shifts in
the first slot use non-adjacent cyclic time shifts (with large separation) in the second slot. A further
benefit of using a different cyclic time shift in each slot is that the non-ideal cross-correlation between
different base RS sequences is averaged (as the cross-correlation is not constant for all time lags).
RS Symbol Duration
The LB RS structure was adopted in LTE for the PUSCH DM RS. The exact position of the single
PUSCH DM RS symbol in each uplink slot depends on whether the normal or extended CP is used.
rd
For normal CP with seven SC-FDMA symbols per slot, the PUSCH DM RS occupies the centre (3 ).
nd
With six SC-FDMA symbols per slot. In extended CP, the 2 SC-FDMA symbol is used. For PUCCH,
the position and number of DM RS depends on the type of PUCCH format.
The DM RS occupies the same RBs as RB allocation for PUSCH or PUCCH. Thus, RS sequence
length, MRSsc = #subcarriers for PUSCH or PUCCH. Since PUSCH RB size is multiples of two, three
and/or five RBs, DM RS sequence lengths are also restricted to the same multiples. For interference
randomization, cyclic time shift hopping is always enabled for DM RS.
SRS Bandwidths
Some of the factors which affect the SRS bandwidth are max power of the UE, number of supportable
sounding UEs, and the sounding bandwidth needed to benefit from uplink channel-dependent
scheduling. Full bandwidth sounding provides the most complete channel information when the UE is
sufficiently close to the eNodeB, but degrades as the path-loss increases when the UE cannot further
increase its transmit power to maintain the transmission across the full bandwidth. Full bandwidth
transmission of SRS also limits the number of simultaneous UEs whose channels can be sounded,
due to the limited number of cyclic time shifts (eight cyclic time shifts per SRS-comb).
To improve the SNR and support a larger number of SRS, up to four SRS bandwidths can be
simultaneously supported in LTE depending on the system bandwidth. To provide flexibility with the
values for the SRS bandwidths, eight sets of four SRS bandwidths are defined for each possible
system bandwidth. RRC signalling indicates which of the eight sets is applicable in the cell by means
of a 3-bit cell-specific parameter „srsBandwidthConfiguration’. This allows some variability in the
maximum SRS bandwidths, which is important as the SRS region does not include the PUCCH region
near the edges of the system bandwidth, which is itself variable in bandwidth.
The specific SRS bandwidth is configured by a further 2-bit UE-specific parameter, „srsBandwidth’.
The smallest sounding bandwidth supported is 4 RBs. Frequency hopping can be enabled or disabled
for an individual UE based on „frequencyDomainPosition‟. The tree structure of the SRS bandwidths
limits the possible starting positions for the different SRS bandwidths, reducing the overhead for
signalling the starting position to 5 bits (for each UE by „Frequency-domain-position’.
UL RS provided in LTE fulfil an important function in facilitating channel estimation and channel
sounding. The ZC-based sequence design can be seen to be a good match to this role, with constant
amplitude in the frequency domain and the ability to provide a large number of sequences with zero or
low correlation. This enables both interference randomization and interference coordination
techniques to be employed in LTE system deployments, as appropriate to the scenario. A high degree
of flexibility is provided for configuring the reference signals, especially for the sounding reference
signals, where the overhead arising from their transmission can be traded off against the
improvements in system efficiency which may be achievable from frequency-selective uplink
scheduling.
Uplink cell capacity is constrained by interference levels from other active UEs. The ratio between the
total received power spectral density Io(signal+interference), and thermal noise level N0 at eNodeB
receiver j in the time-frequency region k composed of NULsymb symbols and NULsc subcarriers, is
denoted IoT (j, k) and defined as IoT (j, k) = (Io(j, k) + N0)/N0. Without loss of generality the indices j
and k can be dropped, and IoT represents the level for an arbitrary time-frequency region.
Cell IoT levels must be managed to maintain cell-edge coverage for uplink control channels including
random access channels (vital for scheduling data channels and handovers), as well as maintaining
the minimum cell edge data rates for crucial services such as VoIP.
Average IoT varies versus VoIP loading with two different scheduling methods, A(SPS) and
B(Persistent) with a signalled resource assignment bitmap. As the load increases, it becomes more
likely that the same time and frequency resource regions will be occupied in neighbouring cells, such
that average IoT increases.
For very high SINR scenarios, the supportable VoIP users in a cell reaches a plateau in case of 1 ms
TTI, as the minimum RB allocation is larger than necessary to transmit a single VoIP packet. To
optimize this a RB is limited to 12 subcarrier. Higher subcarrier in a RB will give opposite effect of the
benefits of „TTI bundling‟, whereby a long TTI has the potential to increase the received energy per
packet, and therefore to improve coverage at the cell edge when the UE is power-limited. The choice
of a 1 ms TTI length in LTE is therefore a compromise between high capacity in high-SINR
conditions and good coverage at the cell edge. A choice of a small RB size (only 12 subcarriers)
helps to ensure that the minimum resource allocation size does not unduly limit capacity.
The link budget results are for deployment scenario Case 3, and assume a log-normal
shadowing margin of 12.1 dB corresponding to 98-percentile single-cell area coverage
reliability and a propagation model given by
Propagation loss = 128.1 + 37.6 log(distance(m)) (18.7)
Only an average throughput of 5 kbps can be supported based on a single (1 ms) RB PUSCH
transmission. The downlink SINR at 1000m given the corresponding downlink conditions is about −8.3
dB, which is also the SINR required for a 1% BLER on the Physical Broadcast CHannel (PBCH)
SINR. At 1000 m the transmission loss (propagation loss minus the antenna gains plus the log-normal
shadowing margin plus the penetration/body loss) is 146.2 dB, which must be supported by the uplink
and downlink control channels in order to achieve the 98-percentile area coverage reliability.
3. PRACH coverage.
A repeated RACH preamble burst (2 × 800 μs) is needed for the PRACH to achieve 98-
percentile or better area coverage reliability, since PRACH format 2, only supports a cell
radius of about 0.8 km which is close to the range supported by one Zadoff–Chu root
sequence (0.78 km). With a single sequence a total received preamble energy per sequence
of approximately 18 dB (Es/N0 ∼−11.5 dB) is required to meet missed detection and false
alarm probabilities of less than 1%. PRACH format 2 with repetition is slightly better with a
required Es/N0 =−13.5 dB for the same error probabilities.
We highlight the main factors affecting uplink capacity and coverage. It is possible to observe how
uplink coverage issues have led to particular design choices in LTE, such as the size of a RB, the
length of TTI, and the design of the control channels. Evaluations can be carried out to examine the
effect of each relevant factor, leading to the LTE performance being able to be characterized by a
variety of metrics. Such metrics include average throughput, cell-edge throughput, number of VoIP
users, and FTP download time. Depending on the requirements of particular deployments, the
eNodeB has the freedom to balance average throughput against cell-edge coverage and fairness.
The initial preamble Tx power setting is based on open-loop estimation with full compensation for the
path-loss, so that received power of the preambles is independent of the path-loss; designed to help
the eNB to detect several simultaneous preamble Tx in the same PRACH RB resource.
UE estimates the path-loss by DL RSRP. The eNodeB may also configure an additional power offset.
The eNodeB may configure preamble power ramping so that the transmission power for each
retransmitted preamble is increased by a fixed step. RA preambles are normally orthogonal to other
uplink transmissions. Therefore, the proportion of RA attempts which succeed at the first preamble
transmission is likely to be higher than in WCDMA, and the need for power ramping is likely to be
reduced.
PRACH Formats
Fig 3.5.3.1 – Preable format versus Cell Radius Max and Recommended
Sequence Duration
The sequence duration, TSEQ, is driven by the following factors:
1. Trade-off between sequence length and overhead: a single sequence must be as long as
possible to maximize the number of orthogonal preambles, while still fitting within a single
subframe to keep PRACH overhead small;
2. Compatibility with the maximum expected round-trip delay;
3. Compatibility between PRACH and PUSCH subcarrier spacings;
4. Coverage performance.
where TSYM = 66.67 μs is the symbol duration. In other words, the preamble duration must be an
integer multiple of UL subframe symbol duration:
TSEQ = kTSYM = k/ df , k∈ N. Here k=12 makes Tseq=12*66.67 = 800μs.
The FFT/IFFT components can be reused from SC-FDMA processing for the scheduled data. For
example, an n · 2m DFT can be implemented with an FFT of 2m samples combined with a DFT of n
samples, since NDFT = kfsTSYM = kNFFT, k∈ N where NFFT is the FFT size for a PUSCH symbol.
Generic Parameters:
Parameter Value
-------------- ---------
Carrier frequency (f ) 2000 MHz
Antenna height (hb) 30 m / 60 m
UE antenna height (hm) 1.5 m
UE transmitter EIRPa (Pmax) 24 dBm (250 mW)
eNodeB Receiver Antenna Gain (Ga) 14dBi
Receiver noise figure (Nf) 5.0 dB
Thermal noise density (N0) −174 dBm/Hz
Percentage of area covered by buildings (α) 10%
Required Ep/N0 (eNodeB with 2 Rx antenna) 18 dB (six-path Typical Urban
channel model)
Penetration loss (PL) 0dB
Log-normal fade margin (LF) 0 dB
PRACH preamble sequence duration TSEQ is then derived from the required preamble sequence
energy to thermal noise ratio Ep/N0 to meet a target missed detection and false alarm probability, as
follows:
where N0 is the thermal noise power density (in mW/Hz) and Nf is the receiver noise figure (in linear
scale). Assuming that Ep/N0 = 18 dB is required to meet missed detection and false alarm
probabilities of 10−2 to 10−3, It is observed that the potential coverage performance of a 1 ms PRACH
preamble is in the region of 14 km. As a consequence, the required CP and GT lengths are
approximately (2 · 14000)/(3 · 108) = 93.3 μs, so that the upper bound for TSEQ is given by TSEQ ≤
1000 − 2 · 93.33 = 813 μs.
Therefore, the longest sequence is TSEQ = 800 μs, as used for preamble formats 0 and 1.
The resulting PRACH subcarrier spacing is dfRA = 1/TSEQ = 1.25 kHz. The 1600 μs preamble
sequence of formats 2 and 3 is implemented by repeating the baseline 800 μs preamble sequence.
These formats can provide up to 3 dB link budget improvement, which is useful in large cells and/or to
balance PUSCH/PUCCH and PRACH coverage at low data rates.
CP and GT Duration
Having chosen TSEQ, the CP and GT dimensioning can be specified more precisely. For formats 0
and 2, the CP is dimensioned to maximize the coverage, given a maximum delay spread d: TCP =
(1000 − 800)/2 + d/2 μs, with d ≈ 5.2 μs (corresponding to the longest normal CP of a PUSCH
symbol). The maximum delay spread is used as a guard period at the end of CP, thus providing
protection against multipath interference even for the cell-edge UEs.
For a cell-edge UE, the delay spread energy at the end of the preamble is replicated at the end of the
CP and is therefore within the observation interval. Consequently, there is no need to include the
maximum delay spread in the GT dimensioning. Hence, instead of locating the sequence in the centre
of the PRACH slot, it is shifted later by half the maximum delay spread, allowing the maximum
Round-Trip Delay (RTD) to be increased by the same amount. The residual delay spread at the end
of the preamble from a cell-edge UE spills over into the next subframe, but this is taken care of by the
CP at the start of the next subframe to avoid any inter-symbol interference.
For formats 1 and 3, the CP is dimensioned to address the maximum cell range in LTE, 100 km, with
a maximum delay spread of d ≈ 16.67 μs. In practice, format 1 is expected to be used with a 3-
Extending this principle, the available slot configurations are designed to facilitate a PRACH receiver
which may be used for multiple cells of an eNodeB, assuming a periodic pattern with period 10 ms or
20 ms.
Assuming an operating collision probability per UE, pcoll = 1%, one PRACH resource (with 64
signatures) per 10 ms per 5 MHz can handle an offered load G=−64 ln(1 − pcoll) = 0.6432 average
PRACH attempts, which translates into 128 attempts per second in 10 MHz. This is expected to be a
typical PRACH load in LTE.
Assuming a typical PRACH load, Resource configurations 0–2 and 15 use a 20 ms PRACH period,
which can be desirable for small BW (1.4 MHz) in order to reduce the PRACH overhead at the price
of higher waiting times.
Preamble Bandwidth
To ease multiplexing of PRACH and PUSCH RB, a PRACH slot must be allocated a bandwidth
BWPRACH equal to an integer multiple of Resource Blocks (RBs), i.e. an integer multiple of 180 kHz.
For simplicity, BWPRACH in LTE is constant for all BW; to optimize both detection performance and
timing estimation accuracy. Timing estimation drives the lower bound of the PRACH bandwidth.
Indeed, a minimum bandwidth of approx 1 MHz is necessary to provide a one-shot accuracy of about
±0.5 μs, which is an acceptable timing accuracy for PUCCH/PUSCH.
We have observed that the best detection performance is achieved by preambles of 6 RBs and 12
RBs for low and high SNRs respectively. The 25-RB preamble has the overall best performance
considering the whole SNR range. Thus the diversity gain of large bandwidths only compensates the
increased detection threshold in the high SNR region corresponding to misdetection performances in
the range of 10−3 and below. At a typical 10−2 detection probability target, the 6-RB allocation only has
0.5 dB degradation with respect to the best case.
Therefore, a PRACH allocation of 6 RBs provides a good trade-off between PRACH overhead,
detection performance and timing estimation accuracy. Note that for the smallest system bandwidth
(1.4 MHz, 6 RBs) the PRACH overlaps with the PUCCH; it is left to the eNodeB implementation
whether to implement scheduling restrictions during PRACH slots to avoid collisions, or to let PRACH
collide with PUCCH and handle the resulting interference. Finally, the exact preamble transmission
bandwidth is adjusted to isolate PRACH slots from surrounding PUSCH/PUCCH allocations through
guard bands, as elaborated in the following section.
Sequence Length
The sequence length design should address the following requirements:
1. Maximize the number of ZC sequences with optimal cross-correlation properties;
2. Minimize the interference to/from the surrounding scheduled data on the PUSCH.
The interference from PUSCH to PRACH is further amplified by the fact that the operating Es/N0 of
PUSCH (where Es is the PUSCH symbol energy) is much greater than that of the PRACH (typically as
much as 24 dB greater if we assume 13 dB Es/N0 for 16QAM PUSCH, while the equivalent ratio for
the PRACH would be −11 dB assuming Ep/N0 = 18 dB and adjusting by −10 log10 (864) to account for
the sequence length).
The PRACH uses guard bands to avoid the data interference at preamble edges. A cautious design of
preamble sequence length not only retains a high inherent processing gain, but also allows avoidance
of strong data interference. In addition, the loss of spectral efficiency (by reservation of guard
subcarriers) can also be well controlled at a fine granularity (dfRA = 1.25 kHz).
In the absence of interference, there is no significant performance difference between sequences of
similar prime length. In the presence of interference, it can be seen that reducing the sequence length
below 839 gives no further improvement in detection rate. No effect is observed on the false alarm
rate.
Therefore the sequence length of 839 is selected for LTE PRACH, corresponding to 69.91 PUSCH
subcarriers in each symbol, and offers 72 − 69.91 = 2.09 PUSCH subcarriers protection, which is very
close to one PUSCH subcarrier protection on each side of the preamble. Note that the preamble is
positioned centrally in the block of 864 available PRACH subcarriers, with 12.5 null subcarriers on
each side.
Larger the cell, the larger the cyclic shift required to generate orthogonal sequences, and
consequently, the larger the number of ZC root sequences necessary to provide the 64 required
preambles. The relationship between cell size and the required number of ZC root sequences allows
for some system optimization. In general, the eNodeB should configure NCS independently in each
cell, because the expected inter-cell interference and load (user density) increases as cell size
decreases; therefore smaller cells need more protection from co-preamble interference than larger
cells.
Some examples may show different scenarios with different numbers of parameters such as:
(Number of cyclic shifts per ZC sequence * Number of ZC sequences) >= 64, Cyclic shift size Ncs, Cell Radius
etc. For each scenario, the total number of sequences is 64, but resulting from different combinations of
the number of root sequences and cyclic shifts.
There can be some of the following cases:
1. Case 1: Only one UE transmits a preamble;
2. Case 2: Two UEs transmit a preamble, and the two preamble sequences are generated from
the same root ZC sequence;
Figure shows the range of NCS values and their usage with the various preamble formats. NCS values
are designed for use in low-speed cells.
As can be observed, frequency offsets as large as one PRACH subcarrier (δf =±_fRA =
±1/TSEQ = ±1.25 kHz) result in cyclic shifts on the ZC sequence. This frequency offset δf can be due
to the accumulated frequency uncertainties at both UE transmitter and eNodeB receiver, δfLO, and the
Doppler shift resulting from the UE motion in a Line of Sight (LOS) radio propagation. The impact of
the cyclic shift distortion on the received Power Delay Profile (PDP) is, it creates false alarm peaks
whose relative amplitude to the correct peak depends on the frequency offset. The solution adopted in
LTE to address this issue is referred to as „cyclic shift restriction‟ and consists of „masking‟ some
cyclic shift positions in the ZC root sequence. This makes it possible to retain an acceptable false
alarm rate, while also combining the PDPs of the three uncertainty windows, thus also maintaining a
high detection performance even for very high-speed UEs.
It should be noted that at |δf| = _fRA, the preamble peak completely disappears at the desired
location. However, the false image peak begins to appear even with |δf |<_fRA. Another impact of the
side peaks is that they restrict the possible cyclic shift range so as to prevent from side peaks from
falling into the cyclic shift region.
We use C−1 and C+1 to denote the two wrong cyclic shift windows arising from the frequency offset,
while C0 denotes the correct cyclic shift window. The cyclic restriction rule must be such that the two
Sequence Ordering
A UE using the contention-based random access procedure needs to know which sequences are
available to select from. The full set of 64 sequences may require the use of several ZC root
sequences, the identity of each of which must be broadcast in the cell. Given the existence of 838
root sequences, signalling each individual sequence index requires 10 bits per root sequence, which
can lead to a large signalling overhead. Therefore in LTE the signalling is streamlined by broadcasting
only the index of the first root sequence in a cell, and the UE derives the other preamble signatures
from it given a predefined ordering of all the sequences.
Two factors are taken into account for the root sequence ordering, namely the CM of the sequence,
and maximum supportable cell size for high-speed cells (or equivalently the maximum supported
cyclic shift). Since CM has a direct impact on cell coverage, the first step in ordering the root
sequences is to divide the 838 sequences into a low CM group and a high CM group, using the CM of
QPSK (1.2 dB) as a threshold. The low CM group would be used first in sequence planning (and also
for high-speed cells) since it is more favourable for coverage.
Then, within each CM group, the root sequences are classified into subgroups based on their
maximum supportable cell radius, to facilitate sequence planning including high-speed cells.
Specifically, a sequence subgroup g is the set of all root sequences with their maximum allowed cyclic
shifts (Smax) lying between two consecutive high-speed NCS values according to
Signature Detection
Collision detection.
In any cell, the eNodeB can be made aware of the maximum expected delay spread. As a result,
whenever the cell size is more than twice the distance corresponding to the maximum delay spread,
the eNodeB may in some circumstances be able to differentiate the PRACH transmissions of two UEs
if they appear distinctly apart in the PDP. Collision detection is never possible, while the lower PDP
represents a larger cell where it may sometimes be possible to detect two distinct preambles within
the same ZCZ. If an eNodeB detects a collision, it would not send any random access response, and
the colliding UEs would each randomly reselect their signatures and retransmit.
Timing Estimation
The primary role of the PRACH preamble is to enable the eNodeB to estimate a UE‟s transmission
timing. One can observe that the timing of 95% of UEs can be estimated to within 0.5 μs, and more
than 98% within 1 μs. No collision detection algorithm is implemented here. The IFFT size is 2048 and
the system sampling rate 7.68 MHz, giving an oversampling rate of 2.44.
Preamble Format 4
For preamble format 4, a ZC sequence of length 139 is used. The preamble starts 157 μs before the
end of the UpPTS field at the UE. Unlike preamble formats 0 to 3, a restricted preamble set for high-
speed cells is not necessary for preamble format 4, which uses a 7.5 kHz subcarrier spacing. With a
random access duration of two OFDM symbols (157 μs), the preamble format 4 is mainly used for
small cells with a cell radius less than 1.5 km, and where cyclic shift restrictions for high UE velocities
It can be seen how the PRACH preamble addresses the high performance targets of LTE, such as
high user density, very large cells, very high speed, low latency and a plurality of use cases, while
fitting with minimum overhead within the uplink SC-FDMA transmission scheme. Many of these
aspects benefit from the choice of ZC sequences for the PRACH preamble sequences in place of the
pseudo-noise sequences used in earlier systems. The properties of these sequences enable
substantial numbers of orthogonal preambles to be transmitted simultaneously.
Considerable flexibility exists in the selection of the PRACH slot formats and cyclic shifts of the ZC
sequences to enable the LTE PRACH to be dimensioned appropriately for different cell radii and
loadings. Some options for the implementation are available, by which the complexity of the PRACH
transmitter and receiver can be minimized without sacrificing the performance.
UEs located near the edge of a cell, may disrupt UL transmissions in neighbouring cells. A frequency-
dependent „overload indicator‟ may be signalled directly between eNodeBs to warn a neighbouring
eNodeB of high uplink interference levels in specific RBs. In response, neighbouring eNodeB may
reduce the permitted energy per RB of the UEs scheduled in the corresponding RBs in its cell(s). It is
also possible to cooperate to avoid scheduling cell-edge UEs in neighbouring cells to transmit in the
same RB.
In summary, Basic operating point = P0 + α · PL
For low-rate PUCCH (with ACK/NACK and CQI), path-loss compensation is handled separately from
PUSCH, as PUCCH from different users are code-division-multiplexed. To provide good control of
interference between different users, and to maximize number of users simultaneously on PUCCH,
Dynamic Offset
The dynamic offset (Per RB) = MCS dependent component + TPC commands.
Uplink RBs allocated to a UE in a subframe may not be matched to the desired data rate and SIR.
Enable transmit power to be reduced if data transmitted is less than the rate supported by the radio
channel in a single RB. The MCS-dependent component for the PUSCH can be set to zero if fast
Adaptive Modulation and Coding (AMC) is used instead.
PUCCH bandwidth for a UE does not vary (although it may be ranging from a 1 bit SR/ACK/NACK, to
22 bits combined ACK/NACK and CQI together). Magnitude of power offset for each control can be
adjusted semi-statically by eNodeB to set a suitable error-rate operating point.
UE-specific power control commands. Other dynamic offset is UE-specific TPC commands-
accumulative (for PUSCH, PUCCH and SRS) and absolute (for PUSCH only). For PUSCH, switch
between these two modes is done by RRC for each UE.
1. Accumulative - each TPC signals a power step relative to previous level. This is default and
is suited when successive subframes received as closed loop. In LTE, two sets of power step
values are provided: either {−1, +1} dB or {−1, 0, +1, +3} dB configured by TPC command and
RRC configuration. Maximum power step size= +3/−1 dB, adjustable upto maximum and
minimum power limits according to the UE power class. A 0 dB step size means transmit
power to be kept constant if needed.
2. Absolute - TPC command is independent of sequence of TPC commands received
previously; depends only on the most recently-received absolute TPC command, a power
offset relative to the semi-static operating point. Absolute TPC commands offset set is {−4,
−1, +1, +4} dB.
The absolute power control mode can only control power within ±4 dB from operating point, but a
relatively power step can be triggered by a single command (up to ±8 dB), suited to scenarios of
intermittent transmission. Absolute TPC command adjusts to a suitable level in a single step.
Transmit power will be calculated at each TTI (subframe i) for each carrier „c‟ as follows:
Alpha(j) = {0, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 or 1}, 3bit factor for path loss
PL = DLPathLoss calculated by UE in dB = “Ref Sig Pwr – RSRP”.
(PathLossReferenceLinking)