You are on page 1of 125

Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme

bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Syst ` emes de t ´ el ´ ecommunication
C. Poulliat
21 novembre 2011
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Plan
1
Canaux de propagation et mod´ elisation
2
Syst ` eme multi-utilisateurs : acc` es multiple et duplexage
3
GSM : un exemple de syst ` eme bas´ e FDMA/TDMA
4
Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et
SC-FDMA
5
M´ ecanismes d’adaptation et de diversit ´ e
6
3GPP-LTE
7
UMTS : W-CDMA et HXPA
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Canaux de propagation et mod´ elisation
Effet
3.5 Modeling Broadband Fading Channels 99
precedent. The scalable nature of the WiMAX physical layer—notably, variable numbers of sub-
carriers and guard intervals—will allow custom optimization of the system for various environ-
ments and applications.
3.5.3 Empirical Channel Models
The parametric statistical channel models discussed thus far in the chapter do not take into
account specific wireless propagation environments. Although exactly modeling a wireless
channel requires complete knowledge of the surrounding scatterers, such as buildings and
plants, the time and computational demands of such a methodology are unrealistic, owing to the
near-infinite number of possible transmit/receive locations and the fact that objects are subject to
movement. Therefore, empirical and semiempirical wireless channel models have been devel-
oped to accurately estimate the pathloss, shadowing, and small-scale fast fading. Although these
models are generally not analytically tractable, they are very useful for simulations and to fairly
compare competing designs. Empirical models are based on extensive measurement of various
propagation environments, and they specify the parameters and methods for modeling the typi-
cal propagation scenarios in various wireless systems. Compared to parametric channel models,
the empirical channel models take into account such realistic factors as angle of arrival (AoA),
Figure 3.17 The dispersion/electivity duality: Dispersion in time causes frequency selectivity;
dispersion in frequency causes time selectivity.
Dispersive in
Time
Dispersive in
Frequency
Selective in
Frequency
Selective in
Time
signal rec¸u =
att ´ enuation
+ perte par effet de masquage bilan de liaison r ´ esultant
+ perte fading
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Canaux de propagation et mod´ elisation
Effet de la propagation en espace libre : att ´ enuation de puissance
Att ´ enuation par propagation en espace libre
P
r
= P
t
λ
2
G
t
G
r
(4πd)
2
, avec λ = c/f
c
Path Loss (dB) :P
L
= 10 log
10
(
P
r
P
t
)
o` u P
r
/P
t
repr ´ esente le gain moyen du canal
Mod` ele empirique d’att ´ enuation en espace libre
P
r
= P
t
P
0

d
0
d

α
Path Loss (dB) : P
L
= 10 log
10
(
P
r
P
t
)
o` u P
0
repr ´ esente la puissance ´ emise ` a la distance de r ´ ef ´ erence d
0
.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Canaux de propagation et mod´ elisation
Effet de masquage de l’environnement : shadowing
Att ´ enuation due au Shadowing : large scale fading
P
r
= P
t
P
0
χ

d
0
d

α
o` u χ = 10
x
10
avec x ∼ N(0, σ
2
x
)
Att ´ enuation totale (en dB) = Path loss + shadowing
P
db
= 10 log
10
(P
0
) −10αlog
10
(
d
d
0
) + χ
db
3.2 The Broadband Wireless Channel: Pathloss and Shadowing 75
Example 3.2
Consider a WiMAX base station communicating to a subscriber; the channel
parameters are , dB, and m, and dB. We
assume a transmit power of watt (30 dBm) and a bandwidth of
MHz. Owing to rate convolutional codes, a received SNR of 14.7
dB is required for 16 QAM, but just 3 dB is required for BPSK.
4
Finally, we
Figure 3.3 Shadowing causes large random fluctuations about the pathloss model: Figure from
[28], courtesy of IEEE.
(a) (b)
Figure 3.4 Shadowing causes large random fluctuations about the pathloss model. Figure from
[28], courtesy of IEEE.
4. These values are both 3 dB from the Shannon limit.
W
e
a
k
e
n
e
d
P
a
th
T–R Separation (km)
0.1
22 3 4
1 10
2 3 4
n = 1
n = 3
n = 2
140
130
120
110
100
70
80
90
n = 2.7
σ = 11.8 dB
n = 4
n = 5
All Measurement Locations
P
a
t
h
l
o
s
s

(
d
B
)
PA Bldg.
Stuttgart
Dusseldorf
Bank Bldg.
Kranberg
Hamburg
T–R Separation (km)
0.1
140
130
120
110
100
70
80
90
PA Bldg.
Stuttgart
Dusseldorf
Bank Bldg.
Kranberg
Hamburg
All Measurement Locations
22 3 4
1 10
2 3 4
n = 3.0
σ = 8.9 dB
n = 1
n = 3
n = 2
n = 4
n = 5
P
a
t
h
l
o
s
s

(
d
B
)
α = 3 P
o
= 40 − d
0
=1 σ
s
= 6
P
t
=1
B =10 1/2
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Canaux de propagation et mod´ elisation
Effet des trajets multiples
Mod` ele du signal rec¸u ´ equivalent en bande de base
y(t ) =
L(t )

n=0
α
n
(t )e
−j φ
n
(t )
x(t −τ
n
(t )) + b(t )
avec
φ
n
(t ) = 2πf
c
τ
n
(t ) −φ
D
n
φ
D
n
=

2πf
D
n
(t )dt , tel que f
D
n
(t ) = vf
c
cos (θ
n
(t ))/c
Canal ´ equivalent en bande de base
h(t , τ) =
L(t )

n=0
α
n
(t )e
−j φ
n
(t )
δ(τ −τ
n
(t ))
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Canaux de propagation et mod´ elisation
Effet des trajets multiples
P1: OTA/XYZ P2: ABC
c02 JWBK249/Myung August 27, 2008 19:33 Printer Name: Yet to Come
20 Single Carrier FDMA
Figure 2.3 Multipath propagation
2.2.1.5 Flat Fading and Frequency-Selective Fading
Signal scattering and multipath propagation together produce rapid fluctua-
tions in the strength of signals received at a base station as a cellular phone
moves through its service area. These fluctuations are due to differences in
received signal strength at locations spaced on the order of the wavelength
of the carrier frequency of the transmitted signal. This phenomenon is usu-
ally referred to as fast fading to distinguish it from shadow fading. The
differences in received signal strength associated with shadow fading are
noticeable at locations spaced in the order of tens of meters whereas fast
fading signals result from signal strength differences at locations spaced on
the order of a few centimeters.
The effect of fast fading on received cellular signals depends on the rela-
tionship of signal bandwidth to the width of the frequency response of the
channel. The frequency response is the Fourier transform of the impulse
response and its width is inversely proportional to the rms delay spread
of the multipath profile. When the signal bandwidth B
S
Hz is small com-
pared to the width of the frequency response, the fast fading is referred to
as “flat” because all the frequency components of the transmitted signal are
attenuated approximately equally. Otherwise the fast fading is “frequency-
selective”. Flat fading occurs when the product B
S
τ
rms
is small. Although
the nature of the fading changes gradually with changes in B
S
τ
rms
, it is
customary to refer to the fading as flat when B
S
τ
rms
< 0.02. At this value,
the correlation between two signal components at extreme ends of the
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Canaux de propagation et mod´ elisation
Caract ´ erisation du canal
3.5 Modeling Broadband Fading Channels 99
precedent. The scalable nature of the WiMAX physical layer—notably, variable numbers of sub-
carriers and guard intervals—will allow custom optimization of the system for various environ-
ments and applications.
3.5.3 Empirical Channel Models
The parametric statistical channel models discussed thus far in the chapter do not take into
account specific wireless propagation environments. Although exactly modeling a wireless
channel requires complete knowledge of the surrounding scatterers, such as buildings and
plants, the time and computational demands of such a methodology are unrealistic, owing to the
near-infinite number of possible transmit/receive locations and the fact that objects are subject to
movement. Therefore, empirical and semiempirical wireless channel models have been devel-
oped to accurately estimate the pathloss, shadowing, and small-scale fast fading. Although these
models are generally not analytically tractable, they are very useful for simulations and to fairly
compare competing designs. Empirical models are based on extensive measurement of various
propagation environments, and they specify the parameters and methods for modeling the typi-
cal propagation scenarios in various wireless systems. Compared to parametric channel models,
the empirical channel models take into account such realistic factors as angle of arrival (AoA),
Figure 3.17 The dispersion/electivity duality: Dispersion in time causes frequency selectivity;
dispersion in frequency causes time selectivity.
Dispersive in
Time
Dispersive in
Frequency
Selective in
Frequency
Selective in
Time
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Canaux de propagation et mod´ elisation
Mod´ elisation du comportement de h(t , τ)
Autocorr ´ elations
Autocorr ´ elation du canal temps/retard :
ρ
h
(t
1
, t
2
; τ
1
, τ
2
) = E(h(t
1
, τ
1
)h

(t
2
, τ
2
)).
Autocorr ´ elation du canal temps/fr ´ equence :
ρ
h
(t
1
, t
2
; f
1
, f
2
) = E(H(t
1
, f
1
)H

(t
2
, f
2
)). avec
H(t , f ) = TF
τ
(h(t , τ)(f )
WSSUS hypothesis
Autocorr ´ elation du canal temps/retard :
ρ
h
(t ; τ) = E(h(t , τ)h

(t +t , τ +τ))
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Canaux de propagation et mod´ elisation
Mod´ elisation du comportement de h(t , τ)
Etalement temporel
Delay spread T
d
≈ τ
max
Coherence Bandwith B
c
≈ 1/T
d
Etalement fr ´ equentiel
Doppler spread f
D
≈ vf
c
/c
Coherence time T
c
≈ 1/f
d
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Canaux de propagation et mod´ elisation
Classification
124 CHAPTER 2. MOBILE RADIO CHANNELS
mitted signal duration (the symbol period for digital transmissions) plus
the channel delay spread, a(t) (defined in Equation 2.95). As the channel
response appears to be constant for the duration of the signal's flight, the
channel is referred to as time-flat.
When a channel is flat in both frequency and time, it is called a flat-
flat channel. When a channel is flat neither in frequency nor in time, it
is often referred to as a doubly dispersive channel. This nomenclature,
however, is somewhat misleading because such a channel need only cause
signal distortion not dispersion. Hence we shall refer to a channel that is
neither time-flat nor frequency-flat as a non-flat channel. Figure 2.17 shows
the classification of channels following the above approach. The shaded
region of the figure indicates the physical restriction that it is impossible
for the time bandwidth product of a signal to be less than 1/2 [21]. A
more rigorous system of classification, emphasising the differences between
distorting and dispersive channels, is shown in Figure 2.18.
B
x
Time-Flat Non-Flat
Frequency-Flat
Flat-Flat
Be : .
Time Duration
Figure 2.17: Channel classifications.
The flat-flat, or doubly-flat channel does not fade with either time or
frequency. Using the approximate values derived above, it is seen that a
signal of bandwidth less than Be(t) Hz and duration less than Te(t) seconds
will observe a flat-flat channel at time t. For example, in this category re-
2.4. CLASSIFICATION OF CHANNELS 125
;'
~ Frequency Dispersive Frequency Disperaive . Frequency Diapenive
./ Time Flat Time Diatortlng Time Diapeuive
./
./
1
4T
tn
Frequency Distorting . Frequency Distorting . Frequency Diatorting
Time Flat Time Distorting Time Diapenive
Signal Duration
Frequency Flat
Time Diaperaive
1
4/m
Frequency Flat
Time Distorting
Be .
Figure 2.18: Detailed channel classifications.
sides the uplink from the mobile stations (MS) to the base station (BS) of
the MATS-D system [22]. MATS-D is a hybrid mobile radio system that
was put forward as a contender for the pan-European cellular radio sys-
tem. It employs a narrowband frequency division multiple access (FDMA)
scheme for the uplink. The transmissions employ generalised tamed fre-
quency modulation (GTFM) to give a bandwidth of approximately 25 kHz
for a bit rate of 19.5 kb/s,
The frequency-flat fading channel is observed by narrowband channel
sounders. These sounders transmit a monochromatic (single tone) signal
continuously, and so approximate to a signal of infinitesimal bandwidth
and infinite duration. The envelope of the received process therefore varies
in sympathy with the channel.
For a delay spread of 2J-ts, the time-flat fading channel applies to all
mobile radio systems using digital transmissions with bit rates in excess
of 80 kb /s (see page 121). Many of the systems put forward for the
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
M´ ethodes d’acc` es multiples
G´ en´ eralit ´ es 1/3
Motivations
Partager le canal radio (spectre et temps) entre diff ´ erents
utlisateurs,
Permettre l’acc` es pour que les utilisateurs communiquent
simultan´ ement
⇒cr ´ eation d’”un canal” (“channelization”), portion de temps
et/ou de fr ´ equence allou´ ee temporairement ` a un utilisateur pour
communiquer et transmettre ses donn´ ees
Servir le plus grand nombre d’utilisateurs (rentabilit ´ e) tout en
garantissant une ceratine qualit ´ e de service (QoS), ie. taux
d’erreur, d´ ebits, connectivit ´ es, d´ elai, etc....
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
M´ ethodes d’acc` es multiples
G´ en´ eralit ´ es 2/3
Principales m´ ethodes d’acc´ es
M´ ethodes Acc` es Multiples centralis´ ees :
Frequency Division Multiple Access, FDMA,
Time Division Multiple Access, TDMA,
Code Division Multiple Access, CDMA,
Space Division Multiple Access, SDMA,
Orthogonale Frequency Division Multiple Access, OFDMA,
Single-Carrier Frequency Division Multiple Access, SC-FDMA
M´ ethodes Acc` es Distribu´ ees et/ou Al ´ eatoires (Packet
Radio/Random Access) :
m´ ethodes bas´ ees sur ALOHA,
Packet Reservation Multiple Acces, PRMA,
Carrier Sense Multiple Access, CSMA
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
M´ ethodes d’acc` es multiples
G´ en´ eralit ´ es 3/3
D´ efinitions et acronymes courants
Emetteur-R´ ecepteur :
Station de base (Base Station, access point, BS) li ´ ee au r ´ eseau de
l’op´ erateur (points d’entr ´ ee),
Terminal Mobile (Mobile terminal/Mobile Station, MS)
Voie descendante, VD (downlink/forward link) :
communication de BS vers MS
Voie montante, VM (uplink/reverse link) :
communication de MS vers BS
g´ en´ eralement d´ ebits sur voie VD bien sup´ erieure ` a VM.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Duplexage
Duplexage
le duplexage permet de s´ eparer les communications voies montante
et descendante
Duplexage total ou partiel
Syst ` eme de communication bidirectionnel (resp.
mono-directionnel) :
syst ` eme pouvant ´ emettre et recevoir (resp. ´ emettre ou recevoir),
Duplexage total (Full-Duplex) : le terminal peut ´ emettre et
recevoir des donn´ ees de mani ` ere simultan´ ee,,
Semi-Duplexage (Half-Duplex) : le terminal peut ´ emettre et
recevoir des donn´ ees de mani ` ere non simultan´ ee (g´ en´ eralement
s´ equentielle),
Mode Simplex : pour les syst ` emes mono-directionnels.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Duplexage en temps
Duplexage en temps (Time Division Duplex, TDD)
s´ eparation des voies Down/Up en temps en assignant des slots
temporels (time-slots, TS) diff ´ erents pour la communication
(mˆ eme bande de fr ´ equence),
temps/p´ eriode de garde requis pour minimiser interf ´ erence entre
´ emission et r ´ eception, fonction du temps max. pour
communication “aller-retour”,
estimation de canal simplifi´ ee par “r ´ eciprocit ´ e” (channel
reciprocity assumption) si temps de duplage entre les TS inf.
temps de coh´ erence du canal,
⇒possible sur certains syst ` emes et certaines conditions de
mobilit ´ es (canaux dˆıts quasi-statiques)
non full-duplex au sens strict.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Duplexage en temps
Principe g´ en´ erale du duplexage TDD
312 MULTIPLE-ACCESS TECHNIQUES
Time
D
1
U
1
D
2
U
3
D
3
D
4
U
4
U
2
Guard time (50 to 200 µs)
Figure 8.15 Time division duplexing (TDD) with downlink transmission intervals D
n
and uplink
transmission intervals U
n
with guard time. Transmission interval length can vary
with adaptive TDD.
increased efficiency is given at about 16% in a recent report. Because of basic overhead
information, all TDD systems (like UTRA TDD, for example) use a fixed frame length and
TS or ‘burst’ structure of some sort that becomes the quantum unit for uplink/downlink
transmission. The flexibility to adjust the symmetry of the uplink/downlink flow is there-
fore not unlimited but restricted to possible ratios as discrete values. The limited number
of discrete values still provide for considerable range in adjusting the uplink/downlink
time slot allocations.
For multiple access or adaptive antenna techniques that use channel estimation, the fact
that the uplink and the downlink channels are the same can be used to some advantage.
TDD systems are also more flexible in administrative frequency allocations in which only
a single rather than a pair of frequency channels are available. Of course, for equivalent
full rate downlink and uplink transmission, the bandwidth of the TDD channel must be
equal to the bandwidth of the sum of the downlink and uplink channel bandwidths used
for FDD.
8.8.2.1 TDD interference calculations
Although TDD has many benefits, one drawback of a TDD system that is rarely pointed
out is the doubling of the number of interference sources that must be dealt with as
compared to an FDD system. This situation is illustrated in Figure 8.16(a) for FDD with
a simple two-hub system with one remote terminal connected to each hub using the
same downlink and uplink frequencies f
d
and f
u
in each case. The interference paths
for the FDD interference analysis are shown by dotted lines for the uplink and dashed
lines for the downlink. The interference analysis for the downlink (interference to remote
terminals) is confined to interference from other system hubs only. Similarly, uplink
interference calculations (interference to hubs) are confined to interference from other
remote terminals only.
The TDD interference situation is shown in Figure 8.16(b). Because both the hub
and remote terminals are transmitting on the channel f
d
, both must be considered in
the interference calculations at the hub and remote terminal (uplink and downlink) on
Cas de trafic asym´ etrique
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Duplexage en fr ´ equence
Duplexage en fr ´ equence (Frequency Division Duplex, TDD) :
s´ eparation des voies Down/Up en fr ´ equence en assignant des
bandes de fr ´ equence (frequency bands) de communication
pendant la dur ´ ee d’une communications,
Bande de garde requis pour r ´ eduire les interf ´ erences entre voies
Down/Up.,
si s´ eparation entre deux bandes sup. ` a la bande de coh´ erence
du canal, gains fr ´ equentiels diff ´ erents,
⇒toujours le cas en pratique, donc “r ´ eciprocit ´ e” du canal non
valide
full-duplex au sens strict,
perte d’efficacit ´ e pour syst ` eme fortement asym´ etrique en trafic,
duplexage le plus utilis´ e.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Duplexage en fr ´ equence
Principe g´ en´ erale du duplexage FDD
310 MULTIPLE-ACCESS TECHNIQUES
networks. The duplexing method chosen for a system is essentially independent of the
modulation type and the multiple-access-technology. The main issues that influence the
duplexing method selection are the nature of the available spectrum allocations and
the nature of the traffic that is to be communicated.
The two duplexing methods primarily used in fixed broadband wireless are FDD and
TDD. Both are discussed in the following sections along with the lesser-used methods of
polarization duplexing and space duplexing.
8.8.1 Frequency division duplexing (FDD)
Frequency division duplexing is the method most widely used in traditional fixed broad-
band point-to-point and point –to-multipoint systems. A drawing illustrating FDD is shown
in Figure 8.14. The primary parameters with FDD are the channel bandwidths themselves
for downlink and uplink and the duplex frequency separation between pairs of downlink
and uplink frequencies. The downlink and uplink bandwidths are chosen to accommodate
the required system data rate and modulation scheme, or they may also be mandated by
administrative rules as in the case of first- and second-generation cellular systems. The
total duplex bandwidth available to a user is D
n
+U
n
. If asymmetrical downlink and
uplink traffic volume is anticipated, D
n
does not necessarily have to be the same as U
n
,
although in conventional systems it is typical for the downlink and the uplink bandwidths
to be equal.
The duplex frequency separation is chosen largely to make the design of duplexing
waveguides or circuitry at a terminal practical. For a full duplex system, a terminal will
be sending and receiving simultaneously so that both the downlink and the uplink chan-
nels are active. If a single antenna is used, which is the economical approach, the device
known as a duplexer at the remote terminal must separate the signals being received on D
n
from those being transmitted on U
n
. The design of duplex filters to accomplish this task
becomes more difficult and problematic as f decreases. Since the uplink and the down-
link signals are different, signal leakage of one into the other represents interference and
Frequency
D
1
D
2
D
3
D
N
Duplex frequency separation ∆f
U
2
U
3
U
N
U
1
Guard band
Figure 8.14 Frequency division duplexing (FDD) with downlink channels D
n
and uplink channels
U
n
with duplex frequency separation f .
Cas de plusieurs canaux duplex´ es en fr ´ equence
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiple Par R´ epartition en Fr ´ equence
FDMA
Principe g´ en´ erale
chaque utilisateur se voit assigner une fr ´ equence ou bande de
fr ´ equences (g´ en´ eralement contigues) pour la durr ´ ee de sa
communication,
FREQUENCY DIVISION MULTIPLE ACCESS (FDMA) 279
Frequency
Time
Code
Radio channel N
Radio channel 3
Radio channel 2
Radio channel 1
Figure 8.1 Frequency division multiple access (FDMA).
Frequency
1
Channel
2
Channel Channel
3
Channel
N
Channel center
frequency
Channel
bandwidth
Power spectral
density (PSD)
Figure 8.2 FDMA channelization of allocated spectrum with different occupied bandwidths.
flexibility, adequate bandwidth to support the data rates required by the terminal. An
effective channel assignment process is fundamental for achieving efficient use of the
spectrum. Accordingly, frequency assignment and planning will be discussed in some
detail in Chapter 12 for particular network configurations.
Essentially all wireless systems use FDMA in some form, even systems that are
normally referred to as time division multiple access (TDMA) systems. For example,
the IS-136 TDMA cellular system used in the United States uses the same FDMA chan-
nelization scheme as the first generation advanced mobile phone service (AMPS) cellular
Principe g´ en´ erale du FDMA
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiple Par R´ epartition en Fr ´ equence
FDMA : duplexage associ ´ e
Time
Frequency
A
m
p
l
i
t
u
d
e
Downlink Uplink
f
1
f
2
f
1
*
f
2
*
f
3
f
4
f
3
*
f
4
*
Time
Frequency
A
m
p
l
i
t
u
d
e
Downlink Uplink
f
1
f
2
f
1
*
f
2
*
(a)
(c)
t
1
t
2
t
3
t
4
t
1
t
2
t
1
t
2
t
1
t
2
t
3
t
4
t
3
t
4
t
3
t
4
Time
Frequency
A
m
p
l
i
t
u
d
e
f
1
f
2
f
3
f
4
(b)
(d)
t
1
t
1
t
1
*
t
1
*
Uplink
Downlink
user1 user2 user3 user4
Time
Frequency
a
m
p
l
i
t
u
d
e
f
1
f
2
t
1
t
1
*
t
2
*
t
3
*
t
4
*
t
2
t
3
t
4
t
1
t1
*
t
2
*
t
3
*
t
4
*
t
2
t
3
t
4
t
1
t
1
t
1
*
t
1
*
t
1
t
1
t
1
*
t
1
*
t
1
t
1
t
1
*
t
1
*
FIGURE 5.1 (a) FDMA/FDD; (b) FDMA/TDD; (c) TDMA/FDD with multiple carriers;
(d) TDMA/TDD with multiple carriers.
(a)
… …
Downlink Uplink
824
849 869 894 MHz
30 kHz
(b)
Time
frequency


One user one-way frequency channel
1 ms
Channel 1
Channel 40
Spectrum
Uplink Downlink
FIGURE 5.2 (a) FDMA/FDD in AMPS and (b) FDMA/TDD in CT-2.
MEDIUM ACCESS METHODS 169
Time
Frequency
A
m
p
l
i
t
u
d
e
Downlink Uplink
f
1
f
2
f
1
*
f
2
*
f
3
f
4
f
3
*
f
4
*
Time
Frequency
A
m
p
l
i
t
u
d
e
Downlink Uplink
f
1
f
2
f
1
*
f
2
*
(a)
(c)
t
1
t
2
t
3
t
4
t
1
t
2
t
1
t
2
t
1
t
2
t
3
t
4
t
3
t
4
t
3
t
4
Time
Frequency
A
m
p
l
i
t
u
d
e
f
1
f
2
f
3
f
4
(b)
(d)
t
1
t
1
t
1
*
t
1
*
Uplink
Downlink
user1 user2 user3 user4
Time
Frequency
a
m
p
l
i
t
u
d
e
f
1
f
2
t
1
t
1
*
t
2
*
t
3
*
t
4
*
t
2
t
3
t
4
t
1
t1
*
t
2
*
t
3
*
t
4
*
t
2
t
3
t
4
t
1
t
1
t
1
*
t
1
*
t
1
t
1
t
1
*
t
1
*
t
1
t
1
t
1
*
t
1
*
FIGURE 5.1 (a) FDMA/FDD; (b) FDMA/TDD; (c) TDMA/FDD with multiple carriers;
(d) TDMA/TDD with multiple carriers.
(a)
… …
Downlink Uplink
824
849 869 894 MHz
30 kHz
(b)
Time
frequency


One user one-way frequency channel
1 ms
Channel 1
Channel 40
Spectrum
Uplink Downlink
FIGURE 5.2 (a) FDMA/FDD in AMPS and (b) FDMA/TDD in CT-2.
MEDIUM ACCESS METHODS 169
FDMA-FDD FDMA-TDD
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiple Par R´ epartition en Fr ´ equence
FDMA : avantages-inconv´ enients
Avantages
G´ en´ eralement, moins de traitement de signal requis
(communications bandes ´ etroites),
synchronisation temporelle facilit ´ e
D´ esavantages
sensibilit ´ e au fading fr ´ equentiel (pas de diversit ´ e fr ´ equentielle),
interf ´ erence des canaux adjacents (en part. VM), produits
d’intermodulation (BS),
n´ ecessit ´ e ´ eventuelle d’intervalle fr ´ equentiel de garde
Type de syst ` emes utlisant le FDMA
syst ` emes de communications analogiques,
utilis´ e souvent en combinaison avec d’autres m´ ethodes d’acc` es
(ex : GSM),
syst ` emes ` a forts d´ ebits
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiple Par R´ epartition en Temps
TDMA : Principe g´ en´ erale
sur une une mˆ eme bande de fr ´ equence, N utilisateurs
communiquent sur un intervalle de temps fixe (Time-Slot, TS),
sur chaque TS, un utilisateur unique peut communiquer en
utilisant la bande allou´ ee,
le processus est r ´ ep´ et ´ e dans le temps de mani ` ere p´ eriodique.
Ex du GSM : N = 8 utilisateurs sur 200 kHz de bande.
TIME DIVISION MULTIPLE ACCESS (TDMA) 287
Frequency
Time
Code
Radio Channel N
Radio Channel 3
Radio Channel 2
TS 1
TS 1
Radio Channel 1 TS 1
TS 2
TS 2
TS 1 TS 2 TS 1 TS 2 TS K
TS K TS K
TS K
TS 1
TS 1
TS 2
TS 2
TS 1 TS 2 TS 2 TS K TS K
TS K
TS K
Figure 8.6 Time division multiple access (TDMA) with schematic of time slots (TS).
Time Slot N Time Slot 4 Time Slot 3 Time Slot 2 Time Slot 1
Trailing bits Information bits Frame preamble
TDMA frame
Guard
bits
Information bits
Synchronization &
training bits
Guard
bits
Figure 8.7 Generic TDMA frame and time slot structure.
The TS shown schematically in Figure 8.6 are actually grouped together in frames to
ensure that they are coordinated with one another. Figure 8.7 shows a generic TDMA
frame with preamble and trailing bits. The information frame is divided into TS that are
self-contained bursts of data including training and synchronization bits. The training
bit sequence is known to the terminal and used to train the equalizer as discussed in
Section 7.4.4. Depending on the modulation type, these bits may also be used to establish
a synchronous reference to facilitate data symbol detection.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiples Par R´ epartition en Temps
TDMA : duplexage associ ´ e
Time
Frequency
A
m
p
l
i
t
u
d
e
Downlink Uplink
f
1
f
2
f
1
*
f
2
*
f
3
f
4
f
3
*
f
4
*
Time
Frequency
A
m
p
l
i
t
u
d
e
Downlink Uplink
f
1
f
2
f
1
*
f
2
*
(a)
(c)
t
1
t
2
t
3
t
4
t
1
t
2
t
1
t
2
t
1
t
2
t
3
t
4
t
3
t
4
t
3
t
4
Time
Frequency
A
m
p
l
i
t
u
d
e
f
1
f
2
f
3
f
4
(b)
(d)
t
1
t
1
t
1
*
t
1
*
Uplink
Downlink
user1 user2 user3 user4
Time
Frequency
a
m
p
l
i
t
u
d
e
f
1
f
2
t
1
t
1
*
t
2
*
t
3
*
t
4
*
t
2
t
3
t
4
t
1
t1
*
t
2
*
t
3
*
t
4
*
t
2
t
3
t
4
t
1
t
1
t
1
*
t
1
*
t
1
t
1
t
1
*
t
1
*
t
1
t
1
t
1
*
t
1
*
FIGURE 5.1 (a) FDMA/FDD; (b) FDMA/TDD; (c) TDMA/FDD with multiple carriers;
(d) TDMA/TDD with multiple carriers.
(a)
… …
Downlink Uplink
824
849 869 894 MHz
30 kHz
(b)
Time
frequency


One user one-way frequency channel
1 ms
Channel 1
Channel 40
Spectrum
Uplink Downlink
FIGURE 5.2 (a) FDMA/FDD in AMPS and (b) FDMA/TDD in CT-2.
MEDIUM ACCESS METHODS 169
Time
Frequency
A
m
p
l
i
t
u
d
e
Downlink Uplink
f
1
f
2
f
1
*
f
2
*
f
3
f
4
f
3
*
f
4
*
Time
Frequency
A
m
p
l
i
t
u
d
e
Downlink Uplink
f
1
f
2
f
1
*
f
2
*
(a)
(c)
t
1
t
2
t
3
t
4
t
1
t
2
t
1
t
2
t
1
t
2
t
3
t
4
t
3
t
4
t
3
t
4
Time
Frequency
A
m
p
l
i
t
u
d
e
f
1
f
2
f
3
f
4
(b)
(d)
t
1
t
1
t
1
*
t
1
*
Uplink
Downlink
user1 user2 user3 user4
Time
Frequency
a
m
p
l
i
t
u
d
e
f
1
f
2
t
1
t
1
*
t
2
*
t
3
*
t
4
*
t
2
t
3
t
4
t
1
t1
*
t
2
*
t
3
*
t
4
*
t
2
t
3
t
4
t
1
t
1
t
1
*
t
1
*
t
1
t
1
t
1
*
t
1
*
t
1
t
1
t
1
*
t
1
*
FIGURE 5.1 (a) FDMA/FDD; (b) FDMA/TDD; (c) TDMA/FDD with multiple carriers;
(d) TDMA/TDD with multiple carriers.
(a)
… …
Downlink Uplink
824
849 869 894 MHz
30 kHz
(b)
Time
frequency


One user one-way frequency channel
1 ms
Channel 1
Channel 40
Spectrum
Uplink Downlink
FIGURE 5.2 (a) FDMA/FDD in AMPS and (b) FDMA/TDD in CT-2.
MEDIUM ACCESS METHODS 169
TDMA-FDD TDMA-TDD
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiples Par R´ epartition en Temps
TDMA : avantages-inconv´ enients
Avantages
diversity fr ´ equentielle possible,
pas d’intervalle fr ´ equentiel de garde,
gestion de diff ´ erents services par assignation de plusieurs TS,
p´ eriodes d’ ´ ecoute pendant la phase de silence (handover
facilit ´ e).
D´ esavantages
interf ´ erence entre symbole : ´ egalisation n´ ecessaire,
intervalle temporel de garde n´ ecessaire (temps de mont ´ ee et
descente pour puissance d’ ´ emission+compensation des temps
de trajets h´ et ´ erog` enes entre terminaux) pour synchronisation
temporelle des utlisateurs (surtout uplink),
estimation de canal pour chaque TS possiblement requises,
overhead induit par ces traitements
Type de syst ` emes utlisant le TDMA
GSM/EDGE/GPRS, DECT,
IS-54/136 (US)
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiples Par R´ epartition en Temps
TDMA : trame g´ en´ erique
Exemple g´ en´ erique de trame TDMA
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiples : Sch´ emas Hybrides FDMA/TDMA
Exemple du GSM : FDMA/TDMA/FDD
Example 5.10: TDMA in GSM Figure 5.3 shows an FDMA/TDMA/FDDchannel used in
2G digital cellular in Europe (GSM). The particular example shows the eight-slot TDMA
scheme used in the GSM system. Forward and reverse channels use separate carrier
frequencies (FDD). Each carrier can support up to eight simultaneous users via TDMA,
eachusing13 kb/s encodeddigital speech, withina200 KHzcarrier bandwidth. Atotal of 124
frequency carriers (FDMA) are available in the 25 MHz allocated band in each direction.
100 kHz of band is allocated as a guard band at each edge of the overall allocated band.
Example 5.11: TDMAinDECT Figure 5.4 shows an FDMA/TDMA/TDDsystemused in
the pan-European digital PCS standard DECT. Since distances are short, a TDD format
allows use of the same frequency for forward and reverse operations. The bandwidth per
.
.
.
10 MHz
11 9 8 … 1 0
1.728 MHz
Time slots allocated to users on one carrier
Carrier 1
Carrier 2
Carrier 5
10 12 21 … 13 22 23 20
Downlink 5ms Uplink 5ms
32 bit
preamble
388 bit
data
60 bit
guard
64 control
bits
320 bit
channel
4 parity
bits
FIGURE 5.4 FDMA/TDMA/TDD in DECT.
.
.
.
25 MHz
100 kHz
guard band
8 7 6 5 4 3 2 1
200 kHz
Time slots allocated to
users on one carrier
Carrier 1
Carrier 2
Carrier 124
Uplink channels
Downlink channels
FIGURE 5.3 FDMA/TDMA/FDD in GSM.
172 NETWORKING FUNDAMENTALS
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiples : Sch´ emas Hybrides FDMA/TDMA
Exemple du DECT : FDMA/TDMA/TDD
Example 5.10: TDMA in GSM Figure 5.3 shows an FDMA/TDMA/FDDchannel used in
2G digital cellular in Europe (GSM). The particular example shows the eight-slot TDMA
scheme used in the GSM system. Forward and reverse channels use separate carrier
frequencies (FDD). Each carrier can support up to eight simultaneous users via TDMA,
eachusing13 kb/s encodeddigital speech, withina200 KHzcarrier bandwidth. Atotal of 124
frequency carriers (FDMA) are available in the 25 MHz allocated band in each direction.
100 kHz of band is allocated as a guard band at each edge of the overall allocated band.
Example 5.11: TDMAinDECT Figure 5.4 shows an FDMA/TDMA/TDDsystemused in
the pan-European digital PCS standard DECT. Since distances are short, a TDD format
allows use of the same frequency for forward and reverse operations. The bandwidth per
.
.
.
10 MHz
11 9 8 … 1 0
1.728 MHz
Time slots allocated to users on one carrier
Carrier 1
Carrier 2
Carrier 5
10 12 21 … 13 22 23 20
Downlink 5ms Uplink 5ms
32 bit
preamble
388 bit
data
60 bit
guard
64 control
bits
320 bit
channel
4 parity
bits
FIGURE 5.4 FDMA/TDMA/TDD in DECT.
.
.
.
25 MHz
100 kHz
guard band
8 7 6 5 4 3 2 1
200 kHz
Time slots allocated to
users on one carrier
Carrier 1
Carrier 2
Carrier 124
Uplink channels
Downlink channels
FIGURE 5.3 FDMA/TDMA/FDD in GSM.
172 NETWORKING FUNDAMENTALS
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiples : Sch´ emas Hybrides FDMA/TDMA
Exemple du standard IS-136 (US) : FDMA/TDMA/TDD
carrier is 1.728 MHz, which can support up 12 ADPCM-coded speech channels via TDMA.
The total allocated band in Europe is 10 MHz, which can support five carriers (FDMA).
Figure 5.4 shows the details of the TDMA/TDD time slots use in the DECT system. The
frame duration is 10 ms, with 5 ms for portable-to-fixed station and 5 ms for fixed-to-
portable. The transmitter transfers information in signal bursts which it transmits in slots of
duration 10/24 ¼0.417 ms. With 480 bits per slot (including a 60-bit guard time), the total
bit rate is 1.152 Mb/s. Each slot contains 64 bits for systemcontrol (C, P, Qand Mchannels)
and 320 bits for user information (I channel).
Example 5.12: TDMA in IS-136 Figure 5.5 shows the frame format for the TDMA/FDD
with six slots considered for IS-136 both for the forward (base to mobile) and reverse
(mobile to base) channels. In IS-136, each 30 kHz digital channel has a channel transmission
rate of 48.6 kb/s. The 48.6 kb/s stream is divided into six TDMA channels of 8.1 kb/s each.
The IS-136 slot and frame format, shown in Fig. 5.5, is much simpler than that of the GSM
standard. The 40 ms frame is composed of six 6.67 ms time slots. Each slot contains 324 bits,
including 260 bits of user data, and 12 bits of systemcontrol information in a slowassociated
control channel (SACCH). There is also a 28-bit synchronization sequence and a 12-bit
digital verification color code (DVCC) used to identify the frequency channel to which the
mobile terminal is tuned. In the mobile-to-base direction, the slot also contains a guard time
interval of 6-bit duration, when no signal is transmitted, and a 6-bit ramp interval to allowthe
transmitter to reach its full output power level.
Owing to the near--far problem, the received signal on the reverse channel from a user
occupying a time slot can be muchlarger than the receivedpower fromthe terminal using the
adjacent time slot. In such a case, the receiver will have difficulty in distinguishing the
weaker signal from the background noise. In a manner similar to FDMA systems, TDMA
systems also use power control to handle this near--far problem.
Capacity of Time-/Frequency-Division Multiple Access Cellular Systems. The capacity
of a cellular system in the case of TDMA or FDMA depends primarily on the number of
.
.
.
25 MHz
6 5 4 3 2 1
30 kHz
Time slots allocated to
users on one carrier
Carrier 1 (used as a guard band)
Carrier 2
Carrier 833
Uplink channels
Downlink channels
FIGURE 5.5 FDMA/TDMA/FDD in IS-136 standard.
MEDIUM ACCESS METHODS 173
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiples Par R´ epartition de Codes
Principe g´ en´ erale
On attribue ` a chaque utilisateur un code dˆıt d’ ´ etalement pour
moduler son signal d’information,
Les utilisateurs occupent la mˆ eme bande au mˆ eme instant,
Etalement par S´ equences Directes (Direct Sequence) ou par
saut de fr ´ equences (Frequency Hopping)
M´ ethode d’acc` es orthogonale ou non-orthogonale
292 MULTIPLE-ACCESS TECHNIQUES
Frequency
Time
Code
Code1
Code 2
Code 1
Code 2
Radio channel 1
Radio channel 2
Radio channel 3
Time slot (for TD-CDMA)
Code K
Radio channel N
Code K
Figure 8.9 Code division multiple access (CDMA) on multiple radio channels.
W, resulting in a transmitted signal symbol rate at the spreading code rate and an occupied
bandwidth essentially the same as the spreading code bandwidth (provided W R). The
transmitted symbols are known as chips with the rate of the spreading code called the
chipping rat e. The ratio of the chipping rate to the data rate, W/R, is called the spreading
factor (also called processing gain). The spreading factor is one fundamental indication of
the degree of processing gain in the system, and consequently, the degree of interference
rejection (multiuser access) that can be achieved.
As stated in the introduction to this chapter, the core issue with all multiple-access
schemes is their ability to operate in the presence of interference. The following sections
will present standard interference calculations for CDMA systems. As will be shown,
CDMA systems generally rely on careful control of the interference levels from all termi-
nals communicating within a given hub on a radio channel; specifically, best (unbiased)
detection of the uplink signals from all terminals connected to a hub is realized when all
are received with equal power. This means that CDMA systems must have the ability to
control the transmitting power of the remote terminals so that the signals received from
terminals close to the hub do not dominate the signals received from more distant ter-
minals. This is the so-called near–far problem in CDMA systems. It is dealt with using
effective APC that may need a dynamic range of 80 dB or more to equalize the effects
of wide path loss differences that can occur with near and far terminals connected to the
same hub. The use of advanced techniques such as joint detection (JD) (see Section 8.4.5)
can relax the demands on the uplink APC adjustments.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiples Par R´ epartition de Codes
Propri ´ et ´ es
Utlisation du code d’ ´ etalement pour s´ eparer les utilisateurs en
r ´ eception (d´ etection multi-utilisateurs possible en uplink),
Utilisation de codes orthogonaux en liaison descendante (type
Walsh-Hadamard),
Utilisation de codes non-orthogonaux en liaison montante :
gestion des trajets multiples, peu de coordination car s´ eparation
par codes,
Contr ˆ ole de puissance requis pour effet near-far,
Utilisation de multi-codes pour accroitre les d´ ebits d’un
utilisateur.
Standards
IS-95, CDMA-2000,
UMTS-WCDMA/UMTS-HSXPA
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiples Par R´ epartition de Codes
transmitter and receiver use the same carrier frequency (Fig. 5.7b), the system is CDMA/
TDD.
In CDMA, each user is a source of noise to the receiver of other users, and if we increase
the number of users beyond a certain value, then the entire system collapses because the
signal received in each specific receiver will be buried under the noise caused by many other
users. An important question is how many users can simultaneously use a CDMA system
before the system collapses. We investigate the answer to this below.
Capacity of Code-Division Multiple Access. CDMA systems are implemented based on
spread-spectrum technology that was presented in Chapter 3. In its most simplified form, a
spread-spectrum transmitter spreads the signal power over a spectrum N times wider than
the spectrum of the message signal. In other words, an information bandwidth of R
b
occupies a transmission bandwidth of W, where
W ¼ N
p
R
b
ð5:1Þ
The spread-spectrumreceiver processes the received signal with a processing gain of N
p
.
This means that, during the processing at the receiver, the power of the received signal
having the code of that particular receiver will be increased Ntimes beyond the value before
processing.
Let us consider the situation of a single cell in a cellular system employing CDMA.
Assume that we have M simultaneous users on the reverse channel of a CDMA network.
Further, let us assume that we have an ideal power control enforced on the channel so that the
received power of signals fromall terminals has the same value P. Then, the received power
fromthe target user after processing at the receiver is NP and the received interference from
M À 1 other terminals is (M À 1)P. If we also assume that a cellular system is interference
limited and the background noise is dominated by the interference noise from other users,
then the received SNR for the target receiver will be
S
r
¼
N
p
P
ðMÀ1ÞP
¼
N
p
MÀ1
ð5:2Þ
Code
Frequency
Time
User 1
User 2
User 3
User 4
User 5
User 6
User 7
CDMA/TDD
Uplink
Downlink
Downlink
Uplink
CDMA/FDD
(a)
(b)
FIGURE 5.7 (a) FDD and (b) TDD with CDMA.
MEDIUM ACCESS METHODS 175
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es Multiples Par R´ epartition Spatiale
Principe g´ en´ erale
m´ ethode d’acc` es pour am´ eliorer l’efficacit ´ e des m´ ethodes
FDMA/TDMACDMA en utilisant la r ´ epartition spatiale
utilisation d’antennes multiples directionnelles pour cr ´ eer des
canaux orthogonaux
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
GSM : Pr ´ esentation g´ en´ erale
standard ouvert,
3 versions de GSM :
GSM-900,
GSM-1800 (DCS-1800),
GSM-1900(PCS1900), surtrout USA.
Evolutions :
EDGE,
GPRS
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Interface Air
Duplexage FDD
GSM-900 :
Uplink : 890 to 915 Mhz,
Downlink : 935 to 960 Mhz,
s´ eparation fr ´ equentielle pour duplexage : 45Mhz,
GSM-1800 :
Uplink : 1710 to 1785 Mhz,
Downlink : 1805 to 1880 Mhz,
s´ eparation fr ´ equentielle pour duplexage : 9 :ml5Mhz,
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Interface Air
Acc` es multiple primaire : FDMA
Partitionnement des bandes allou´ ees pour uplink et downlink en
124 sous-bandes de 200 kHz,
Bandes de garde fr ´ equencielles de 100 kHz au deux bords de
chaque sous-bande,
les sous bandes de 200 kHz sont num´ erot ´ es de mani ` ere
cons´ ecutive en utlisant la num´ erotation Absolute Radio
Frequency Channel Numbers, ARFCNs.
GSM – Global System for Mobile Communications 591
24.3 The Air Interface
GSM employs a combined FDMA/TDMA approach which further combines with Frequency
Domain Duplexing (FDD) (see Chapter 17). Let us elaborate on these acronyms.
FDD
In the first GSM version, frequencies from 890 to 915 MHz and from 935 to 960 MHz were
available. The lower band is used for the uplink (connection from the MS to the BS). The upper
band is used for the downlink. The frequency spacing between the uplink and downlink for any
given connection is 45 MHz. Therefore, relatively cheap duplex filters are sufficient for achieving
very good separation between the uplink and downlink.
For GSM1800, the frequency ranges are 1,710–1,785 MHz for the uplink, and 1,805–1,880 MHz
for the downlink. In North America, 1,850–1,910 MHz are used for the uplink and 1,930–
1,990 MHz for the downlink. Other bands are added as they become available, see also Chapter 27.
FDMA
Both uplink and downlink frequency bands are partitioned into a 200-kHz grid. The outer 100 kHz
of each 25-MHz band are not used,
4
as they are guard bands to limit interference in the adjoined
spectrum, which is used by other systems. The remaining 124 200-kHz subbands are numbered
consecutively by the so-called Absolute Radio Frequency Channel Numbers (ARFCNs).
TDMA
Due to the very-bandwidth-efficient modulation technique (GMSK, see below), each 200-kHz sub-
band supports a data rate of 271 kbit/s. Each subband is shared by eight users. The time axis
is partitioned into timeslots, which are periodically available to each of the possible eight users
(Figure 24.2). Each timeslot is 576.92 μs long, which is equivalent to 156.25 bits. A set of eight
TDMA/FDMA
A physical channel
is denoted by timeslot
index and ARFCN
Amplitude
ARFCN
Frequency
Time
T
i
m
e
s
l
o
t

i
n
d
e
x
1 2 3 4 5 6
7
0
1
2
3
4
5
6
7
Figure 24.2 Time Division Multiple Access/Frequency Division Multiple Access system.
Adapted with permission from HP [1994] © Hewlett Packard.
4
This applies to a GSM900 system, and analogously to other frequency bands.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Interface Air
Acc` es multiple secondaire : TDMA
Chaque sous-bande de 200 kHz permet des d´ ebits de 271 kbit/s,
Chaque sous-bande est divis´ ee en 8 time-slots de env. de 156
bits, qui sont allouables p´ eriodiquement,
Une frame (trame) est compos´ ee de 8 time-slots, num´ erot ´ e 0 ` a 7,
un utilisateur acc` ede p´ eriodiquement ` a un time-slot sp´ ecifique de
chaque trame d’une bande de fr ´ equence.
Canal Physique : num´ ero de time-slot +bande de fr ´ equence,
les donn´ ees transitant sur un canal physique proviennent de
canaux dˆıts logiques
les sous bandes de 200 kHz sont num´ erot ´ ees de mani ` ere
cons´ ecutives en utilisant la num´ erotation Absolute Radio
Frequency Channel Numbers, ARFCNs.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Interface Air : exemple
Example 5.10: TDMA in GSM Figure 5.3 shows an FDMA/TDMA/FDDchannel used in
2G digital cellular in Europe (GSM). The particular example shows the eight-slot TDMA
scheme used in the GSM system. Forward and reverse channels use separate carrier
frequencies (FDD). Each carrier can support up to eight simultaneous users via TDMA,
eachusing13 kb/s encodeddigital speech, withina200 KHzcarrier bandwidth. Atotal of 124
frequency carriers (FDMA) are available in the 25 MHz allocated band in each direction.
100 kHz of band is allocated as a guard band at each edge of the overall allocated band.
Example 5.11: TDMAinDECT Figure 5.4 shows an FDMA/TDMA/TDDsystemused in
the pan-European digital PCS standard DECT. Since distances are short, a TDD format
allows use of the same frequency for forward and reverse operations. The bandwidth per
.
.
.
10 MHz
11 9 8 … 1 0
1.728 MHz
Time slots allocated to users on one carrier
Carrier 1
Carrier 2
Carrier 5
10 12 21 … 13 22 23 20
Downlink 5ms Uplink 5ms
32 bit
preamble
388 bit
data
60 bit
guard
64 control
bits
320 bit
channel
4 parity
bits
FIGURE 5.4 FDMA/TDMA/TDD in DECT.
.
.
.
25 MHz
100 kHz
guard band
8 7 6 5 4 3 2 1
200 kHz
Time slots allocated to
users on one carrier
Carrier 1
Carrier 2
Carrier 124
Uplink channels
Downlink channels
FIGURE 5.3 FDMA/TDMA/FDD in GSM.
172 NETWORKING FUNDAMENTALS
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Interface Air : fonctionnalit ´ es
Num´ erotation des time-slots en uplink et downlink
Mˆ eme indexes de time-slot entre UP et DOWN links, mais
d´ ecalage de 3 slots en liaison montante
592 Wireless Communications
timeslots is called a frame; it has a duration of 4.615 ms. Within each frame, the timeslots are
numbered from 0 to 7. Each subscriber periodically accesses one specific timeslot in every frame
on one frequency subband. The combination of timeslot number and frequency band is called the
physical channel . The kind of data that are transmitted over one such physical channel depends on
the logical channel (see also Section 24.4).
The important features of the air interface are now described in a step-by-step manner.
The Assignment of Timeslots in the Uplink and Downlink
A subscriber utilizes the timeslots with the same number (index) in the uplink and downlink.
However, numbering in the uplink is shifted by three slots relative to the numbering in the downlink.
This facilitates the design of the MS transmitter/receiver, because reception and transmission do
not occur at the same time (compare Figure 24.3).
2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1
0 1 2 3 4 5 6 7 0 1
Timeslots for uplink and downlink
Timeslot index
4
5

M
H
z
ARFCN
Timeslot index
Frame
D
o
w
n
l
i
n
k
U
p
l
i
n
k
Figure 24.3 The alignment of timeslots in the uplinks and downlink.
Adapted with permission from HP [1994] © Hewlett Packard.
The Modulation Technique
GSM uses GMSK as a modulation format. GMSK is a variant of Minimum Shift Keying (MSK);
the difference is that the data sequence is passed through a filter with a Gaussian impulse response
(time bandwidth product B
G
T = 0.3) (see Chapter 11).
This filtering is rather hard. Therefore, the spectrum is rather narrow, but there is a significant
amount of Inter Symbol Interference (ISI). On the other hand, the ISI due to delay dispersion of
the wireless channel is usually much more severe. Thus, some kind of equalization has to be used
anyway. Figure 24.4 illustrates a typical example of a phase trellis of this kind of GMSK and
of pure MSK for comparison. The detection method is not specified by the standard. Differential
detection, coherent detection, or limiter–discriminator detection might be employed.
Power Ramping
Were a transmitter to start data transmission right at the beginning of each timeslot, it would have
to be able to switch on its signal within a very short time (much shorter than a symbol period).
Similarly, at the end of a timeslot, it would have to stop transmitting abruptly, so as not to create
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Interface Air : fonctionnalit ´ es
Structure d’un time-slot (burst)
Payload transmise dans 2 time-slots de 57 bits,
Midambule : s´ equence CAZAC de synchronisation
d’apprentissage de 26 bits, servant ´ egalement d’identifiant pour
la BS,
d’autres types de bursts ont ´ et ´ e d´ efinis : access bursts,
frequency correction burst, synchronization burst,...
GSM – Global System for Mobile Communications 595
Timeslot (normal burst)
156.25 bits
576.92μs
Tail
bits
Data
Control
bit
Midamble
Control
bit
Data
Guard
period
Tail
bits
Bits 8.25 3 57 1 26 1 57 3
Figure 24.6 Functions of the bits of a normal transmission burst.
of the detection of burst data. This reduces the complexity and increases the performance of
decoding (see also Chapter 14). The timeslots end with a guard period of 8.25 bits. Apart from
“normal” transmission bursts, there are other kinds of bursts. MSs transmit access bursts to establish
initial contact with the BS. Frequency correction bursts enable frequency correction of the MSs.
Synchronization bursts allow MSs to synchronize to the frame timing of BSs. These bursts will be
explained in more detail in Section 24.4.2.
24.4 Logical and Physical Channels
In addition to the actual payload data, GSM also needs to transmit a large amount of signaling
information. These different types of data are transmitted via several logical channels. The name
stems from the fact that each of the data types is transmitted on specific timeslots that are parts of
physical channels. The first part of this section discusses the kind of data that is transmitted via
logical channels. The second part describes the mapping of logical channels to physical channels.
24.4.1 Logical Channels
Traffic CHannels (TCHs)
Payload data are transmitted via the TCHs. The payload might consist of encoded voice data or
“pure” data. There is a certain flexibility regarding the data rate: Full-rate Traffic CHannels (TCH/F)
and Half-rate Traffic CHannels (TCH/H). Two half-rate channels are mapped to the same timeslot,
but in alternating frames.
Full-Rate Traffic CHannels
• Full-rate voice channels: the output data rate of the voice encoder is 13 kbit/s. Channel coding
increases the effective transmission rate to 22.8 kbit/s.
• Full-rate data channels: the payload data with data rates of 9.6, 4.8, or 2.4 kbit/s are encoded with
Forward Error Correction (FEC) codes and transmitted with an effective data rate of 22.8 kbit/s.
Half-Rate Traffic CHannels
• Half-rate voice channels: voice encoding with a data rate as low as 6.5 kbit/s is feasible. Channel
coding increases the transmitted data rate to 11.4 kbit/s.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Interface Air : fonctionnalit ´ es
Power Ramping : gabarit dynamique en puissance
d´ efinition d’un masque en puisdance associ ´ e au time-slot
594 Wireless Communications
Principle of "power rampings"
Power
+4dB
+1dB
−1dB
–6dB
–30dB
–70dB
10μs 10μs 8μs
3 57 1 26 1 57 3
148 "active" bits (546μs) Time
10μs 10μs 8μs
−70dB
−30dB
−6dB
Figure 24.5 Power ramping during a timeslot.
Adapted with permission from HP [1994] © Hewlett Packard.
Out-of-Band Emission and Intermodulation Products
The limits for out-of-band emissions are not as severe as, e.g., for analog systems. The maximum
permitted out-of-band signal power at both BS and MS is roughly −30 dBm, which is a very high
value for wireless communications. However, in the band from 890 to 915 MHz (the uplink band),
the power emitted by the BS must not exceed −93 dBm. This is necessary because the BS has
to receive signals from MSs, with signal levels as low as −102 dBm, in this band. Furthermore,
transmit antennas are located close to receive antennas (or even colocated) at the BS, and therefore
any out-of-band emission in this band causes severe interference.
5
Similar limits apply for the
intermodulation products.
6
Structure of a Timeslot
Figure 24.6 illustrates the data contained in a timeslot with a length of 148 bits. However, not all
of these bits are payload data. Payload data are transmitted over two blocks of 57 bits. Between
these blocks is the so-called midamble. This is a known sequence of 26 bits and provides the
training for equalization, which will be covered in Section 24.7. Furthermore, the midamble serves
as an identifier of the BS. There is an extra control bit between the midamble and each of the two
data-containing blocks; the purpose of these control bits are explained in Section 24.4. Finally, the
transmission burst starts and ends with three tail bits. These bits are known, and enable termination
of Maximum Likelihood Sequence Estimation (MLSE) in defined states at the beginning and end
5
Similarly, emissions in the bands used by other systems such as UMTS have strict limits.
6
Note that intermodulation products can only be found at the BS, as only the BS transmits on several frequencies
simultaneously.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Interface Air : chaˆıne
Modulation et traitements en r ´ eception
GMSK (B
N
T = 0.3),
528 CHAPTER 6. PARTIAL-RESPONSE MODULATION
5 4 3 2 1

o
1
0.2
9(t)
1.2 ---------------------------,
0.4
0.6
0.8
Bit Periods
Figure 6.13: Impulse response g(t) for GMSK for different values of BN.
enabling us to rewrite Equation 6.37 as
Un = [ erf (n BN + ) - erf (n BN - ) ] .
(6.39)
The error functions are now expressed by the following series expansion
2 00 (-l)k z2k+l
erf(z) = yl7Tf; k!(2k + 1)
(6.40)
yielding after simplification
[
00 (-1) k b" (( TJ ) a ( TJ ) a) ]
n« = C f; ak! n + 2" - n - 2"
(6.41)
Egalisation requise,
Contr ˆ ole adaptatif de puissance : reduction jusqu’ ` a 30dB de la
puissance ´ emise, par pas de 2 dB,
Saut de fr ´ equence pour diversit ´ e fr ´ equentielle
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Mapping entre canaux logiques et physiques
Canaux logiques du GSM
~
C'"
(S"
00
~
o
en
~
0-
crq
(;.
~
n
:::r-
~
~
~
~
~
@
t!j
~
- . . ~
~
~
::s
N
o
~
::s
0..
en
~
~
~
-..Cb
......
~
~
~
Logical channels
Duplex BS H MS Control channels: CCH
Traffic channels: TCH
FEC-coded FEC-coded Broadcast CCH Common CCH Stand-alone Associated CCH
Speech Data BCCH CCCH Dedicated CCH ACCH
TCH/F TCH/F9.6 BS -t MS SDCCH BS H MS
22.8 kbit./s TCH/F4.8 BS f-t MS
TCH/F2.4 Freq.Corr.Ch: Paging Ch: PCH SDCCH/4 Fast ACCH:
22.8 kbitj's FCCH BS -t MS FACCH/F
FACCH/H
TCH/H TCH/H4.8 Synchron. Ch: Random Access Ch: SDCCH/8 Slow ACCH:
11.4 kbit/s TCH/H2.4 SCH RACH SACCH/TF
11.4 kbit/s MS -+ BS SACCH/TH
SACCH/C4
SACCH/C8
General Inf. Access Grant Ch:
AGCH
BS -+ MS
~
....
o
o
=:
>
~
8
t;!j
~
~
c
tn
~
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Mapping entre canaux logiques et physiques
Structure multi-trame
.
1 hyps-r Fr-one = 2048 superfro.l"'les = 2,715,648 TDMA f r o ne s ( 3 hours, 28 Minutes, .,,)

-...

>

....
2
o
e-
o
o
....
o
>
t"4
o
== >
2
Z
t;!j
t""
en
58 Encrypted bits
0.5i7MS)
--
26 bits Training Seg.
1 til"'leslot = 056.25 bit duro. t rons
58 Encrypted bits
TB
3
tIEl - -0
./ -
./
./
/./
./
/./
./
./ -
./ --
/? ( 1 bit duro.tion s.69us) - - - -
GITEI- -
\ ------
\
\\ 1 super-Fr-one = 1326 TDMA fro.Mes (6.125)
- ----
egTCHIFS -003 bffij
e.g.BCCH 0, 1 24 25
/ " /"
/, /"/" -,
/ ", ,.- /" ""
MultifrQMe = 26 TDMA Fr c.ne s <l20MS) 1 l"'Iultifro.l"'le = 51 TDMA fro.Mes (2351"'15)
I ,. IcHe/SACCH ...
ffiJ IT]
/
1 TDMA fro.l"IE? = 8 til"lE?slots (4.615I"1s)

ciQ-
=
11
It
00

l:J'"
(t)
C1
sr:

t:'

>

S
en
<::"t-
t-t

o
<::"t-

t-t

@
tJj
1-3

§
N
o
§
c,
o:

C'D

CD
CD

Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Sch´ ema traitement couche physique
Structure ´ emetteur-r ´ ecepteur
8.3. MAPPING LOGICAL CHANNELS 675
Chu.el
Figure 8.4: Block diagram of the TCH/FS channel.
ure 8.2 is provided to prevent burst overlapping due to delay fluctuations.
Finally, a 26-bit equaliser training segment is included in the centre of the
normal traffic burst. This segment is constructed by a 16-bit Viterbi chan-
nel equaliser training pattern surrounded by five quasi-periodically repeated
bits on both sides. These bits provide a sufficiently long quasi-periodic ex-
tension of the training sequence before useful data are entered into the
modulator, to keep the side-lobes of its autocorrelation function, as well as
those of its spectra, sufficiently low. The 16-bit pattern was determined
by evaluating the autocorrelation of the modulated signals due to all 2
16
candidate training sequences and selecting the one with the highest auto-
correlation peak, while maintaining a low main lobe to side-lobes ratio. For
GMSK with BT==0.3 and modulation index of 0.5 several good sequences
can be found. Since the MS has to be informed about which BS it com-
municates with, for neighbouring BSs different training patterns are used.
Therefore, in the GSM system the eight best training patterns are used to
be associated with eight different BS colour codes.
This 156.25 bit duration TCH/FS normal burst (NB) constitutes the
basic timeslot of the TDMA frame structure, which is input to the GMSK
modulator at a bit rate of approximately 271 kbit / s. Since the bit inter-
val is 3.69 /-ts, the timeslot duration is 156.25· 3.69 ~ 0.577 ms. Eight
such normal bursts of eight appropriately staggered users are multiplexed
onto one RF carrier giving a TDMA frame of 8 . 0.577 ~ 4.615 ms
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Codage canal pour la voix
Codage avec protection in´ egale contre les erreurs
614 CHAPTER 8. GSM
260 bits/20MS 13 kbps
C1b
132 bits
/
/
/
/
Po.r rt y /
Check
/
/
/
/
/
132
/
/
/
189 bits
/
/
'/
/
/
Convolutional Code
/
/
r = 1/2, k = 5
I
/
/
378
/
78 /
Figure 8.3: Mapping the TCH/FS logical channel onto a physical channel.
@ETT, Hanzo and Steele, 1994.
down smoothly, as will be highlighted in the context of the power ramping
mask of Figure 8.29, which is necessary for mitigating the spurious adjacent
channel emissions.
This mapping process is also summarised in the form of a hardware
oriented block diagram in Figure 8.4, where the Voice Activity Detector
(VAD) is included to enable or disable transmissions depending on whether
speech is deemed to be present at the input of the RPE speech encoder.
This allows a substantial reduction of the power consumption as well as that
of the interferences imposed on other users. The effects of subjectively
annoying silent periods are mitigated at the receiver by adding comfort
noise during these intervals.
The 8.25 bit-interval duration guard space (GP) at the bottom of Fig-
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Sch´ emas de codage canal et d´ etection d’erreur
Code d´ etecteur d’erreur :
604 Wireless Communications
results in 378 bits which are transmitted together with the 78 bits of class 2. Thus, for every 20 ms
of the voice signal, 456 bits have to be transmitted. In the following, the details of the different
encoder blocks will be discussed.
Block Encoding
Block Encoding of Voice Data As discussed above, only class-1a bits of the voice data are
encoded using a (53,50) block code. This is a very “weak” block code. It is only supposed to
detect bit errors and cannot detect more than three bit errors within the 50 class 1a bits reliably.
However, this is sufficient, since a block is completely discarded if an error is detected within the
class-1a bits; the receiver then smoothes the resulting signal by “inventing” a block. Figure 24.12
shows the linear shift register representation of the block encoder. As the code is systematic, the
50 data bits pass through the encoder unchanged. However, each of them impacts the state of the
shift register. The final state of the shift register determines the 3 parity bits which are attached to
the 50 class-1a bits. Class 1a, 1b, and parity check bits are then reordered and interleaved. Finally,
four all-zero tail bits are attached, which are needed for the convolutional decoder (see below).
Generator-Polynomial
G
4
(D) = D
3
+ D + 1
1 ... 50 CKL: SW closed
51 ... 53 CKL: SW open
D D D
+ +
Data
Last 3 parity bits
First 50 bits
SW
Figure 24.12 Shift register structure for voice block encoding, Cla (53,50) systematic, cyclic block encoder.
In this figure: CKL, Clock; SW, Switch.
Reproduced with permission from Steele and Hanzo [1999] © J. Wiley & Sons, Ltd.
Block Encoding of Signaling Data As mentioned in Section 24.4.1, the signaling information has
to have stronger protection against bit errors than the voice data. While a bit error in voice-related
data might lead to an unintelligible audio signal for 20 ms, a bit error in signaling bits can have a
more severe impact – e.g., handover to a wrong cell and therefore loss of connection. Thus, higher
redundancy is required. For most of the control channels, only 184 signal bits are transmitted within
20 ms (instead of 260 for speech). This allows better error correction. Signaling bits are encoded
with a (224,184) Fire code. The Fire code is defined by the generator polynomial:
G(D) = D
40
+D
26
+D
23
+D
17
+D
3
+1 (24.1)
Fire codes are block codes which are particularly capable of correcting burst errors. Burst errors are
defined as a series of bit errors, meaning that two or more consecutive bits are wrong; such error
bursts occur, e.g., when Viterbi decoding fails (see Chapter 14). A total of 4 tail bits are attached
to the resulting 224 bits. The result is fed into the convolutional encoder at code rate-1/2, which is
the same as that used for class 1 of the voice signal. For selected logical signaling channels, such
as RACH and SCH, different generator polynomials are used. The interested reader is referred to
Steele and Hanzo [1999] and the GSM specifications.
Code Fire :
604 Wireless Communications
results in 378 bits which are transmitted together with the 78 bits of class 2. Thus, for every 20 ms
of the voice signal, 456 bits have to be transmitted. In the following, the details of the different
encoder blocks will be discussed.
Block Encoding
Block Encoding of Voice Data As discussed above, only class-1a bits of the voice data are
encoded using a (53,50) block code. This is a very “weak” block code. It is only supposed to
detect bit errors and cannot detect more than three bit errors within the 50 class 1a bits reliably.
However, this is sufficient, since a block is completely discarded if an error is detected within the
class-1a bits; the receiver then smoothes the resulting signal by “inventing” a block. Figure 24.12
shows the linear shift register representation of the block encoder. As the code is systematic, the
50 data bits pass through the encoder unchanged. However, each of them impacts the state of the
shift register. The final state of the shift register determines the 3 parity bits which are attached to
the 50 class-1a bits. Class 1a, 1b, and parity check bits are then reordered and interleaved. Finally,
four all-zero tail bits are attached, which are needed for the convolutional decoder (see below).
Generator-Polynomial
G
4
(D) = D
3
+ D + 1
1 ... 50 CKL: SW closed
51 ... 53 CKL: SW open
D D D
+ +
Data
Last 3 parity bits
First 50 bits
SW
Figure 24.12 Shift register structure for voice block encoding, Cla (53,50) systematic, cyclic block encoder.
In this figure: CKL, Clock; SW, Switch.
Reproduced with permission from Steele and Hanzo [1999] © J. Wiley & Sons, Ltd.
Block Encoding of Signaling Data As mentioned in Section 24.4.1, the signaling information has
to have stronger protection against bit errors than the voice data. While a bit error in voice-related
data might lead to an unintelligible audio signal for 20 ms, a bit error in signaling bits can have a
more severe impact – e.g., handover to a wrong cell and therefore loss of connection. Thus, higher
redundancy is required. For most of the control channels, only 184 signal bits are transmitted within
20 ms (instead of 260 for speech). This allows better error correction. Signaling bits are encoded
with a (224,184) Fire code. The Fire code is defined by the generator polynomial:
G(D) = D
40
+D
26
+D
23
+D
17
+D
3
+1 (24.1)
Fire codes are block codes which are particularly capable of correcting burst errors. Burst errors are
defined as a series of bit errors, meaning that two or more consecutive bits are wrong; such error
bursts occur, e.g., when Viterbi decoding fails (see Chapter 14). A total of 4 tail bits are attached
to the resulting 224 bits. The result is fed into the convolutional encoder at code rate-1/2, which is
the same as that used for class 1 of the voice signal. For selected logical signaling channels, such
as RACH and SCH, different generator polynomials are used. The interested reader is referred to
Steele and Hanzo [1999] and the GSM specifications.
Code convolutif :
GSM – Global System for Mobile Communications 605
Convolutional Encoding
Both the class-1 bits of the voice data and all of the signaling information are encoded with a
convolutional coder at code rate-1/2 (see Section 14.3). The bits are fed into a 5-bit shift register.
For each new input bit, two codebits are calculated according to the generator polynomials
G1(D) = 1 +D +D
3
+D
4
G2(D) = 1 +D
3
+D
4

(24.2)
and transmitted. The 4 final tail bits attached to the input sequence ensure that the encoder terminates
in the all-zero state at the end of each encoded block.
Interleaving
Due to the nature of fading channels, bit errors may occur in bursts in some transmission
blocks – e.g., if those blocks were transmitted during a deep fade. Interleaving orders the bits
in such a manner that the burst errors due to the channel are (hopefully) distributed evenly (see
Section 14.7.1) Evidently, the more the interleaver distributes corrupted bits, the better. However,
latency of the speech signal puts an upper limit on interleaver depth: In order to give acceptable
speech quality, the delay of the signal should be less than 100 ms.
GSM interleaves the data of two blocks (henceforth called “a” and “b”) in the following way:
first, each of the blocks is divided into eight subblocks. Specifically, each bit receives an index
i ∈ {0, . . . , 455}, and the bits are sorted into subblocks with index k ∈ {0, . . . , 8} according to
k = i mod 8. Each subblock of block “a” contributes one half of the bits in a transmission burst
(114 bits). The other half is associated with subblocks of either a previous or a succeeding block
“b.” Figure 24.13 illustrates diagonal interleaving.
Frame number
i + 0 i + 1 i + 2 i + 3 i + 4 i + 5 i + 6 i + 7
0a 4b 1a 5b 2a 6b 3a 7b 4a 0b 5a 1b 6a 2b 7a 3b
114 Bits 114 Bits
Figure 24.13 Diagonal interleaving for traffic channel/slow associated control channel/fast associated control
channel data.
Reproduced with permission from Rappaport [1996] © IEEE.
24.6.3 Cryptography
One of the most severe shortcomings of analog mobile communications was the ease with which
it could be intercepted. Anybody with a frequency scanner was able to eavesdrop on phone
conversations. This posed a threat – e.g., for business people dealing with confidential material.
Furthermore, even political scandals have been known to develop as a consequence of eavesdropped
conversations.
In a digital system, this problem can be solved by “standard” means: once the audio signal
is represented by a bitstream, cryptographic procedures, which had long before been developed
for military applications, can be easily applied. For GSM, intercepting a conversation requires a
man-in-the-middle attack, which involves implementing a BTS, to which the target MS would log
Entrelacement :
GSM – Global System for Mobile Communications 605
Convolutional Encoding
Both the class-1 bits of the voice data and all of the signaling information are encoded with a
convolutional coder at code rate-1/2 (see Section 14.3). The bits are fed into a 5-bit shift register.
For each new input bit, two codebits are calculated according to the generator polynomials
G1(D) = 1 +D +D
3
+D
4
G2(D) = 1 +D
3
+D
4

(24.2)
and transmitted. The 4 final tail bits attached to the input sequence ensure that the encoder terminates
in the all-zero state at the end of each encoded block.
Interleaving
Due to the nature of fading channels, bit errors may occur in bursts in some transmission
blocks – e.g., if those blocks were transmitted during a deep fade. Interleaving orders the bits
in such a manner that the burst errors due to the channel are (hopefully) distributed evenly (see
Section 14.7.1) Evidently, the more the interleaver distributes corrupted bits, the better. However,
latency of the speech signal puts an upper limit on interleaver depth: In order to give acceptable
speech quality, the delay of the signal should be less than 100 ms.
GSM interleaves the data of two blocks (henceforth called “a” and “b”) in the following way:
first, each of the blocks is divided into eight subblocks. Specifically, each bit receives an index
i ∈ {0, . . . , 455}, and the bits are sorted into subblocks with index k ∈ {0, . . . , 8} according to
k = i mod 8. Each subblock of block “a” contributes one half of the bits in a transmission burst
(114 bits). The other half is associated with subblocks of either a previous or a succeeding block
“b.” Figure 24.13 illustrates diagonal interleaving.
Frame number
i + 0 i + 1 i + 2 i + 3 i + 4 i + 5 i + 6 i + 7
0a 4b 1a 5b 2a 6b 3a 7b 4a 0b 5a 1b 6a 2b 7a 3b
114 Bits 114 Bits
Figure 24.13 Diagonal interleaving for traffic channel/slow associated control channel/fast associated control
channel data.
Reproduced with permission from Rappaport [1996] © IEEE.
24.6.3 Cryptography
One of the most severe shortcomings of analog mobile communications was the ease with which
it could be intercepted. Anybody with a frequency scanner was able to eavesdrop on phone
conversations. This posed a threat – e.g., for business people dealing with confidential material.
Furthermore, even political scandals have been known to develop as a consequence of eavesdropped
conversations.
In a digital system, this problem can be solved by “standard” means: once the audio signal
is represented by a bitstream, cryptographic procedures, which had long before been developed
for military applications, can be easily applied. For GSM, intercepting a conversation requires a
man-in-the-middle attack, which involves implementing a BTS, to which the target MS would log
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Sch´ ema de codage canal pour les donn´ ees de signalisation
8.3. MAPPING LOGICAL CHANNELS
184 Bits
677
/
/
Fire-Code (224,184) /
/
/
/
40 26 23 17 3 /
G (D)= D + D + D + D + D +
/
tQiling
5
/
.nf or nc t ron bits: 184 po r rt y 40 4
456
Figure 8.5: Mapping of SACCR, FACCH, BCCH, SDCCH, PCH and AGCH
logical channels.
the SACCH, as also shown in Figure 8.5, but its 456-bit frame is mapped
onto eight consecutive 114-bit TDMA traffic bursts, exactly as specified for
TCR/FS. This is carried out by stealing the even bits of the first four and
the odd bits of the last four bursts, which is signalled by setting hu = 1,
hl = 0 and hu = 0, hl = 1 in the first and last bursts, respectively. The
unprotected FACCR information rate is 184 bits/20 ms=9.2 kbps, which is
transmitted after concatenated error protection at a rate of 22.8 kbps. The
repetition delay is 20 ms and the interleaving delay is 8 . 4.615 = 37 ms,
resulting in a total delay of 57 ms.
In a subsequent stage of Figure 8.2 51 TCH/FS multiframes are amal-
gamated into one superframe lasting 51 . 120 ms = 6.12 s, which contains
26·51 = 1326 TDMA frames. There would be no need for any further levels
of TDMA hierarchy, if it was not for the encryption, which uses the TDMA
frame number (FN) as a parameter in its algorithm. However, with 1326
FNs only the encryption rule is not sufficiently secure. Therefore 2048 su-
perframes are concatenated to form a hyperframe of 1326·2048 = 2 715 648
TDMA frames lasting 2048·6.12 s ~ 3 h 28 min, using a satisfactorily high
number of FNs in the encryption algorithm. This step now concludes our
example of mapping the TCH/FS and its SACCH logical channel onto an
appropriate physical channel constituted by a specific timeslot dedicated
to a specific user of a specific RF channel carrying the messages of eight
TDMA users. To reduce the complexity of MSs they do not have to receive
and transmit simultaneously, their receive and transmit timeslots carrying
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Global System for Mobile communications
Structure des bursts de contr ˆ ole et signalisation
602 Wireless Communications
Normal
3 start
bits
58 data bits
(encrypted)
26 training
bits
58 data bits
(encrypted)
3 stop
bits
8.25 bits
guard period
FCCH burst
3 start
bits
142 zeros
3 stop
bits
8.25 bits
guard period
SCH burst
3 start
bits
39 data bits
(encrypted)
64 training
bits
3 stop
bits
39 data bits
(encrypted)
8.25 bits
guard period
RACH burst
8 start
bits
41 synchronization
bits
36 data bits
(encrypted)
3 stop
bits
68.25 bits extended
guard period
Dummy burst
3 start
bits
58 mixed bits
26 training
bits
58 mixed bits
3 stop
bits
8.25 bits
guard period
Figure 24.10 Structure of timeslots in the Global System for Mobile communications.
Reproduced with permission from Rappaport [1996] © IEEE.
24.6 Coding
To transmit speech via the physical GSM channel the “speech signals” have to be translated into
digital signals. This process should maintain a certain speech quality while keeping the required
data rate as low as possible (see also Chapter 15). Different forms of speech coding were considered
for GSM, and finally a Regular Pulse Excited with Long Term Prediction (RPE-LTP) solution was
chosen (see Chapter 15). The digitized speech that is obtained in such a way then has to be protected
by FEC in order to remain intelligible when transmitted over typical cellular channels (uncoded
Bit Error Rates (BERs) of ∼10
−3
to 10
−1
). Both block and convolutional codes are used for this
purpose in GSM.
Thus, voice transmission in GSM represents a typical example of the paradox of speech com-
munications. First, redundancy is removed from the source data stream during the speech-coding
process, and then redundancy is added in the form of error-correcting coding before transmission.
The reason for this approach is that the original redundancy of the speech signal is rather inefficient
at ensuring intelligibility of speech when transmitted over wireless channels. In this section, we
first describe voice encoding, and subsequently channel coding; these can be seen as important
applications of the principles expounded in Chapters 15 and 14, respectively.
24.6.1 Voice Encoding
Like most voice encoders (also referred to as vocoders), the GSM vocoder is not a classical source-
coding processes like, e.g., the Huffman code. Rather, GSM uses a lossy compression method,
meaning that the original signal cannot be reconstructed perfectly, but that the compression and
decompression procedures lead to a signal which is similar enough to the original one to allow
comfortable voice communications. As GSM has evolved, so has the speech coder. For the first
release of GSM, an RPE-LTP approach was used. The idea behind this approach is to consider
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
OFDM : principe de base
Orthogonal Frequency Division
Multiplexing
Originellement introduit pour
traiter efficacement les
interf ´ erences entre symboles pour
les canaux fortement dispersifs,
Principe : transformer un canal
large bande en un certains
nombre de canaux bande ´ etroite
de largeur plus petite que la
bande de coh´ erence du canal.
⇒flat fading sur chaque canal
116 Chapter 4 • Orthogonal Frequency Division Multiplexing
Although this simple type of multicarrier modulation is easy to understand, it has several
crucial shortcomings. First, in a realistic implementation, a large bandwidth penalty will be
inflicted, since the subcarriers can’t have perfectly rectangular pulse shapes and still be time lim-
ited. Additionally, very high quality (and hence, expensive) low-pass filters will be required to
Figure 4.2 A basic multicarrier receiver: Each subcarrier is decoded separately, requiring L inde-
pendent receivers.
Figure 4.3 The transmitted multicarrier signal experiences approximately flat fading on each
subchannel, since , even though the overall channel experiences frequency-selective
fading: .
P/S
Demod..
1
cos(2 f
c
)
Demod
2
Demod.
L
cos(2 f
c
+(L–1) f)
.
.
.
cos(2 f
c
+ f)
LPF
LPF
LPF
s p b R ) t ( y
f
1
f
2
f
L
B
B
c
f
|H(f)|
B/L
B L B
c
/
B B
c
>
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
OFDM : principe du pr ´ efixe cyclique
4.3 An Example: OFDM in WiMAX 123
3. In order for the IFFT/FFT to decompose the ISI channel into orthogonal subcarriers, a
cyclic prefix of length must be appended after the IFFT operation. The resulting
symbols are then sent in serial through the wideband channel.
4. At the receiver, the cyclic prefix is discarded, and the L received symbols are demodulated,
using an FFT operation, which results in L data symbols, each of the form
for subcarrier l.
5. Each subcarrier can then be equalized via an FEQ by simply dividing by the complex
channel gain for that subcarrier. This results in .
We have neglected a number of important practical issues thus far. For example, we have
assumed that the transmitter and the receiver are perfectly synchronized and that the receiver
perfectly knows the channel, in order to perform the FEQ. In the next section, we present the
implementation issues for OFDM in WiMAX.
4.3 An Example: OFDM in WiMAX
To gain an appreciation for the time- and frequency-domain interpretations of OFDM, WiMAX
systems can be used as an example. Although simple in concept, the subtleties of OFDM can be
confusing if each signal-processing step is not understood. To ground the discussion, we con-
sider a passband OFDM system and then give specific values for the important system
parameters.
Figure 4.8 shows a passband OFDM modulation engine. The inputs to this figure are L
independent QAM symbols (the vector X), and these L symbols are treated as separate subcarri-
ers. These L data-bearing symbols can be created from a bit stream by a symbol mapper and
serial-to-parallel convertor (S/P). The L-point IFFT then creates a time-domain L-vector x that is
cyclic extended to have length , where G is the fractional overhead. This longer vector
is then parallel-to-serial (P/S) converted into a wideband digital signal that can be amplitude
modulated with a single radio at a carrier frequency of .
Figure 4.7 An OFDM system in vector notation.
n
x
X
L-pt
IDFT
L-pt
DFT
Y
P/S
Add
CP
+ h[n]
Delete
CP
S/P
y
FEQ X
^
Time Domain
Frequency Domain
A circular channel: y = h x + n
v L v +
Y H X N
l l l l
= +
H i [ ]
ˆ
l l l l
X N H = / + X
L G (1 ) +
f
c c
= /2 ω π
124 Chapter 4 • Orthogonal Frequency Division Multiplexing
This procedure appears to be relatively straightforward, but in order to be a bit less abstract,
we now use some plausible values for the parameters. (Chapter 8 enumerates all the legal values
for the OFDM parameters B, L, L
d
, and G.) The key OFDM parameters are summarized in
Table 4.1, along with some potential numerical values for them. As an example, if 16 QAM
modulation were used (M = 16), the raw (neglecting coding) data rate of this WiMAX system
would be
(4.12)
(4.13)
In words, each data-carrying subcarriers of bandwidth carries bits of data.
An additional overhead penalty of must be paid for the cyclic prefix, since it consists of
redundant information and sacrifices the transmission of actual data symbols.
4.4 Timing and Frequency Synchronization
In order to demodulate an OFDM signal, the receiver needs to perform two important synchroni-
zation tasks. First, the timing offset of the symbol and the optimal timing instants need to be
determined. This is referred to as timing synchronization. Second, the receiver must align its car-
rier frequency as closely as possible with the transmitted carrier frequency. This is referred to as
frequency synchronization. Compared to single-carrier systems, the timing-synchronization
requirements for OFDM are in fact somewhat relaxed, since the OFDM symbol structure natu-
rally accommodates a reasonable degree of synchronization error. On the other hand, frequency-
synchronization requirements are significantly more stringent, since the orthogonality of the
data symbols is reliant on their being individually discernible in the frequency domain.
Figure 4.8 Closeup of the OFDM baseband transmitter
IFFT
P/S
Speed = B/L Hz
L Subcarriers
Speed = B/L Hz
L(1 + G) Samples
Serial
Stream at
B(1 + G) Hz
Cyclic Prefix of
LG Samples
QAM
Symbols
(X)
D/A X
Analog
Baseband
Multicarrier
Signal
RF
Multicarrier
Signal
exp(j
c
)
R
B
L
L M
G
d
=
( )
1
2
log
+
=
10
1024
768 (16)
1.125
= 24 .
7
2
log MHz
Mbps
L
d
B L /
2
( ) log M
(1 ) + G
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
OFDM : structure ´ emetteur-r ´ ecepteur en mono-utilisateur
4.2 OFDM Basics 117
maintain the orthogonality of the subcarriers at the receiver. Most important, this scheme
requires L independent RF units and demodulation paths. In Section 4.2, we show how OFDM
overcomes these shortcomings.
4.2 OFDM Basics
In order to overcome the daunting requirement for L RF radios in both the transmitter and the
receiver, OFDM uses an efficient computational technique, discrete Fourier transform (DFT),
which lends itself to a highly efficient implementation commonly known as the fast Fourier
transform (FFT). The FFT and its inverse, the IFFT, can create a multitude of orthogonal subcar-
riers using a single radio.
4.2.1 Block Transmission with Guard Intervals
We begin by grouping L data symbols into a block known as an OFDM symbol. An OFDM sym-
bol lasts for a duration of T seconds, where . In order to keep each OFDM symbol inde-
pendent of the others after going through a wireless channel, it is necessary to introduce a guard
time between OFDM symbols:
This way, after receiving a series of OFDM symbols, as long as the guard time is larger
than the delay spread of the channel , each OFDM symbol will interfere only with itself:
Put simply, OFDM transmissions allow ISI within an OFDM symbol. But by including a suffi-
ciently large guard band, it is possible to guarantee that there is no interference between subse-
quent OFDM symbols.
4.2.2 Circular Convolution and the DFT
Now that subsequent OFDM symbols have been rendered orthogonal with a guard interval, the
next task is to attempt to remove the ISI within each OFDM symbol. As described in Chapter 3,
when an input data stream is sent through a linear time-invariant Finite Impulse Response
(FIR) channel , the output is the linear convolution of the input and the channel:
. However, let’s imagine computing in terms of a circular convolution:
(4.1)
T LT
s
=
OFDM Symbol OFDM Symbol OFDM Symbol Guard Guard
T
g
τ
OFDM Symbol OFDM Symbol OFDM Symbol
Delay Spread
x n [ ]
h n [ ]
y n x n h n [ ] = [ ]* [ ] y n [ ]
y n x n h n h n x n [ ] = [ ] [ ] = [ ] [ ],
Intervales de garde ⇒pas d’IES inter-symboles OFDM
4.2 OFDM Basics 119
4.2.3 The Cyclic Prefix
The key to making OFDM realizable in practice is the use of the FFT algorithm, which has
low complexity. In order for the IFFT/FFT to create an ISI-free channel, the channel must
appear to provide a circular convolution, as seen in Equation (4.4). Adding cyclic prefix to the
transmitted signal, as is shown in Figure 4.4, creates a signal that appears to be , and so
.
Let’s see how this works. If the maximum channel delay spread has a duration of sam-
ples, adding a guard band of at least samples between OFDM symbols makes each OFDM
symbol independent of those coming before and after it, and so only a single OFDM symbol can
be considered. Representing such an OFDM symbol in the time domain as a length L vector gives
(4.8)
After applying a cyclic prefix of length , the transmitted signal is
(4.9)
The output of the channel is by definition y
cp
= h * x
cp
, where h is a length vector describ-
ing the impulse response of the channel during the OFDM symbol.
4
The output y
cp
has
samples. The first samples of y
cp
contain interference from the
preceding OFDM symbol and so are discarded. The last samples disperse into the subsequent
OFDM symbol, so also are discarded. This leaves exactly L samples for the desired output y,
which is precisely what is required to recover the L data symbols embedded in x.
Our claim is that these L samples of y will be equivalent to y = h ⊗ x. Various proofs are
possible; the most intuitive is a simple inductive argument. Consider , the first element in y.
As shown in Figure 4.5, owing to the cyclic prefix, depends on and the circularly wrapped
values . That is:
Figure 4.4 The OFDM cyclic prefix
4. It can generally be reasonably assumed that the channel remains constant over an OFDM symbol,
since the OFDM symbol time T is usually much less than the channel coherence time, T
c
.
x n
L
[ ]
y n x n h n [ ] = [ ] [ ]
x
L-v
x
L-v+1
... x
L-1
x
0
x
1
x
2
... x
L-v-1
x
L-v
x
L-v+1
... x
L-1
Cyclic Prefix
Copy and paste last v symbols.
OFDM Data Symbols
v +1
v
x = [ ].
1 2
x x x
L

v
x
cp L v L v L
x x x x x x = [
1 1 0 1 − − + −
… …

Cyclic Prefix
LL−1
].
Original Data

v +1
( ) ( 1) 1 = 2 L v v L v + + + − + v
v
y
0
y
0
x
0
x x
L v L − −

1
Pr ´ efixe cyclique : rendre la convolution avec le canal circulaire
4.4 Timing and Frequency Synchronization 127
On the other hand, if the timing offset is not within this window ,
intersymbol interference occurs regardless of whether the phase shift is appropriately accounted
for. This can be confirmed intuitively for the scenario that and for . For the
case , the receiver loses some of the desired energy, since only the delayed version of the
early samples is received, and incorporates undesired energy from the subsequent sym-
bol. Similarly for : Desired energy is lost while interference from the preceding sym-
bol is included in the receive window. For both of these scenarios, the SNR loss can be
approximated by
(4.16)
which makes intuitive sense and has been shown more rigorously in the literature on synchroni-
zation for OFDM [40]. Important observations from this expression follow.
• SNR decreases quadratically with the timing offset.
• Longer OFDM symbols are increasingly immune from timing offset; that is, more subcar-
riers help.
• Since in general , timing-synchronization errors are not that critical as long as the
induced phase change is corrected.
In summary, to minimize SNR loss owing to imperfect timing synchronization, the timing errors
should be kept small compared to the guard interval, and a small margin in the cyclic prefix
length is helpful.
4.4.2 Frequency Synchronization
OFDM achieves a higher degree of bandwidth efficiency than do other wideband systems. The
subcarrier packing is extremely tight compared to conventional modulation techniques, which
require a guard band on the order of 50 percent or more, in addition to special transmitter archi-
tectures, such as the Weaver architecture or single-sideband modulation, that suppress the redun-
dant negative-frequency portion of the passband signal. The price to be paid for this bandwidth
Figure 4.10 Timing-synchronization margin
CP CP L Data Symbols L Data Symbols
Delay Spread (v samples, T
m
sec)
Synchronization Margin (N
g
– v samples, T
g
– T
m
sec)
τ 0 ≤ ≤ − τ T T
m g
τ > 0 τ < T T
m g

τ > 0
x x
0 1
, ,...
τ < T T
m g

∆SNR
LT
s
( ) 2 ,
2
τ
τ
≈ −






τ LT
s
Ajout Pr ´ efixe cyclique : plus IES intra symbole OFDM
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
Egalisation Monoporteuse dans le domaine fr ´ equentiel : SC-FDE vs OFDM
P1: OTA/XYZ P2: ABC
c02 JWBK249/Myung August 27, 2008 19:33 Printer Name: Yet to Come
Channel Characteristics and Frequency Multiplexing 31
Using DFT, the frequency domain equalization can be easily imple-
mented using modern digital signal processors (DSP). Because the DFT
size does not grow linearly with the length of the channel response, the
complexity of FDE is much lower than that of the equivalent time domain
equalizer for broadband channels.
Single carrier modulation with frequency domain equalization
(SC/FDE) is a practical technique for mitigating the effects of frequency-
selective fading. It delivers performance similar to OFDM with essentially
the same overall complexity, even for a long channel impulse response
[6],[7]. Figure 2.13 shows the block diagrams of an SC/FDE receiver and,
for comparison, an OFDM receiver. We can see that both systems use the
same communication component blocks and the only difference between
the two diagrams is the location of the IDFT block. Thus, one can expect
the two systems to have similar link level performance and spectral effi-
ciency.
An SC/FDE modulator transmits modulation symbols sequentially. It di-
vides the sequence of modulation symbols into blocks and adds a cyclic
prefix (CP) to the beginning of each block. The CP is a copy of the last
part of the block as shown in Figure 2.14. As in OFDM, the CP prevents
inter-block interference. It also ensures that the convolution of the channel
impulse response with the modulated symbols has the form of a circular
convolution. This matches the signal processing performed by the channel
with the signal processing performed by the FDE because multiplication in
the DFT-domain is equivalent to circular convolution in the time domain.
Channel
N-
point
IDFT
Equalization
N-
point
DFT
SC/FDE
OFDM
Detect
Remove
CP
Add
CP/
PS
* CP: Cyclic Prefix, PS: Pulse Shaping
Channel Equalization
N-
point
DFT
Detect
Remove
CP
N-
point
IDFT
Add
CP/
PS
Figure 2.13 Block diagrams of SC/FDE and OFDM systems
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
SC-FDE vs OFDM : r ´ ecepteurs
P1: OTA/XYZ P2: ABC
c02 JWBK249/Myung August 27, 2008 19:33 Printer Name: Yet to Come
Channel Characteristics and Frequency Multiplexing 33
Equalizer
Equalizer
Equalizer
Detect
Detect
Detect
Equalizer IDFT Detect SC/FDE
OFDM
DFT
DFT
Figure 2.15 OFDM and SC/FDE receivers
Also, the duration of the modulated time symbols is expanded in the case
of OFDM with parallel transmission of the data block during the elongated
time period, as shown in Figure 2.16. Also, the system bandwidth B
s
Hz
is subdivided into smaller-bandwidth subcarriers and the individual data is
conveyed on each subcarrier.
In summary, SC/FDE has advantages over OFDM as follows:
r
low PAPR due to single carrier modulation at the transmitter;
r
robustness to spectral null;
T seconds
B
S
H
z
B
S
H
z
OFDM
Single carrier
Figure 2.16 OFDM and SC/FDE symbols: there are six data symbols and the
system bandwidth is B
s
Hz
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
SC-FDE vs OFDM : interpr ´ etation dans le plan temps-fr ´ equence
P1: OTA/XYZ P2: ABC
c02 JWBK249/Myung August 27, 2008 19:33 Printer Name: Yet to Come
Channel Characteristics and Frequency Multiplexing 33
Equalizer
Equalizer
Equalizer
Detect
Detect
Detect
Equalizer IDFT Detect SC/FDE
OFDM
DFT
DFT
Figure 2.15 OFDM and SC/FDE receivers
Also, the duration of the modulated time symbols is expanded in the case
of OFDM with parallel transmission of the data block during the elongated
time period, as shown in Figure 2.16. Also, the system bandwidth B
s
Hz
is subdivided into smaller-bandwidth subcarriers and the individual data is
conveyed on each subcarrier.
In summary, SC/FDE has advantages over OFDM as follows:
r
low PAPR due to single carrier modulation at the transmitter;
r
robustness to spectral null;
T seconds
B
S
H
z
B
S
H
z
OFDM
Single carrier
Figure 2.16 OFDM and SC/FDE symbols: there are six data symbols and the
system bandwidth is B
s
Hz
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
Acc` es multiples bas´ es OFDM
OFDM-FDMA
On assigne diff ´ erentes sous-porteuses ` a chaque utilisateurs,
La fac¸on d’allouer varie suivant la strat ´ egie d’optimisation mise
en oeuvre :
Localized FDMA/ Block FDMA (LFDMA) : les sous-porteuses sont
attribu´ ees par sous-blocs,
Interleaved FDMA (IFDMA) : les sous-porteuses sont attribu´ ees de
mani ` ere entrelac´ ee,
Allocation dynamique des porteuses possible pour gain en
diversit ´ e en utilisant des algorithmes de scheduling
180 5 Principles of OFDMA
time
f
r
e
q
u
e
n
c
y
TDMA-OFDM
U1 U2 U3 U4 U5 U6 U1
t t t t t t
round robin
Fig. 5.3 TDMA-OFDM: There are 6 users (U), and TDMA scheme has fixed time intervals be-
tween frames (t) and fixed packet sizes
time
f
r
e
q
u
e
n
c
y
FDMA-OFDM
U1-U6 U1-U6 U1-U6
t t
U1-U6 U1-U6 U1-U6
t t t t
U1-U6
Blocked FDMA
Fig. 5.4 FDMA-OFDM: There are 6 users (U), and Block-FDMA scheme has fixed time intervals
between frames (t) and fixed subcarrier allocation
applied to all subcarriers. If the channel is low quality for a user in the time slot,
either transmission experiences high bit error rate with the higher modulation, or
lower modulation is selected by reducing the bit rate. Also, if there is no data to
transmit, the bandwidth is wasted.
5.1.3 Frequency Division: FDMA-OFDM
FDMA-OFDM is similar to TDMA-OFDM. FDMA-OFDM creates orthogonal re-
source by dividing the available subcarriers into fixed sets where each set is used by
a user as seen in Fig. 5.4. Fixed allocation does not change over time but the num-
ber of subcarriers associated to a user may differ. Albeit, allocation is fixed, which
prevents leveraging the multiuser diversity.
There are two types of allocation: localized FDMA (LFDMA) or interleaved
FDMA (IFDMA). Block FDMA allocates sets to adjacent subcarriers. Interleaved
FDMA, on the other hand, interleaves the subcarrier when defining the set.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
Acc` es multiples bas´ es OFDM
OFDM-TDMA
Les utilisateurs sont slott ´ es,
Un utilisateur utilise toute la bande pendant un ou plusieurs
symboles OFDM puis attend ` a nouveau son tour,
Appropri ´ e pour des applications ` a d´ ebits constants,
Allocation TDMA Statique : Round Robin scheduling
180 5 Principles of OFDMA
time
f
r
e
q
u
e
n
c
y
TDMA-OFDM
U1 U2 U3 U4 U5 U6 U1
t t t t t t
round robin
Fig. 5.3 TDMA-OFDM: There are 6 users (U), and TDMA scheme has fixed time intervals be-
tween frames (t) and fixed packet sizes
time
f
r
e
q
u
e
n
c
y
FDMA-OFDM
U1-U6 U1-U6 U1-U6
t t
U1-U6 U1-U6 U1-U6
t t t t
U1-U6
Blocked FDMA
Fig. 5.4 FDMA-OFDM: There are 6 users (U), and Block-FDMA scheme has fixed time intervals
between frames (t) and fixed subcarrier allocation
applied to all subcarriers. If the channel is low quality for a user in the time slot,
either transmission experiences high bit error rate with the higher modulation, or
lower modulation is selected by reducing the bit rate. Also, if there is no data to
transmit, the bandwidth is wasted.
5.1.3 Frequency Division: FDMA-OFDM
FDMA-OFDM is similar to TDMA-OFDM. FDMA-OFDM creates orthogonal re-
source by dividing the available subcarriers into fixed sets where each set is used by
a user as seen in Fig. 5.4. Fixed allocation does not change over time but the num-
ber of subcarriers associated to a user may differ. Albeit, allocation is fixed, which
prevents leveraging the multiuser diversity.
There are two types of allocation: localized FDMA (LFDMA) or interleaved
FDMA (IFDMA). Block FDMA allocates sets to adjacent subcarriers. Interleaved
FDMA, on the other hand, interleaves the subcarrier when defining the set.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en fr ´ equences :
OFDMA et SC-FDMA
Acc` es multiples bas´ es OFDM
OFDMA
Combinaison d’un acc` es FDMA et TDMA,
On alloue un “pav´ e” temps-fr ´ equence (slice, resource grids) :
N
mc
sous porteuses sur N
s
symboles OFDM cons´ ecutifs,
Diff ´ erents types d’allocation :
distribu´ ee (distributed/comb/diversity allocation), type IFDMA,
localis´ ee (localized/block/grouped/band AMC cluster), type
LFDMA.
utilis´ e pour WIMAX et 3GPP-LTE liaison descendante
182 5 Principles of OFDMA
time
f
r
e
q
u
e
n
c
y
OFDMA
t t
U1, U2, U3, U4, U6 U1, U2, U3, U4, U6 U1, U2, U3, U4, U5, U6
Fig. 5.7 OFDMA: There are 6 users (U), and OFDMA scheme has fixed distance between frames
(t) and flexible slot and subcarrier allocation
5.1.6 OFDMA
OFDMA inherits the advantages of TDMA and FDMA schemes. It also provides
flexibility as in CSMA in terms of packet sizes. User is scheduled dynamically only
if it has packet to transmit according to how much it needs to transmit.
In OFDMA, on the other hand, a transmission is packed as seen in Fig. 5.7. Frame
size is fixed as in TDMA and FDMA, but frame is shared by users. As a result,
efficiency is increased considerably and air link efficiency becomes isolated from
the traffic pattern of users. Sophisticated QoS scheduling can be applied with respect
to widely varying applications, data rates, etc.
OFDMA scheduling may exploit diversity to increase the capacity of the air link.
For instance, each user experiences different channel, as a result, a slot (subcarrier ×
time symbol) that is low in quality might not be low for others. OFDMA may exploit
multiuser diversity by assigning the subcarriers according to channel quality of each
user if channel side information is present in the transmitter. Also, modulation and
power control may be tuned according to each subcarrier with respect to channel
condition.
OFDMA scheduling may also be based on frequency diversity. An orthogonal
code may be defined for each user to determine their hopping pattern across time
and frequency. This way, fading of a user may be averaged over without any channel
information. Also, in a typical cell deployment, same frequency is used in multiple
spatially located cells. OFDMA may reduce the interference coming from the adja-
cent cells operating with the same frequency by assigning different sets of orthogo-
nal codes in each cell.
This may lead to utilize OFDMA to deploy a single frequency network in several
ways. In one way, the same frequency is used by the base stations but slot allocation
is done with respect to a code that is user specific. Codes need not to be strictly
orthogonal, which would otherwise result in low capacity system, since users are
separated in distance, which makes the probability of interference minimal. In an-
other way, some slots in a frame can be restricted to be used for the users that are not
in the overlapping area of cells. As a result, the remaining slots are used to create
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
SC-FDMA : principe g´ en´ erale
P1: OTA/XYZ P2: ABC
c03 JWBK249/Myung August 27, 2008 19:34 Printer Name: Yet to Come
Single Carrier FDMA 57
3.8 Summary
Single carrier FDMA (SC-FDMA) is a multiple access technique that uti-
lizes single carrier modulation, orthogonal frequency multiplexing, and fre-
quency domain equalization. It has similar performance and essentially the
same overall complexity as orthogonal frequency division multiple access
(OFDMA). One prominent advantage over OFDMA is that the SC-FDMA
signal has better peak power characteristics because of its inherent sin-
gle carrier structure. SC-FDMA has drawn great attention as an attractive
alternative to OFDMA, especially in the uplink communications where bet-
ter peak power characteristics greatly benefit the mobile terminal in terms
of transmit power efficiency and manufacturing cost. Because of its merits,
SC-FDMA has been chosen as the uplink multiple access scheme in 3GPP
Long Term Evolution (LTE). Also, a modified form of SC-FDMA is used
for the uplink control channel in 3GPP2 Ultra Mobile Broadband (UMB).
So why do we call this new multiple access scheme “Single Carrier”
“FDMA”? As shown in Figure 3.22, SC-FDMA symbols are transmitted
sequentially over a single carrier as opposed to the parallel transmission
of OFDM/OFDMA over multiple carriers. Also, the users are orthogonally
multiplexed and de-multiplexed in the frequency domain, which gives SC-
FDMA an aspect of FDMA.
In this chapter, we first gave an overview of SC-FDMA and explained
the transmission and reception operations in detail. SC-FDMA has two dif-
ferent approaches to subcarrier mapping: distributed and localized. In a dis-
tributed subcarrier mapping scheme, a user’s data symbols occupy a set of
subcarriers distributed over the entire frequency range of the channel and
Time
domain
Frequency
domain
Time
domain
“FDMA”
“Single Carrier” : Sequential transmission of the symbols
over a single frequency carrier.
: User multiplexing in the frequency domain.
S
u
b
c
a
r
r
i
e
r
m
a
p
p
i
n
g
M-
point
DFT
N-
point
IDFT
A
d
d

c
y
c
l
i
c

p
r
e
f
i
x
/

P
u
l
s
e

s
h
a
p
i
n
g
D
i
g
i
t
a
l
-
t
o
-
a
n
a
l
o
g
/

R
a
d
i
o

f
r
e
q
u
e
n
c
y
S
e
r
i
a
l
-
t
o
-
p
a
r
a
l
l
e
l
P
a
r
a
l
l
e
l
-
t
o
-
s
e
r
i
a
l
Figure 3.22 Why call it “Single Carrier” “FDMA”?
P1: OTA/XYZ P2: ABC
c03 JWBK249/Myung August 27, 2008 19:34 Printer Name: Yet to Come
Single Carrier FDMA 39
S
u
b
c
a
r
r
i
e
r
m
a
p
p
i
n
g
Channel
M-
point
IDFT
N-
point
DFT
Detect
M-
point
DFT
N-
point
IDFT
A
d
d

c
y
c
l
i
c

p
r
e
f
i
x
/

P
u
l
s
e

s
h
a
p
i
n
g
D
i
g
i
t
a
l
-
t
o
-
a
n
a
l
o
g
/

R
a
d
i
o

f
r
e
q
u
e
n
c
y
R
a
d
i
o

f
r
e
q
u
e
n
c
y
/

A
n
a
l
o
g
-
t
o
-
d
i
g
i
t
a
l
SC-FDMA:
OFDMA:
+
* M < N
S
e
r
i
a
l
-
t
o
-
p
a
r
a
l
l
e
l
P
a
r
a
l
l
e
l
-
t
o
-
s
e
r
i
a
l
S
e
r
i
a
l
-
t
o
-
p
a
r
a
l
l
e
l
P
a
r
a
l
l
e
l
-
t
o
-
s
e
r
i
a
l
S
u
b
c
a
r
r
i
e
r
d
e
-
m
a
p
p
i
n
g
/

E
q
u
a
l
i
z
a
t
i
o
n
R
e
m
o
v
e
c
y
c
l
i
c

p
r
e
f
i
x
Figure 3.1 Transmitter and receiver structure of SC-FDMA and OFDMA
systems
then, the SC-FDMA system can handle up to Q orthogonal source signals
with each source occupying a different set of M orthogonal subcarriers. In
the notation of Figure 3.2, x
m
(m = 0, 1, . . . , M − 1) represents modulated
source symbols and X
k
(k = 0, 1, . . . , M − 1) represents M samples of the
DFT of x
m
. Y
l
(l = 0, 1, . . . , N − 1) represents the frequency domain sam-
ples after subcarrier mapping and y
n
(n = 0, 1, . . . , N − 1) represents the
transmitted time domain channel symbols obtained from the inverse DFT
(IDFT) of Y
l
. The subcarrier mapping block in Figures 3.1 and 3.2 as-
signs frequency domain modulation symbols to subcarriers. The mapping
DFT
(M-point)
IDFT
(N-point)
Subcarrier
Mapping
{ }
m
x
{ }
k
X
{ }
n
y
{ }
l
Y
*M, N: number of data symbols
M M N N
Figure 3.2 Generation of SC-FDMA transmit symbols; there are N subcarriers
among which M (< N) subcarriers are occupied by the input data
structure ` a l’ ´ emetteur
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
SC-FDMA : architecture r ´ ecepteur
P1: OTA/XYZ P2: ABC
c03 JWBK249/Myung August 27, 2008 19:34 Printer Name: Yet to Come
42 Single Carrier FDMA
N-point
DFT
R
a
d
i
o

f
r
e
q
u
e
n
c
y
/

A
n
a
l
o
g
-
t
o
-
d
i
g
i
t
a
l
* M < N
S
u
b
c
a
r
r
i
e
r
d
e
-
m
a
p
p
i
n
g
/

U
s
e
r

s
e
p
a
r
a
t
i
o
n
R
e
m
o
v
e
c
y
c
l
i
c

p
r
e
f
i
x
Detect Equalization
Detect Equalization
M-
point
IDFT
M-
point
IDFT
M-
point
IDFT
Detect Equalization
Terminal 1
Terminal 2
Terminal Q
Terminal 1 data
Terminal 2 data
Terminal Qdata
Base station
Figure 3.4 SC-FDMA receiver structure from a multiple user access perspective
with Q terminals in the uplink
3.3 Subcarrier Mapping
Figure 3.5 shows two methods of assigning the M frequency domain modu-
lation symbols to subcarriers: distributed subcarrier mapping and localized
subcarrier mapping. In the localized subcarrier mapping mode, the mod-
ulation symbols are assigned to M adjacent subcarriers. In the distributed
mode, the symbols are equally spaced across the entire channel bandwidth.
In both modes, the IDFT in the transmitter assigns zero amplitude to the
N − M unoccupied subcarriers. We refer to the localized subcarrier map-
ping mode of SC-FDMA as localized FDMA (LFDMA) and distributed
subcarrier mapping mode of SC-FDMA as distributed FDMA (DFDMA).
The case of N =Q×M for the distributed mode with equidistance between
occupied subcarriers is referred to as Interleaved FDMA (IFDMA) [2], [3].
IFDMA is a special case of SC-FMDA and it is very efficient in that the
transmitter can modulate the signal strictly in the time domain without the
use of DFT and IDFT. In Section 3.4, we show that for IFDMA, the com-
bination of the DFT and IDFT reduces to a single complex multiplication,
equivalent to a phase rotation of each modulation symbol at the input to the
transmitter.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
Comparaison OFDMA vs SC-FDMA : structure
P1: OTA/XYZ P2: ABC
c03 JWBK249/Myung August 27, 2008 19:34 Printer Name: Yet to Come
Single Carrier FDMA 39
S
u
b
c
a
r
r
i
e
r
m
a
p
p
i
n
g
Channel
M-
point
IDFT
N-
point
DFT
Detect
M-
point
DFT
N-
point
IDFT
A
d
d

c
y
c
l
i
c

p
r
e
f
i
x
/

P
u
l
s
e

s
h
a
p
i
n
g
D
i
g
i
t
a
l
-
t
o
-
a
n
a
l
o
g
/

R
a
d
i
o

f
r
e
q
u
e
n
c
y
R
a
d
i
o

f
r
e
q
u
e
n
c
y
/

A
n
a
l
o
g
-
t
o
-
d
i
g
i
t
a
l
SC-FDMA:
OFDMA:
+
* M < N
S
e
r
i
a
l
-
t
o
-
p
a
r
a
l
l
e
l
P
a
r
a
l
l
e
l
-
t
o
-
s
e
r
i
a
l
S
e
r
i
a
l
-
t
o
-
p
a
r
a
l
l
e
l
P
a
r
a
l
l
e
l
-
t
o
-
s
e
r
i
a
l
S
u
b
c
a
r
r
i
e
r
d
e
-
m
a
p
p
i
n
g
/

E
q
u
a
l
i
z
a
t
i
o
n
R
e
m
o
v
e
c
y
c
l
i
c

p
r
e
f
i
x
Figure 3.1 Transmitter and receiver structure of SC-FDMA and OFDMA
systems
then, the SC-FDMA system can handle up to Q orthogonal source signals
with each source occupying a different set of M orthogonal subcarriers. In
the notation of Figure 3.2, x
m
(m = 0, 1, . . . , M − 1) represents modulated
source symbols and X
k
(k = 0, 1, . . . , M − 1) represents M samples of the
DFT of x
m
. Y
l
(l = 0, 1, . . . , N − 1) represents the frequency domain sam-
ples after subcarrier mapping and y
n
(n = 0, 1, . . . , N − 1) represents the
transmitted time domain channel symbols obtained from the inverse DFT
(IDFT) of Y
l
. The subcarrier mapping block in Figures 3.1 and 3.2 as-
signs frequency domain modulation symbols to subcarriers. The mapping
DFT
(M-point)
IDFT
(N-point)
Subcarrier
Mapping
{ }
m
x
{ }
k
X
{ }
n
y
{ }
l
Y
*M, N: number of data symbols
M M N N
Figure 3.2 Generation of SC-FDMA transmit symbols; there are N subcarriers
among which M (< N) subcarriers are occupied by the input data
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
SC-FDMA : allocation de sous porteuses
P1: OTA/XYZ P2: ABC
c03 JWBK249/Myung August 27, 2008 19:34 Printer Name: Yet to Come
Single Carrier FDMA 43
Localized Distributed
0
X
1 M
X

1
X
Zeros
Zeros
0
Y
1 N
Y

Zeros
0
X
Zeros
1
X
2
X
1 M
X

0
Y
1 N
Y

Zeros
Figure 3.5 Subcarrier mapping modes; distributed and localized
Figure 3.6 illustrates three examples of SC-FDMA transmit symbols in
the frequency domain for M = 4 symbols per block, N = 12 subcarriers,
and Q = N/M = 3 terminals. In the localized mode, the four modulation
symbols occupy subcarriers 0, 1, 2, and 3: Y
0
= X
0
, Y
1
= X
1
, Y
2
= X
2
,
Y
3
= X
3
, and Y
i
= 0 for i = 0, 1, 2, 3. In the distributed mode with modu-
lation symbols equally spaced over all the subcarriers, Y
0
= X
0
, Y
2
= X
1
,
Y
4
= X
2
, Y
6
= X
3
, and in the interleaved mode, Y
0
= X
0
, Y
3
= X
1
, Y
6
= X
2
,
Y
9
= X
3
.
0 0 0 0 0 0 0 0
X
0
X
1
X
2
X
3
frequency
0 0 0 0 0 0 0 0
X
0
X
1
X
2
X
3
{ }
:
k
X X
0
X
1
X
2
X
3
{ }
:
m
x x
0
x
1
x
2
x
3
DFT
2
1
0
M
j mk
M
k m
m
X x e
π


=
 
= , M = 4
 
 

0 0 0 0 0 0 0 0
X
0
X
1
X
2
X
3
{ }
l
Y
Current
implementation
in 3GPP LTE
IFDMA
DFDMA
LFDMA
Figure 3.6 An example of different subcarrier mapping schemes for M = 4,
Q = 3, and N = 12
P1: OTA/XYZ P2: ABC
c03 JWBK249/Myung August 27, 2008 19:34 Printer Name: Yet to Come
Single Carrier FDMA 43
Localized Distributed
0
X
1 M
X

1
X
Zeros
Zeros
0
Y
1 N
Y

Zeros
0
X
Zeros
1
X
2
X
1 M
X

0
Y
1 N
Y

Zeros
Figure 3.5 Subcarrier mapping modes; distributed and localized
Figure 3.6 illustrates three examples of SC-FDMA transmit symbols in
the frequency domain for M = 4 symbols per block, N = 12 subcarriers,
and Q = N/M = 3 terminals. In the localized mode, the four modulation
symbols occupy subcarriers 0, 1, 2, and 3: Y
0
= X
0
, Y
1
= X
1
, Y
2
= X
2
,
Y
3
= X
3
, and Y
i
= 0 for i = 0, 1, 2, 3. In the distributed mode with modu-
lation symbols equally spaced over all the subcarriers, Y
0
= X
0
, Y
2
= X
1
,
Y
4
= X
2
, Y
6
= X
3
, and in the interleaved mode, Y
0
= X
0
, Y
3
= X
1
, Y
6
= X
2
,
Y
9
= X
3
.
0 0 0 0 0 0 0 0
X
0
X
1
X
2
X
3
frequency
0 0 0 0 0 0 0 0
X
0
X
1
X
2
X
3
{ }:
k
X X
0
X
1
X
2
X
3
{ }:
m
x x
0
x
1
x
2
x
3
DFT
2
1
0
M
j mk
M
k m
m
X x e
π


=
 
= , M = 4
 
 

0 0 0 0 0 0 0 0
X
0
X
1
X
2
X
3
{ }
l
Y
Current
implementation
in 3GPP LTE
IFDMA
DFDMA
LFDMA
Figure 3.6 An example of different subcarrier mapping schemes for M = 4,
Q = 3, and N = 12
P1: OTA/XYZ P2: ABC
c03 JWBK249/Myung August 27, 2008 19:34 Printer Name: Yet to Come
44 Single Carrier FDMA
subcarriers
Terminal 1
Terminal 2
Terminal 3
subcarriers
Interleaved (distributed) Mode Localized Mode
Figure 3.7 Subcarrier allocation methods for multiple users (3 users, 12 subcar-
riers, and 4 subcarriers allocated per user)
Figure 3.7 shows IFDMA and LFDMA demonstrating that the signals
of the three different terminals arriving at a base station occupy mutually
exclusive sets of subcarriers.
From a resource allocation point of view, subcarrier mapping methods
are further divided into static scheduling and channel-dependent schedul-
ing (CDS) methods. CDS assigns subcarriers to users according to the
channel frequency response of each user. For both scheduling methods,
distributed subcarrier mapping provides frequency diversity because the
transmitted signal is spread over the entire bandwidth. With distributed
mapping, CDS incrementally improves performance. By contrast, CDS is
of great benefit with localized subcarrier mapping because it provides sig-
nificant multi-user diversity. Chapter 5 contains a detailed description of
channel dependent scheduling.
3.4 Time Domain Representation of SC-FDMA Signals
For IFDMA, LFDMA, and DFDMA, the three operations in Figure 3.2 can
be viewed as one linear operation on the sequence of modulation symbols
{x
m
: m = 0, 1, 2, . . . , M − 1}. Therefore each element of the output se-
quence {y
n
: n = 0, 1, 2, . . . , N − 1} is a weighted sum of the elements of
the input sequence, where the weights are complex numbers. In the case of
IFDMA, the weights are zero for all but one element of the input sequence.
The set of operations reduces to multiplying each input symbol by a com-
plex number with unit magnitude and repeating the input sequence with
proper phase rotation Q times, where Q is the bandwidth expansion factor.
Section 3.4.1 derives this property for the example M = 4 symbols
per block, N = 12 subcarriers and Q = 3 terminals. Section 3.4.1 also
shows graphically the spectrum and time sequence of an IFDMA signal
block. The formulas for the time sequences of LFDMA and DFDMA are
more complicated than for IFDMA. Sections 3.4.2 and 3.4.3 show the
spectrum of each of these two subcarrier mapping techniques along with
P1: OTA/XYZ P2: ABC
c03 JWBK249/Myung August 27, 2008 19:34 Printer Name: Yet to Come
44 Single Carrier FDMA
subcarriers
Terminal 1
Terminal 2
Terminal 3
subcarriers
Interleaved (distributed) Mode Localized Mode
Figure 3.7 Subcarrier allocation methods for multiple users (3 users, 12 subcar-
riers, and 4 subcarriers allocated per user)
Figure 3.7 shows IFDMA and LFDMA demonstrating that the signals
of the three different terminals arriving at a base station occupy mutually
exclusive sets of subcarriers.
From a resource allocation point of view, subcarrier mapping methods
are further divided into static scheduling and channel-dependent schedul-
ing (CDS) methods. CDS assigns subcarriers to users according to the
channel frequency response of each user. For both scheduling methods,
distributed subcarrier mapping provides frequency diversity because the
transmitted signal is spread over the entire bandwidth. With distributed
mapping, CDS incrementally improves performance. By contrast, CDS is
of great benefit with localized subcarrier mapping because it provides sig-
nificant multi-user diversity. Chapter 5 contains a detailed description of
channel dependent scheduling.
3.4 Time Domain Representation of SC-FDMA Signals
For IFDMA, LFDMA, and DFDMA, the three operations in Figure 3.2 can
be viewed as one linear operation on the sequence of modulation symbols
{x
m
: m = 0, 1, 2, . . . , M − 1}. Therefore each element of the output se-
quence {y
n
: n = 0, 1, 2, . . . , N − 1} is a weighted sum of the elements of
the input sequence, where the weights are complex numbers. In the case of
IFDMA, the weights are zero for all but one element of the input sequence.
The set of operations reduces to multiplying each input symbol by a com-
plex number with unit magnitude and repeating the input sequence with
proper phase rotation Q times, where Q is the bandwidth expansion factor.
Section 3.4.1 derives this property for the example M = 4 symbols
per block, N = 12 subcarriers and Q = 3 terminals. Section 3.4.1 also
shows graphically the spectrum and time sequence of an IFDMA signal
block. The formulas for the time sequences of LFDMA and DFDMA are
more complicated than for IFDMA. Sections 3.4.2 and 3.4.3 show the
spectrum of each of these two subcarrier mapping techniques along with
P1: OTA/XYZ P2: ABC
c03 JWBK249/Myung August 27, 2008 19:34 Printer Name: Yet to Come
44 Single Carrier FDMA
subcarriers
Terminal 1
Terminal 2
Terminal 3
subcarriers
Interleaved (distributed) Mode Localized Mode
Figure 3.7 Subcarrier allocation methods for multiple users (3 users, 12 subcar-
riers, and 4 subcarriers allocated per user)
Figure 3.7 shows IFDMA and LFDMA demonstrating that the signals
of the three different terminals arriving at a base station occupy mutually
exclusive sets of subcarriers.
From a resource allocation point of view, subcarrier mapping methods
are further divided into static scheduling and channel-dependent schedul-
ing (CDS) methods. CDS assigns subcarriers to users according to the
channel frequency response of each user. For both scheduling methods,
distributed subcarrier mapping provides frequency diversity because the
transmitted signal is spread over the entire bandwidth. With distributed
mapping, CDS incrementally improves performance. By contrast, CDS is
of great benefit with localized subcarrier mapping because it provides sig-
nificant multi-user diversity. Chapter 5 contains a detailed description of
channel dependent scheduling.
3.4 Time Domain Representation of SC-FDMA Signals
For IFDMA, LFDMA, and DFDMA, the three operations in Figure 3.2 can
be viewed as one linear operation on the sequence of modulation symbols
{x
m
: m = 0, 1, 2, . . . , M − 1}. Therefore each element of the output se-
quence {y
n
: n = 0, 1, 2, . . . , N − 1} is a weighted sum of the elements of
the input sequence, where the weights are complex numbers. In the case of
IFDMA, the weights are zero for all but one element of the input sequence.
The set of operations reduces to multiplying each input symbol by a com-
plex number with unit magnitude and repeating the input sequence with
proper phase rotation Q times, where Q is the bandwidth expansion factor.
Section 3.4.1 derives this property for the example M = 4 symbols
per block, N = 12 subcarriers and Q = 3 terminals. Section 3.4.1 also
shows graphically the spectrum and time sequence of an IFDMA signal
block. The formulas for the time sequences of LFDMA and DFDMA are
more complicated than for IFDMA. Sections 3.4.2 and 3.4.3 show the
spectrum of each of these two subcarrier mapping techniques along with
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
SC-FDMA : allocation de sous-porteuses
P1: OTA/XYZ P2: ABC
c03 JWBK249/Myung August 27, 2008 19:34 Printer Name: Yet to Come
Single Carrier FDMA 43
Localized Distributed
0
X
1 M
X

1
X
Zeros
Zeros
0
Y
1 N
Y

Zeros
0
X
Zeros
1
X
2
X
1 M
X

0
Y
1 N
Y

Zeros
Figure 3.5 Subcarrier mapping modes; distributed and localized
Figure 3.6 illustrates three examples of SC-FDMA transmit symbols in
the frequency domain for M = 4 symbols per block, N = 12 subcarriers,
and Q = N/M = 3 terminals. In the localized mode, the four modulation
symbols occupy subcarriers 0, 1, 2, and 3: Y
0
= X
0
, Y
1
= X
1
, Y
2
= X
2
,
Y
3
= X
3
, and Y
i
= 0 for i = 0, 1, 2, 3. In the distributed mode with modu-
lation symbols equally spaced over all the subcarriers, Y
0
= X
0
, Y
2
= X
1
,
Y
4
= X
2
, Y
6
= X
3
, and in the interleaved mode, Y
0
= X
0
, Y
3
= X
1
, Y
6
= X
2
,
Y
9
= X
3
.
0 0 0 0 0 0 0 0
X
0
X
1
X
2
X
3
frequency
0 0 0 0 0 0 0 0
X
0
X
1
X
2
X
3
{ }
:
k
X X
0
X
1
X
2
X
3
{ }
:
m
x
x
0
x
1
x
2
x
3
DFT
2
1
0
M
j mk
M
k m
m
X x e
π


=
 
= , M = 4
 
 

0 0 0 0 0 0 0 0
X
0
X
1
X
2
X
3
{ }
l
Y
Current
implementation
in 3GPP LTE
IFDMA
DFDMA
LFDMA
Figure 3.6 An example of different subcarrier mapping schemes for M = 4,
Q = 3, and N = 12
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
Dynamique des signaux : SC-FDMA vs OFDMA
P1: OTA/XYZ P2: ABC
c03 JWBK249/Myung August 27, 2008 19:34 Printer Name: Yet to Come
50 Single Carrier FDMA
0 10 20 30 40 50 60
0
0.2
0.4
0.6
0 10 20 30 40 50 60
0
0.2
0.4
0.6
0 10 20 30 40 50 60
0
0.2
0.4
0.6
0 10 20 30 40 50 60
0
0.2
0.4
0.6
Sample [time]
A
m
p
l
i
t
u
d
e

[
l
i
n
e
a
r
]
IFDMA
LFDMA
DFDMA
OFDMA
Figure 3.12 Amplitude of SC-FDMA and OFDMA samples for QPSK input
modulation without pulse shaping with N = 64 subcarriers, M = 16 subcarriers
per block, Q = 4 spreading factor for IFDMA with four terminals, and
˜
Q = 3
spreading factor for DFDMA with three terminals
Overall, all three single carrier subcarrier mapping schemes exhibit lower
peak power than OFDMA.
3.5 SC-FDMA and Orthogonal Frequency Division
Multiple Access
Section 2.4 describes the relationship between single carrier transmis-
sion and OFDM for the transmission of one signal occupying the entire
system bandwidth. There is a similar relationship between OFDMA and
SC-FDMA for the transmission of independent signals from dispersed
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Acc` es multiples par R´ epartition en
fr ´ equences :OFDMA et SC-FDMA
SC-FDMA vs OFDMA
SC-FDMA :
PAPR faible,
Moins sensible aux offsets de fr ´ equence,
Robustesse face aux ´ evanoussements fr ´ equentiels,
bit-loading non possible,
OFDMA :
PAPR plus important,
Plus sensible aux offsets de fr ´ equence,
Sensible face aux ´ evanouissements fr ´ equentiels,
bit-loading possible
⇒SC-FDMA plus adapt ´ e pour la liaison montante
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
M´ ecanismes d’adaptation et de diversit ´ e
Principes g´ en´ eraux
Principe
Adapter de mani ` ere dynamique les param` etres de transmission
pour s’adapter aux variations temporelles du canal,
Exploiter la diversit ´ e due aux variations du canal
Principaux param` etres d’adaptation
puissance d’ ´ emission,
taille de constellation,
type de codage canal et rendement associ ´ e,
Principales techniques de diversit ´ e
Techniques de diversit ´ e temporelle, fr ´ equentielle ou spatiale :
OFDMA/SC-FDMA, MIMO, coop´ eration,
Codage de canal et m´ ecanismes de retransmission (H-ARQ),
Diversit ´ e multi-utilisateurs : m´ ecanismes de scheduling, s´ election
de relais
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
M´ ecanismes d’adaptation et de diversit ´ e
Principes g´ en´ eraux
Notions de QoS
Adaptation et m´ ecanismes de diversit ´ e visent ` a garantir une QoS
aux couches OSI sup´ erieures.
Diff ´ erents types de contraintes peuvent ˆ etre impos´ ees pour
garantir la QoS souhait ´ ee : d´ elai, Taux d’erreur paquet r ´ esiduel
lors de la pr ´ esentation aux couches sup´ erieures,
en g´ en´ erale, on traduit les contraintes QoS comme une
contriante de taux d’erreur paquet sur la couche physique
Qui g` ere quoi ?
Couche physique : g` ere l’ensemble des ´ el ´ ements de traitements
physiques des donn´ ees ` a transmettre (puissance, modulations et
codage) en fonction des ´ el ´ ements fournis par la couche MAC,
Couche liaison donn´ ee (MAC layer) : g` ere l’ensemble des
´ el ´ ements de traitements logiques et d´ ecisionnels sur les
donn´ ees ` a transmettre, ie. gestion des m´ ecanismes et
param` etres d’adaptation et allocation (diversit ´ e multi-utilisateurs)
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
M´ ecanismes d’adaptation
Techniques d’adaptation
Adaptive Modulation and coding (AMC),
Hybrid Automatic Repeat reQuest (HARQ),
Allocation/contr ˆ ole de puissance,
Allocation de porteuses/scheduling des utilisateurs,
contr ˆ ole d’admission
Hypoth` eses syst ` eme
Connaissance a priori des caract ´ eristiques du canal ` a l’ ´ emetteur
(Channel Side Infomation),
Voie de retour pour que l’ ´ emetteur informe le r ´ ecepteur,
Donn´ ees de contr ˆ ole et de signalisation pour informer le
r ´ ecepteur des m´ ecanismes utilis´ es ` a l’ ´ emission,
M´ ecanismes de pr ´ ediction pour traquer les variations du canal,
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Adaptation de lien (link adaptation) par AMC
Principe
Eviter de surdimensionner un syst ` eme par consid´ eration du pire
cas,
Utiliser un lien physique de la mani ` ere la plus efficace qui soit de
mani ` ere instantan´ ee,
Crit ` ere d’optimisation possible : maximisation du d´ ebit instantan´ e
sous contraintes de taux d’erreur paquet cible en r ´ eception,
M´ ecanismes AMC pour adaptation de lien
Principe : adapter de mani ` ere conjointe ` a l’ ´ emetteur le couple
codeur canal + modulation afin de garantir un taux d’erreur
paquet cible apr ` es d´ ecodage pour les couches plus hautes,
Une combinaison (code+modulation) est appel ´ ee Modulation
and Coding Scheme, MCS
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Adaptation de lien (link adaptation) par AMC
Crit ` ere d’optimisation
throughput : η = R log
2
(M)(1 −PER)
avec
PER : le taux d’erreurs paquets en sortie de d´ ecodage canal,
R le rendement du code canal,
log
2
(M) le nombre de bit par symbole modul ´ e,
le m´ ecanisme d’AMC s´ electionne le MCS qui maximise le d´ ebit
instantan´ e sous certaines contraintes de QoS
Si la QoS est donn´ ee par un PER cible en sortie de codage
canal alors le probl ` eme de d´ ecision est un probl ` eme ` a seuils
(swithching thresholds) qui d´ eterminent les rapport signaux ` a
bruit pour lesquels on utilise le MCS(k).
Marges ` a consid´ erer pour robustesse aux erreurs d’estimation.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Adaptation de lien (link adaptation) par AMC
Switching thresholds
As mentioned above, SIR is a common criterion for MCS
switching. Evaluations of suitable thresholds for various
channel conditions have been performed through simulation
[6]. The threshold values are function of channel conditions
and depend on which particular MCS set is used. In an ideal
situation, a system knows the channel condition and can
predicted the throughput curve like Fig. 1, the system can
decide the optimal thresholds for that particular channel
condition. However, it is not a simple task to predict the
channel condition, particular in a real operating environment,
where the channel condition is time varying (e.g., due to
user's location). Therefore, it is difficult to set optimum
threshold. If the thresholds are set incorrectly, the
throughput performance will degrade as shown in Fig. 2. In
Fig. 2, SIR threshold between MCS 1 and MCS2 is set lower
than the optimum value compared with Fig. 1. On the other
hand, threshold between MCS(2) and MCS(3) is set higher
than the optimum value.
A
MCS(1)
-.-.
SIR threshold between
MCS(2) and MCS(3)
.
t'
SIR threshold between SIR
MCS(I) and MCS(2)
Fig. 1 Throughput curve and optimum thresholds
A
MCS(I)
SIR threshold between
SIR threshold between SIR
MCS(1) and MCS(2)
Fig. 2 Throughput c we and incorrect threshold
111. PROPOSED THRESHOLD CONTROL
A. Threshold Controlling Scheme According to CRC results
In order to solve the above problem, an adaptive SIR
threshold control scheme is proposed in [7] where the SIR
thresholds are controlled according to CRC results shown in
the following:
CRC OK : SIR threshold is decreased by 6dosw
CRC Error : SIR threshold is increased by 4p
For example, if the target BLER is set to I%, the
corresponding values for 6,, and b;,p are O.OldB (=target
BLER in dB) and 0.99dB (=1.0 - target BLER in dB). In
terms of throughput, BLER of 1% is equivalent to 99% of
the maximum throughput and hence 1 CRC Error and 99
CRC OK is a suitable operating point for SIR threshold
control. When the above control scheme is applied, the key
requirement is to decide a target BLER.
B. Proposed Threshold Controlling Scheme
In this proposal, two different thresholds are used for each
MCS(n), where n = 1, 2, ..., N, with N represents the total
number of MCSs available in the AMC system. The two
thresholds are called the upper threshold T(n) which is used
to switch MCS(n) to MCS(n+l) and the lower threshold
T(n-I) which is used to switch MCS(n) to MCS(n-1). After
the system switched from MCS(n) to MCS(n+l), T(n)
becomes the lower threshold for switching MCS(n+l) to
MCS(n). Also note that, the lower threshold T(0) should not
be defined when MCS(1) is selected and the upper threshold
T(N) should not he defined when MCS(N) is selected.
According to the above definition, the proposed SIR
control can be summarised as follows.
F If CRC OK, then T(n) is decreased by
s,,,, = 0.01 [dB1 Eq. 1
If CRC Error, then T(n) is increased by
=0.99 [dB] Eq.2
F
The SIR threshold T(n-I) is used for switch MCS(n) to
MCS(n-1) is controlled as follows.
P If CRC OK, then T(n-I) is decreased by
[dB] Eq.3
Mar- thpt(n - 1)
Max- thpt(n)
'daw =
b I f CRC Error, then T(n-I) is increased by
[@I Eq.4
Max - thpt(n - 1)
6 =1.0-
UP Mar-thpf(n)
Where, Ma.-thpt(n) represents the maximum throughput of
MCS(n). The following is the detail explanation of the
proposed threshold control.
1352
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Adaptation de lien (link adaptation) par AMC
M´ ethodologie
A l’ ´ emetteur, seule la statistique du canal bas´ e sur un SINR
“ ´ equivalent” est n´ ecessaire pour choisir le MCS.
PER = f (MCS, SINR, N)
o` u N est la taille du mot de code.
En pratique, seul l’index du MCS est fourni. Ce dernier est
calcul ´ e au r ´ ecepteur en fonction des estimations du canal
disponible. Cet index est appel ´ e CQI (channel quality index),
Ensuite, d´ eterminer une relation entre CQI = g(h, N), o` u h est
un vecteur qui d´ epend du canal en r ´ eception.
Cette relation est d´ eterminer en utilisant une mesure de qualit ´ e
du lien (Link Quality Metric, LQI). On a alors la relation
PER = f (MCS, LQI, N)
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Adaptation de lien (link adaptation) par AMC
Link Quality Metrics
Effective SINR Mapping
Le SINR (Signal to Interf ´ erence plus Noise Ratio) effectif est
donn´ e par
SNR
eff
= α
1
I
−1

1
N
N

n=1
I

SINR
k
α
2

o` u I(.) est une fonction qui lie le SINR ` a une mesure LQI,
Mesures LQI possibles :
Identit ´ e (SINR moyen, (LESM)) : I(γ) = γ,
Capacit ´ e de Shannon (CESM) : I(γ) = log
2
(1 + γ),
EESM (Exponential Effective SINR mapping) : I(γ) = e
−γ
,
MIESM (Mutual Information Equivalent SINR Mapping) :
I(γ) = log
2
(M) +
1
2πM
M−1

m=1

e
−γ(y−x
m
)
2
log
2
(
e
−γ(y−x
m
)
2

M−1
m=1
e
−γ(y−x
k
)
2
)dy
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Adaptation de lien (link adaptation) par AMC
Link Quality Metrics : performances
2005IEEE 16th International
Symposium
on Personal, Indoor and Mobile Radio Communications
TABLEI
MODULATION AND CODING SCHEMES
IMCS
coding
Imodulationj I
code w ord
leng thi
I1 CC1I2~
Q PSK 4 08
2 CC1/ 2
16- Q AM
8 24
1
CC
3 / 4
16- Q AM
124 0
TC1/ 3 1
16- Q AM
54
TABLE
PERFORMANCEFIGURES OFCONSIDERED EFFECTIVESINRAPPROACHES
109
cc
10
10
100
0-
10
LESM
...
a,IE
... ....
.. .. .. ... ....
..
.. .. .. .. ..
..[
.. .. ..
CESM
............
................................................
.
I 0
................................................
0 2
SINR
[ dB]
MIESM
lo'
I o-
...............................................
i o
.........................
0 2 3
SINR
elf
[ dB]
4
Fig . 4 . Results f or MCS
using
SCM channel model
cr
w
a.
a:
w
( L
The available results f or a
sing le
receive and transmit
antenna are summarized in Table In order to illustrate the
ef f iciency
of the dif f erent ef f ective SINR
approaches, Fig .
4 - 6
exemplarily
show the
output
of the
individually
trained models
in ten- ns of measured i.e. actual PERversus ef f ective SINR f or
MCS 1, MCS 3 and MCS 4 ,
respectively.
Dif f erent channel
model
parameters
are ref lected
by
dif f erent colors of the data
points using mag enta
f or "suburban macro", blue f or "urban
macro' and
g reen
f or "urban micro",
respectively. Considering
an AWGN based
mapping
here
( depicted by
the solid
line) ,
the
match of the data
points
w ith this curve reveals the
accuracy
of the
respective
link
perf ormance
model.
The results in Table as w ell as the
corresponding plots
indicate an excellent PER
prediction accuracy
f or the MIESM
approach
on all tested conditions.
Only
in case of turbo
coding
the
root- mean- square
error values in Table 6 reveal a
slig htly
superior perf ormance
of CESM, how ever,
having
a look at
Fig .
6 the dif f erence cannot be
reg arded
as
sig nif icant.
On other
conditions
especially
at
hig h
code rates CESM as w ell
as LESM
perf orm
considerable w orse than MIESM and also
EESM. At those code rates w hich coincide w ith a
relatively
hig h
SINR
operating point
f or the considered modulation
scheme the amount of "inf ormation" carried
by
at least some
of the less f aded resource elements
appears
to be overrated.
9 10 1 1 12 13 14 15
SINR,
1f dB]
9 10 1 1 12 13 14 15
SINR.f f [ dBI
Fig .
5. Results f or MCS 3
using
SCMEchannel model
Using
f unctions
I( .)
w hich saturate at
hig h
SINR values, as
f or MIESM and
EESM, seem to model the limitations of
the
applied
modulation more
realistically. Apart
f rom MCS
4 , MIESM and EESM show a
comparable perf ormnance
w ith
respect
to the correlation of the data
points.
How ever dif f erent
f rom
MIESM, EESM obtains a
g ood
match w ith the AWGN
curve
only
at
hig her
error rates. This
potential
draw back
may
be overcome
by using
a
polynomial
based
mapping
f unction
as indicated
by
a
comparison
of
o0log
in Table
Another
arg ument
in f avor of MIESMis its inherent
support
f or
adaptive
modulation. This is
simply
achieved
by sw itching
the mutual- inf ormation curve
according
to the actual modula-
tion scheme
applied
f or
any
resource element
p. Corresponding
results are
presented
in the lef t
g raph
of
Fig .
7
indicating
9 78 - 3 - 8 007- 29 09 - 8 / 05/ $20.00 ©2005IEEE
[
AWGN
based polynomial bas~ ed
]
mapping mapping
MCS model
/ 3 opt
CT 6109
! 3 opt
j
l
Tlog
1 LESM I[ - 0.2- 7 0.04 2 0.268 - 0.28 0.03 1 0.08 9
CESM 0.09 0.04 3 0.14 4 02,9
002 006
EESM IL 6 0j j 120O 0.08 4 166 0j 013 0J 02
MIESM I.1 0j j 016 0.03 1.11 0.013 O002
2 LESM if - 0.18 0.03 2 0.163 - 0.21 01125 0I06
CESM If 0.21 0.03 2 0.09 9 0.61 f t01j j j .j 5j
EESM 5f 25 0. 014 8 5 120019 O004
IMIESM Hii
4 001 .3 1.15 1.9 9
1
0.024
3 LESM I.0.3 7 0.03 8 0. 3 52
CESM Ii0.00 0.123 0.74 0 not
EESM 7.4 7 0.027 0.073 evaluated
MIESM 1.07 0.013 0.028
4 LESM - 0.3 6 0. 012 0.026
CESM 0.9 2 0.009 0.015 not
EESM .24 0.020 0.04 evaluated
MIESM I1.02 0.010 0.018
a
Li
a
a:
w
a-
cr
ui
a.
23 09
Authorized licensed use limited to: Charly Poulliat. Downloaded on May 10,2010 at 09:48:27 UTC from IEEE Xplore. Restrictions apply.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Adaptation de lien (link adaptation) par AMC
AMC pour LTE
210 LTE – THE UMTS LONG TERM EVOLUTION
0 5 10 15 20 25
10
−2
10
−1
SNR
B
L
E
R
QPSK, r=1/3
QPSK, r=1/2
QPSK, r=2/3
QPSK, r=4/5
16QAM, r=1/3
16QAM, r=1/2
16QAM, r=2/3
16QAM, r=4/5
64QAM, r=1/3
64QAM, r=1/2
64QAM, r=2/3
64QAM, r=4/5
Figure 10.1 Typical BLER versus Signal-to-Noise Ratio (SNR) for different modulation and
coding schemes. From left to right, the curves in this example correspond to QPSK, 16QAM
and 64QAM, rates 1/3, 1/2, 2/3 and 4/5.
Table 10.1 CQI table. Reproduced by permission of © 3GPP.
Efficiency
CQI index Modulation Approximate code rate (information bits per symbol)
0 No transmission — —
1 QPSK 0.076 0.1523
2 QPSK 0.12 0.2344
3 QPSK 0.19 0.3770
4 QPSK 0.3 0.6016
5 QPSK 0.44 0.8770
6 QPSK 0.59 1.1758
7 16QAM 0.37 1.4766
8 16QAM 0.48 1.9141
9 16QAM 0.6 2.4063
10 64QAM 0.45 2.7305
11 64QAM 0.55 3.3223
12 64QAM 0.65 3.9023
13 64QAM 0.75 4.5234
14 64QAM 0.85 5.1152
15 64QAM 0.93 5.5547
AMC can exploit the UE feedback by assuming that the channel fading is sufficiently
slow. This requires the channel coherence time to be at least as long as the time between
the UE’s measurement of the downlink reference signals and the subframe containing the
210 LTE – THE UMTS LONG TERM EVOLUTION
0 5 10 15 20 25
10
−2
10
−1
SNR
B
L
E
R
QPSK, r=1/3
QPSK, r=1/2
QPSK, r=2/3
QPSK, r=4/5
16QAM, r=1/3
16QAM, r=1/2
16QAM, r=2/3
16QAM, r=4/5
64QAM, r=1/3
64QAM, r=1/2
64QAM, r=2/3
64QAM, r=4/5
Figure 10.1 Typical BLER versus Signal-to-Noise Ratio (SNR) for different modulation and
coding schemes. From left to right, the curves in this example correspond to QPSK, 16QAM
and 64QAM, rates 1/3, 1/2, 2/3 and 4/5.
Table 10.1 CQI table. Reproduced by permission of © 3GPP.
Efficiency
CQI index Modulation Approximate code rate (information bits per symbol)
0 No transmission — —
1 QPSK 0.076 0.1523
2 QPSK 0.12 0.2344
3 QPSK 0.19 0.3770
4 QPSK 0.3 0.6016
5 QPSK 0.44 0.8770
6 QPSK 0.59 1.1758
7 16QAM 0.37 1.4766
8 16QAM 0.48 1.9141
9 16QAM 0.6 2.4063
10 64QAM 0.45 2.7305
11 64QAM 0.55 3.3223
12 64QAM 0.65 3.9023
13 64QAM 0.75 4.5234
14 64QAM 0.85 5.1152
15 64QAM 0.93 5.5547
AMC can exploit the UE feedback by assuming that the channel fading is sufficiently
slow. This requires the channel coherence time to be at least as long as the time between
the UE’s measurement of the downlink reference signals and the subframe containing the
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Contr ˆ ole d’erreur
M´ ecanismes HARQ
Error Control Techniques
3 / 46
17/12/2009
Error control techniques
Hybrid ARQ (HARQ) protocols
Forward error correcting codes
• Convolutional codes, turbo-codes, LDPC codes,
...
• No need for a feedback channel.
• Fixed rate.
• Efficient for bad channel conditions.
Automatic Repeat reQuest (ARQ)
protocols
• The same message is retransmitted.
• Need for a feed back channel.
• variable rate.
• Efficient for good channel conditions.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Contr ˆ ole d’erreur
M´ ecanismes HARQ
Hybrid Automatic Repeat reQuest (HARQ)
4 / 46
17/12/2009
Feedback channel ACK / NACK (error-free)
Forward Channel
Receiver Transmitter
Type-I Parity Message CRC Parity Message CRC
Chase combining (CC-HARQ)
Type-II
Parity 1
Message CRC
Parity 2
Parity Message CRC
RCPC : Rate Compatible Punctured Codes
Parity Message CRC
Parity Message CRC
Parity Message CRC
Parity 1
Message CRC
Parity 2
Parity 3
Incremental redundancy (IR-HARQ)
Tx1
Tx2
Tx3
Tx2
Tx3
Tx1
signal
combining
code
combining
Parity 1
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Contr ˆ ole d’erreur
Protocoles Stop-and-Wait
fails to detect the presence of errors or fails to determine the
exact locations of the errors. In either case, an erroneous
word is delivered to the user. Since no retransmission is
required in an FEC error-control system, no feedback channel
is needed. The throughput of the system is constant, and is
equal to the rate of the code used by the system.
FEC error-control systems, however, do have some draw-
backs. When a received word is detected in error, it must be
decoded, and the decoded word must be delivered to the user
regardless of whether it is correct or incorrect. Since the
probability of a decoding error is much greater than the
probability of an undetected error, it is harder to achieve
high system reliability with FEC schemes. In order to attain
high reliability, a long, powerful error-correcting code must
be used and a large collection of error patterns must be
corrected. This makes decoding hard to implement and
expensive. For these reasons, ARQ schemes are often
preferred over FEC schemes for error control in data
communications systems such as packet-switching data
networks and computer communications networks. However,
in communications (or data storage) systems where feed-
back channels are not available or retransmission is not
suitable for some reason, FEC is the only choice.
This paper surveys a number of ARQ schemes. They
represent alternative solutions to the design of retrans-
mission protocols, particularly the mode in which the
transmitter stores, orders, and retransmits the codewords
which have been received in error. These different schemes
have arisen primarily in an attempt to combat the problem
that, when the channel error rate increases, the throughput
of an ARQ error-control system may deteriorate very rapidly.
This is caused by the time wasted in retransmitting the
codewords detected in error. This problem becomes particu-
larly severe if there is significant round-trip delay between
the transmission of a codeword and the receipt of its error
status information back at the transmitter. Long delay is
inevitable when satellites or long terrestrial channels are
being used.
Another approach to error control is through the use of
hybrid ARQ schemes which incorporate both FEC and retrans-
mission. Hybrid ARQ schemes offer the potential for better
performance if appropriate ARQ and FEC schemes are
properly combined. Either block or convolutional codes may
be used for FEC. This paper also discusses different classes
of hybrid ARQ schemes and their performance.
Basi c ARQ Schemes
Based on retransmission strategies, there are three basic
types of ARQ schemes: stop-and-wait ARQ, go-back4 ARQ,
and selective-repeat ARQ [3, 16-21]. The st op- and- wai t
scheme represents the simplest ARQ procedure and was
implemented in early error-control systems. The IBM Binary
Synchronous Communication (BISYNC) procedure, for ex-
ample, was of the stop-and-wait type [22]. In a stop-and-wait
ARQ error-control system, the transmitter sends a codeword
to the receiver and waits for an acknowledgment, as shown
in Fig. 1. A positive acknowledgment (ACK) from the receiver
indicates that the transmitted codeword has been success-
fully received, and the transmitter sends the next codeword
in the queue. A negative acknowledgment (NAK) from the
receiver indicates that the transmitted codeword has been
detected in error; the transmitter then resends the codeword
and again waits for an acknowledgment. Retransmissions
continue until the transmitter receives an ACK.
This scheme is simple but inherently inefficient because of
the idle time spent waiting for an acknowledgment of each
transmitted codeword. One possible remedy is to make the
block (or code) length n extremely long. However, the use of
a very long block length does not really provide a solution,
since the probability that a block contains errors increases
with the block length. Hence, using a long block length
reduces the idle time but increases the frequency of
retransmissions for each codeword. Moreover, a long block
length may be impractical in many applications because of
restrictions imposed by the data format.
By the 1970's, ARQ systems were in extensive use in
packet-switched and other data networks. Higher data rates
and utilization of satellite channels with long round-trip
delays established the need for continuous transmission
strategies to replace the stop-and-wait procedures. Inter-
national standards organizations such as CCITT (the Inter-
national Telegraph and Telephone Consultative Committee)
began making efforts for protocol standardization. This
resulted in the high-level data link. control (HDLC) and the
CCITT X.25 standards. These envisaged the use of a go-back-
December 1984-Vo1.22, No. 1 2
IEEE Communications Magazine 6
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Contr ˆ ole d’erreur
Protocoles Go-Back-N
N ARQ system on full duplex links. This remains the standard
for packet-switching networks.
The basic go-back-N ARQ scheme is illustrated in Fig. 2.
The transmitter continuously transmits codewords in order
and then stores them pending receipt of an ACKlNAK for
each. The acknowledgment for a codeword arrives after a
round-trip delay, defined as the time interval between the
transmission of a codeword and the receipt of an acknowl-
edgment for that codeword. During this interval, N - 1 other
codewords are also transmitted. Whenever the transmitter
receives a NAK indicating that a particular codeword, say
codeword i , was received in error, it stops transmitting new
codewords. Then it goes back to codeword i and proceeds to
retransmit that codeword and the N - 1 succeeding
codewords which were transmitted during one round-trip
delay. At the receiving end, the receiver discards the
erroneously received word i and all N - 1 subsequently
received words, whether they are error-free or not. Retrans-
mission continues until codeword i is positively acknowl-
edged. In each retransmission for codeword i , the transmitter
resends the same sequence of codewords. As soon as
codeword i is positively acknowledged, the transmitter
proceeds to transmit new codewords.
The main drawback of go-back-N ARQ is that, whenever a
received word is detected in error, the receiver also rejects
the next N - 1 received words, even though many of them
may be error free. As a result, they must be retransmitted.
This represents a waste of transmissions, which can result in
severe deterioration of throughput performance if large
round-trip delay is involved. For example, consider a
satellite channel with a round-trip delay of approximately
700 ms. If the codeword length n is 1000 bits long and the bit
rate is 1 Mb/s, then in one round-trip delay, N =700
codewords are transmitted. Therefore, when one received
word is detected in error, 700 received words are rejected. If
errors occur often enough, the system throughput may fall
off very rapidly.
The go-back4 ARQ scheme becomes quite ineffective for
communications systems with high data rates and large
round-trip delays. This ineffectiveness is caused by the
retransmission of many error-free codewords following a
codeword detected in error. This can be overcome by using
the selective-repeat ARQ protocol. In a selective-repeat ARQ
error-control system, codewords are also transmitted con-
7
tinuously. However, the transmitter only resends those
codewords that are negatively acknowledged (NAK’ed). After
resending a NAK’ed codeword, the transmitter continues
transmitting new codewords in the transmitter buffer (as
illustrated in Fig. 3). With this scheme, a buffer must be
provided at the receiver to store the error-free codewords
following a received word detected in error, because,
ordinarily, codewords must be delivered to the end user in
correct order, for example, in point-to-point communications.
When the first NAK’ed codeword is successfully received, the
receiver then releases any error-free codewords in consecu-
tive order from the receiver buffer until the next erroneously
received word is encountered. Sufficient receiver buffer
storage must be provided in a selective-repeat ARQ system;
otherwise, buffer overflow may occur and codewords may be
lost.
Reliability and Throughput Efficiencies of the Basic ARQ Schemes
The performance of an ARQ error-control system is
normally measured by its reliability and throughput .effici-
ency. In an ARQ system, the receiver commits a decoding
error whenever it accepts a received word with undetected
errors. Such an event is called an error event. Let P(E) denote
the probability of an error event. Clearly, for an ARQ system
to be reliable, the P(E) should be made very small. The
throughput efficiency (simple throughput) of an ARQ system
is defined as the ratio of the average number of information
December 1984-Vol. 22, No. 12
IEEE Communications Magazine
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Contr ˆ ole d’erreur
Protocoles Selective Repeat
N ARQ system on full duplex links. This remains the standard
for packet-switching networks.
The basic go-back-N ARQ scheme is illustrated in Fig. 2.
The transmitter continuously transmits codewords in order
and then stores them pending receipt of an ACKlNAK for
each. The acknowledgment for a codeword arrives after a
round-trip delay, defined as the time interval between the
transmission of a codeword and the receipt of an acknowl-
edgment for that codeword. During this interval, N - 1 other
codewords are also transmitted. Whenever the transmitter
receives a NAK indicating that a particular codeword, say
codeword i , was received in error, it stops transmitting new
codewords. Then it goes back to codeword i and proceeds to
retransmit that codeword and the N - 1 succeeding
codewords which were transmitted during one round-trip
delay. At the receiving end, the receiver discards the
erroneously received word i and all N - 1 subsequently
received words, whether they are error-free or not. Retrans-
mission continues until codeword i is positively acknowl-
edged. In each retransmission for codeword i , the transmitter
resends the same sequence of codewords. As soon as
codeword i is positively acknowledged, the transmitter
proceeds to transmit new codewords.
The main drawback of go-back-N ARQ is that, whenever a
received word is detected in error, the receiver also rejects
the next N - 1 received words, even though many of them
may be error free. As a result, they must be retransmitted.
This represents a waste of transmissions, which can result in
severe deterioration of throughput performance if large
round-trip delay is involved. For example, consider a
satellite channel with a round-trip delay of approximately
700 ms. If the codeword length n is 1000 bits long and the bit
rate is 1 Mb/s, then in one round-trip delay, N = 700
codewords are transmitted. Therefore, when one received
word is detected in error, 700 received words are rejected. If
errors occur often enough, the system throughput may fall
off very rapidly.
The go-back4 ARQ scheme becomes quite ineffective for
communications systems with high data rates and large
round-trip delays. This ineffectiveness is caused by the
retransmission of many error-free codewords following a
codeword detected in error. This can be overcome by using
the selective-repeat ARQ protocol. In a selective-repeat ARQ
error-control system, codewords are also transmitted con-
7
tinuously. However, the transmitter only resends those
codewords that are negatively acknowledged (NAK’ed). After
resending a NAK’ed codeword, the transmitter continues
transmitting new codewords in the transmitter buffer (as
illustrated in Fig. 3). With this scheme, a buffer must be
provided at the receiver to store the error-free codewords
following a received word detected in error, because,
ordinarily, codewords must be delivered to the end user in
correct order, for example, in point-to-point communications.
When the first NAK’ed codeword is successfully received, the
receiver then releases any error-free codewords in consecu-
tive order from the receiver buffer until the next erroneously
received word is encountered. Sufficient receiver buffer
storage must be provided in a selective-repeat ARQ system;
otherwise, buffer overflow may occur and codewords may be
lost.
Reliability and Throughput Efficiencies of the Basic ARQ Schemes
The performance of an ARQ error-control system is
normally measured by its reliability and throughput .effici-
ency. In an ARQ system, the receiver commits a decoding
error whenever it accepts a received word with undetected
errors. Such an event is called an error event. Let P(E) denote
the probability of an error event. Clearly, for an ARQ system
to be reliable, the P(E) should be made very small. The
throughput efficiency (simple throughput) of an ARQ system
is defined as the ratio of the average number of information
December 1984-Vol. 22, No. 12
IEEE Communications Magazine
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
G´ en´ eralites
Principales Fonctionalit ´ es MAC et couche physique
Acc` es multiple : OFDMA (DL) et SC-FDMA (UL),
Duplexage : FDD et TDD possible,
scheduling/diversit ´ e multi-utilisateurs,
AMC et HARQ,
technologies MIMO,
mitigation de l’interf ´ erence inter-cellule,
Applications multicast et broadcast,
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
Bandes allou´ ees
3GPP Long-Term Evolution 667
This implies that for a considerable number of years, the two systems will coexist, usually in the
same frequency band. Transitions from one system to another will be frequently required, espe-
cially during the initial rollout of LTE, when only parts of the country will be covered by LTE
Base Stations (BSs). Transition times for real-time applications should be less than 300 ms, and for
nonreal-time applications should be 500 ms.
27.2 System Overview
27.2.1 Frequency Bands and Spectrum Flexibility
LTE can be operated in a variety of frequency bands that are assigned by national frequency
regulators, based on the decisions of the World Radio Conference. This spectrum can be, in principle,
used for any member of the IMT-2000 and IMT-Advanced family. It originally encompassed the
frequency bands discussed in Section 26.3; later additional frequencies were assigned, which became
available through the so-called “digital dividend,” – i.e., spectrum that was freed up when TV
was converted to digital transmission techniques that required less spectrum than the old analog
techniques. Tables 27.2 and 27.3 show the bands available by the time of this writing (2009).
However, not all bands are available in all countries. In Europe, band 1 is the same as that assigned
for WCDMA, thus anticipating existing operators to migrate, without new frequency assignments,
from WCDMA to LTE. Migration from GSM to LTE is anticipated by the creation of bands 3
and 8 (in Europe). Similarly, bands 2, 4, 10, cover the Personal Communication System (PCS)
frequencies currently occupied by operators in the U.S.A., while band 5 covers the frequencies
long used by U.S. operators for lower-frequency operation. Bands 6 and 9 cover the traditional
Japanese operator frequencies. Bands that became available through the digital dividend include
several blocks of spectrum in the 700-MHz range have been recently auctioned off (bands 12, 13,
14, 17) in the U.S.A. In Europe and Asia, current activity concentrates on the 2,300–2,700 MHz
range (bands 7, 38, 40); further spectrum in the 3,400–3,600-MHz range will become available
in the near future. Note that some newly available spectra are reserved exclusively for specific
systems, while other frequencies might be used by the operators as they deem fit.
Table 27.2 Bands for FDD operation of LTE
Operating band UL (MHz) DL (MHz) Bandwidth
1.4 3 5 10 15 20
1 1920–1980 2110–2170 Europe, Asia
2 1850–1910 1930–1990 America
3 1710–1785 1805–1880 Europe, Asia
4 1710–1755 2110–2155 America
5 824–849 869–894 America
6 830–840 875–885 Japan
7 2500–2570 2620–2690 Europe, Asia
8 880–915 925–960 Europe, Asia
9 1750–1785 1845–1880 Japan
10 1710–1770 2110–2170 Americas
11 1428–1453 1476–1501 Japan
12 698–716 728–746 Americas
13 777–787 746–756 Americas
14 788–798 758–768 Americas
17 704–716 734–746 Americas
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
Duplexage
IEEE Communications Magazine • April 2009 47
supported in TDD, providing downlink-
uplink periodicities of 5 ms or 10 ms and
downlink-to-uplink ratios from 2:3 to 9:1. As
the number of uplink and downlink sub-
frames in a radio frame can differ, there is
no inherent one-to-one relation between
downlink and uplink subframes, resulting in
some minor differences in the control-signal-
ing design between FDD and TDD.
An essential part of any TDD system is the
provisioning of sufficiently large guard periods
during which equipment can switch between
transmission and reception with no overlap of
signals to be transmitted and received. In LTE,
guard periods are created by splitting one or two
subframes, referred to as special subframes, in
each radio frame into three fields: a downlink
part (DwPTS), a guard period (GP), and an
uplink part (UpPTS).
The downlink part of the special subframe
(DwPTS) can be viewed as an ordinary, albeit
shorter, downlink subframe and is used for
downlink data transmission. Its length can be
varied from three up to twelve OFDM symbols.
Unlike normal subframes, where the control
region can span up to three OFDM symbols, the
maximum control region in DwPTS is two
OFDM symbols. The reason is the location of
the primary synchronization signal in the third
OFDM symbol of the DwPTS in the case of
TDD operation. On the other hand, for FDD
the synchronization signal is located in the mid-
dle of subframe zero and five. The difference in
synchronization-signal location enables the ter-
minals to detect the duplex of the cell already at
initial synchronization to a carrier.
The uplink part of the special subframe
(UpPTS) has a short duration, one or two OFDM
symbols, and can be used for transmission of
uplink sounding-reference signals and random
access. Sounding-reference signals are known sig-
nals, transmitted from the terminals and enabling
the base station to estimate uplink channel-quali-
ty, for example, for the purpose of uplink chan-
nel-dependent scheduling and link adaptation. In
addition to the UpPTS, uplink-channel sounding
also can use the normal subframes in the same
manner as in FDD. Random access typically uses
one of the normal subframes as in FDD, enabling
a relatively long random-access preamble to also
provide coverage and capacity in large cells.
However, in scenarios where random-access cov-
erage is not an issue, a short random-access
preamble in the UpPTS can be used instead.
The remaining symbols in the special subframe,
which have not been allocated to DwPTS or
UpPTS, are used to provide the guard period for
the downlink-to-uplink and the uplink-to-downlink
switch. The length of the guard period depends on
several factors. First, the guard period must be
sufficiently long to handle the propagation delay
in the cells. To be time aligned at the base sta-
tions, terminals closer to the cell edge must start
their transmission earlier in time than those close
to the base stations. Obviously, the transmission at
a cell-edge terminal cannot start until the down-
link has been received completely. Hence, the
guard period must cover the maximum roundtrip
propagation delay within the cell in addition to the
time it takes for a terminal to switch from recep-
tion to transmission. In addition, the guard time
also must be selected by taking base-station-to-
base-station interference into account. Due to the
propagation delay, a downlink transmission from a
distant base station is still “in the air” at the base
station trying to receive uplink transmissions even
though all base stations switched from downlink to
uplink at the same time. With the DwPTS and
UpPTS durations mentioned above, LTE supports
a guard period ranging from two to ten OFDM
symbols (140–667 µs), sufficient for cell sizes up to
and beyond 100 km.
Supporting multiple configurations of the
special subframe is useful not only to support
different cell sizes and propagation conditions as
discussed above, but also to support coexistence
between LTE and other, already deployed TDD
systems, most notably TD-SCDMA [5]. To avoid
inter-system interference (without resorting to
expensive filtering of large guard bands), uplink
and downlink transmissions in LTE and TD-
SCDMA should be mutually aligned, that is, the
downlink-uplink switch-points should coincide
between the two systems. The subframe duration
of LTE and the universal terrestrial radio access
(UTRA) TDD technologies TD-SCDMA and
TD-CDMA are 1 ms, 0.675 ms, and 0.667 ms,
respectively, making such an alignment challeng-
ing. However, the use of the special subframe in
LTE offers an elegant solution to this problem.
By selecting the appropriate length of the
DwPTS and UpPTS, switch-point alignment
between LTE-TDD and TD-SCDMA (and other
TDD-based radio-access schemes) can be
achieved. In fact, coexistence with TD-SCDMA
was one of the main technical reasons for the
introduction of the special subframes.
MULTI-ANTENNA TRANSMISSION
Support for multi-antenna transmission was an
integral part of LTE from the first release, and the
channel quality measurements for link adaptation
and scheduling are designed to cater to this. The
fact that the performance requirements are set,
assuming all terminals support at least two receive
antennas, is important because it enables the net-
works to be planned, assuming at least the pres-
ence of downlink-receive diversity. More advanced
multi-antenna schemes also are supported by
LTE, including transmit diversity, spatial multi-
plexing (including both so-called single-user multi-

Figure 4. LTE spectrum (bandwidth and duplex) flexibility. Half duplex
FDD is seen from a terminal perspective.
f
DL
FDD
Paired spectrum Unpaired spectrum
1.4 MHz
Bandwidth flexibility
Duplex flexibility
20 MHz
f
UL
f
DL
Half-duplex FDD
Reduced UE complexity
f
UL
f
DL/UL
TDD
PARKVALL LAYOUT 3/25/09 2:17 PM Page 47

IEEE Communications Magazine • April 2009 46
paired and unpaired bands. Paired frequency
bands implies that uplink and downlink trans-
missions are assigned separate frequency bands,
whereas in the case of unpaired frequency bands,
uplink and downlink must share the same fre-
quency band. Also, at least in an initial migra-
tion phase, different radio-access technologies
often must be able to operate jointly in the same
spectrum band. Spectrum flexibility, enabling
operation under all these conditions, is one key
feature of the LTE radio access.
LTE is able to operate not only in different
frequency bands, but it also can be deployed
with different bandwidths in order to operate in
spectrum of different sizes, as well as to enable
efficient migration of other radio-access tech-
nologies to LTE. More specifically, as illustrated
in Fig. 4, LTE allows for an overall system band-
width ranging from as small as 1.4 MHz up to 20
MHz, where the later is required to provide the
highest LTE data rates. All terminals support
the widest bandwidth. Unlike previous cellular
systems, LTE provides the possibility for differ-
ent uplink and downlink bandwidths, enabling
asymmetric spectrum utilization.
To enable a terminal to access a cell prior to
knowing the cell bandwidth and the duplexing
scheme, the system information occupies only
the most narrow bandwidth supported by LTE
and is located in subframes guaranteed to be
downlink subframes. After the terminal acquires
the system information, the cell bandwidth and
the duplexing scheme is known, and the terminal
can access the cell based on this knowledge.
SPECTRUM FLEXIBILITY — DUPLEX SCHEME
An important part of spectrum flexibility, as pre-
viously mentioned, is the possibility to operate in
paired, as well as unpaired, spectrum allocations.
Support for paired and unpaired spectrum is in
itself not new to 3GPP. However, in the past this
has been accomplished through different 3G
radio-interface specifications: WCDMA for
FDD and TD-SCDMA, as well as TD-CDMA
for TDD, resulting in dual-mode terminals being
relatively uncommon so far. Therefore, a strong
requirement [5] of the LTE design was to pro-
vide a single radio interface supporting both
FDD and TDD to provide an even larger econo-
my-of-scale benefit to both duplex schemes.
Virtually all of the physical-layer processing is
identical for FDD and TDD, enabling low-cost
implementation of terminals supporting both the
FDD and TDD modes of operation. The differ-
ence between the two is mainly in the frame
structure as illustrated in Fig. 2.
• In the case of FDD operation (upper part
of Fig. 2), there are two carrier frequencies,
one for uplink transmission (f
UL
) and one
for downlink transmission (f
DL
). Thus, dur-
ing each frame, there are ten uplink sub-
frames and ten downlink subframes; and
uplink and downlink transmission can occur
simultaneously within a cell. Inherently
there is a one-to-one relation between
downlink and uplink subframes, which is
exploited in the control-signaling design.
• In the case of TDD operation (lower part of
Fig. 2), there is only a single-carrier frequen-
cy, and uplink and downlink transmissions
always are separated in time, also on a cell
basis. To meet different requirements on
uplink-downlink traffic asymmetries, seven
different uplink-downlink configurations are

Figure 2. LTE frame structure.
(Special subframe)
Subframe #0 #1 #2 #3 #4 #5 #6 #7 #8 #9
GP DwPTS UpPTS
One subframe, T
subframe
= 1 ms
One radio frame, T
frame
= 10 ms
UL
FDD
(Special subframe)
DL
f
UL
f
DL
UL
TDD
DL
f
DL/UL

Figure 3. Channel-quality variations in frequency and time.
Frequency
1
m
s
180 kHz
User #1 scheduled
User #2 scheduled
Time frequency
fading, user #1
Time frrequency
fading, user #2
T
im
e
PARKVALL LAYOUT 3/25/09 2:17 PM Page 46

Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
Architecture L1-L3
670 Wireless Communications
Layer 3
Logical channels
RRC
Medium Access Control
Physical layer
Transport channels
Layer 2
Layer 1
C
o
n
t
r
o
l
/
m
e
a
s
u
r
e
m
e
n
t
s
Figure 27.2 Protocol structure of LTE.
Reproduced from [3GPP LTE] © 2009. 3GPP
TM
TSs and TRs are the property of ARIS, ATIS, CCSA, ETSI, TTA and TTC
who jointly own the copyright in them. They are subject to further modifications and are therefore provided
to you “as is” for information purposes only. Further use is strictly prohibited.
uses all available subcarriers, then this is identical to the single-carrier transmission with Cyclic
Prefix (CP), as described in Section 19.11. However, in LTE, an MS can also use just a subset
of subcarriers and apply the DFT precoding to this subset only.
• The multiple access format for both uplink and downlink is OFDMA combined with Time
Division Multiple Access (TDMA). In other words, the spectral resources, as represented in
the time/frequency plane, are assigned in a flexible manner to the different users. Furthermore,
different users can have different data rates. The transmissions for a specific user are scheduled
to happen in those frequency bands that offer the best propagation conditions, thus exploiting
multiuser diversity.
• Multicast/Broadcast over Single Frequency Network (MBSFN), i.e., transmission of the same
information from different BSs, can be realized in a straightforward manner using OFDM, as
long as the runtime differences of the signals from the different BSs are less than the cyclic prefix.
• LTE provides means for intercell interference coordination, i.e., making sure that signals emitted
in one cell do not interfere catastrophically with signals in the neighboring cells. Note that
LTE, using essentially Frequency Division Multiple Access (FDMA)/TDMA for multiple access,
is more dependent on intercell interference coordination than, e.g., WCDMA. The interference
coordination used in LTE tends to be more sophisticated than the simple “frequency reuse”
discussed in Section 17.6.1.
• Support for the multiple antennas, including receive diversity, various forms of transmit diversity,
and spatial multiplexing (see Chapter 20).
• Adaptive modulation and coding, together with advanced coding schemes.
• FDD or TDD can be used for duplexing, depending on the assigned frequency bands. FDD and
TDD modes are very similar (in contrast to WCDMA), though a number of signaling details and
parameter choices differ. To keep the description compact, this chapter only considers the FDD
mode unless otherwise stated. LTE also foresees half-duplexing, where transmission and reception
for a specific terminal are distinguished both by different times and by different frequencies; this
eases implementation of MSs without affecting spectral efficiency (see Section 17.5).
Generating the PHY signal for the downlink then consists of the following steps (see Figure 27.3):
• Error correction encoding: (see Section 27.3.6).
• Scrambling of coded bits: the bits of all transport channels are scrambled by multiplication with
a Pseudo Noise (PN) (Gold) sequence. Note that – in contrast to WCDMA – it is the bits, and
not the complex-valued symbols, that are scrambled.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
Pile protocolaire de l’interface radio
IEEE Communications Magazine • April 2009 53
the radio channel. In addition to its advantages
for a low-complexity receiver design, the multi-
carrier concept enables the operation of LTE in
various system bandwidths up to 20 MHz by
adapting the number of subcarriers used to the
allocated system bandwidth. Finally, OFDM sup-
ports multi-user access because within a trans-
mission interval, subcarriers can be allocated to
different users.
The LTE uplink employs a discrete Fourier
transform (DFT)-spread OFDM (also denoted
as single-carrier frequency division multiple
access [SC-FDMA]). Compared to conventional
OFDM, this OFDM variant provides an
improved peak-to-average power ratio that
enables more power-efficient terminals.
The basic LTE radio resource that is address-
able for data transmission in the two-dimension-
al time-frequency grid is called a resource block.
This type of resource block assembles 12 subcar-
riers and has a bandwidth of 180 kHz. In the
time domain, the resource block has a subframe
duration of only 1 ms. Such a short subframe
enables the exploitation of channel variations by
scheduling users depending on their current
channel quality. At the same time, a short hybrid
ARQ (HARQ) round-trip time of only 8 ms can
be obtained.
By allocating a variable number of resource
blocks to a certain user and selecting a modula-
tion and coding scheme to meet the current
channel conditions, widely scalable transport
block sizes are possible, resulting in a wide range
of user-data rates. In addition, it is possible to
aggregate up to two streams by utilizing multi-
ple-input multiple-output (MIMO) transmissions
to increase the data rate even further under
favorable radio conditions.
USER PLANE PROTOCOL STACK
While the physical layer essentially provides a bit
pipe, protected by turbo-coding and a cyclic
redundancy check (CRC), the link-layer proto-
cols enhance the service to upper layers by
increased reliability, security, and integrity. In
addition, the link layer is responsible for the
multi-user medium access and scheduling.
One of the main challenges for the LTE link-
layer design is to provide the required reliability
levels and delays for Internet Protocol (IP) data
flows with their wide range of different services
and data rates. In particular, the protocol over-
head must scale. For example, it is widely
assumed that voice over IP (VoIP) flows can tol-
erate delays on the order of 100 ms and packet
losses of up to 1 percent. On the other hand, it
is well-known that TCP file downloads perform
better over links with low bandwidth-delay prod-
ucts. Consequently, downloads at very high data
rates (e.g., 100 Mb/s) require even lower delays
and in addition, are more sensitive to IP packet
losses than VoIP traffic.
Overall, this led to the following design of the
LTE link layer: it consists of three sublayers
(Fig. 2) that are partly intertwined. The Packet
Data Convergence Protocol (PDCP) sublayer [4]
is responsible mainly for IP header compression
and ciphering. In addition, it supports lossless
mobility in case of inter-eNB handovers and pro-
vides integrity protection to higher layer-control
protocols. The radio link control (RLC) sublayer
[5] comprises mainly ARQ functionality and sup-
ports data segmentation and concatenation. The
latter two minimize the protocol overhead inde-
pendent of the data rate, as is explained in more
detail below. Finally, the medium access control
(MAC) sublayer [6] provides HARQ and is
responsible for the functionality that is required
for medium access, such as scheduling operation
and random access.
Figure 3 depicts the data flow of an IP packet
through the link-layer protocols down to the
physical layer. The figure shows that each proto-
col sublayer adds its own protocol header to the
data units.
This figure is used in subsequent sections as a
reference when relevant functionality is intro-
duced.
RETRANSMISSION HANDLING
As in any communication system, there are occa-
sional data transmission errors, for example, due
to noise, interference, and/or fading. Link-layer,
network-layer (IP), and transport-layer protocols
are not prepared to cope with bit errors in head-

Figure 1. Overview of the EPC/LTE architecture.
Control interface
User data interface
S10
EPC
E-UTRAN
S11
S1-MME
MME
S5/S8
P-GW
S-GW
MME
eNB eNB
Internet,
operator services, etc.
S1-U S1-U
X2-U
X2-C

Figure 2. User plane protocol stack.
UE
ARQ
(incl. Seg/Conc.)
Re-ordering
Header
compression
Ciphering
HARQ
PDCP
RLC
MAC
PHY
eNB
PDCP
RLC
S
c
h
e
d
u
l
i
n
g
MAC
PHY
MEYER LAYOUT 3/25/09 2:18 PM Page 53

Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
Pile protocolaire de l’interface radio
Protocoles de l’interface radios
Radio Ressource Control (RRC),
Packet Data Convergence Protocol (PDCP) : ROHC +
chiffrement,
Radio Link Control (RLC) (3 modes disponibles : transparent,
non-acquitt ´ e et acquitt ´ e) : segmentation et concat ´ enation des
PDCP PDU et gestion RLC ARQ,
M´ edium Access Control (MAC) : H-ARQ, mapping entre canaux
logiques et physiques, multiplexing
Couche physique (PHY)
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
Packet flow
ers, and the majority of the protocols are not
capable of handling errors in the payload either.
Therefore, a fundamental design choice for LTE
has been not to propagate any bit errors to high-
er layers but rather to drop or retransmit the
entire data unit containing bit errors. As illus-
trated in Fig. 3, the physical layer attaches a 24-
bit CRC checksum to the data units, thus
allowing the receiver to detect bit errors and to
forward only error-free packets to the IP layer.
Most TCP/IP protocols are designed to cope
only with rather low packet-loss rates. State-of-
the-art voice codecs perform well at error rates
up to at most 10
–2
. High-speed, TCP-based file
downloads requires loss rates on the order of
10
–5
to 10
–6
[7]. The HARQ scheme on the MAC
sublayer performs retransmissions of corrupted
transport blocks and thereby corrects the majori-
ty of all transmission errors. The HARQ mecha-
nism is very similar to the solution adopted for
HSDPA [2], that is, the protocol uses multiple
stop-and-wait HARQ processes. The functionali-
ty and performance is comparable to that of a
window-based selective repeat protocol. In par-
ticular, it allows continuous transmission, which
cannot be achieved with a single stop-and-wait
scheme. Instead of a status message containing a
sequence number, a single-bit HARQ feedback
acknowledgment/negative acknowledgment
(ACK/NACK), with a fixed-timing relation to
the corresponding transmission attempt, pro-
vides information about the successful reception
of the HARQ process. Thereby, it gains in terms
of delay, simplicity, and control overhead com-
pared to a window-based selective repeat proto-
col.
It is important that the HARQ protocol is
fast yet consumes as few radio resources as pos-
sible. The single-bit HARQ feedback fulfils
these requirements, but the probability for mis-
interpreting a negative acknowledgment as a
positive acknowledgment, and thereby causing a
residual packet loss, is in the order of 10
–4
to
10
–3
. It would be expensive, in terms of transmit
power, to reduce the feedback error rate further
and thereby, to ensure the desired very low
residual loss rates required by TCP for achieving
high-data rates. Furthermore, certain errors in
other control signaling, such as scheduling infor-
mation, result in HARQ failures. When such
failures are detected by the receiver, the HARQ
process typically has been re-used for new data,
and the single-bit HARQ feedback is not a valid
reference for the desired retransmission.
Due to the error cases mentioned above, the
fast HARQ protocol with low-overhead,
ACK/NACK feedback and retransmissions with
incremental redundancy is complemented by a
highly reliable window-based selective repeat-
ARQ protocol that resides in the RLC sublayer
as depicted in Fig. 4.
When the CRC check is successful, the MAC
HARQ receiver delivers RLC protocol data
units (PDUs) to the corresponding RLC entity.
If the RLC receiver detects a gap in the
sequence of received PDUs based on the RLC
sequence number, it starts a reordering timer
assuming that the missing packet still is being
retransmitted in the HARQ protocol. Note that
the reordering functionality required on top of a
multi-process stop-and-wait mechanism reuses
the same RLC sequence numbers as the ARQ
mechanism, saving additional overhead com-
pared to a sequence, number-based re-ordering
mechanism in MAC, like in HSPA. In the rare
case that the reordering timer expires, an RLC
acknowledged-mode (AM) receiver sends a sta-
tus message comprising the sequence number of
the missing RLC PDU(s) to its transmitting peer
entity. The MAC layer treats the RLC status
message as any other data, meaning that it also
applies the same HARQ operation and CRC to
IEEE Communications Magazine • April 2009 54

Figure 3. Illustration of data flow through L2 protocol stack.
L1
Coding,
interleaving,
modulation
RLC
Segmentation
concatenation
MAC
Multiplexing
IP
IP
via S1 or from
UE’s stack
PDCP
Header compression
and ciphering
UDP Payload
H H
PDCP PDCP SDU
RLC RLC SDU
MAC MAC SDU
MAC
CRC
MAC SDU
RLC RLC SDU
Segmentation
Multiplexing (padding)
Concatenation
PDCP
PDCP
PDU
RLC
PDU
MAC
PDU
IP TCP Payload
Transport block
The two-layer ARQ
design achieves low
latency and low
overhead without
sacrificing reliability.
Most errors are
captured and
corrected by the
lightweight HARQ
protocol. Only
residual HARQ errors
are detected and
resolved by the more
expensive (in terms
of latency and
overhead) ARQ
retransmissions.
MEYER LAYOUT 3/25/09 2:18 PM Page 54

Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
Link layer : contr ˆ ole d’erreur
IEEE Communications Magazine • April 2009 55
this message. Consequently, errors or loss of the
ARQ feedback can be detected and recovered
by sending another RLC status. Upon reception
of the RLC status message, the ARQ transmitter
triggers a retransmission of the corresponding
RLC PDU(s). The HARQ layer does not
attempt to combine the RLC retransmission
with the previous transmission but treats it as
new data.
The fact that RLC and MAC are terminated
in the same nodes enables a tighter interconnec-
tion of the two protocol sublayers. In LTE, the
HARQ transmitter can indicate a so-called local
NACK to the ARQ transmitter, if it knows or
assumes that a HARQ transmission failed. The
most prominent example is that the maximum
number of HARQ transmission attempts was
reached. This might happen, for example, if the
chosen modulation and coding scheme is too
weak for the given channel quality. The main
benefit of the local NACK is the shorter detec-
tion delay, resulting in improved performance
compared to the gap detection at the ARQ
receiver.
The two-layer ARQ design achieves low
latency and low overhead without sacrificing reli-
ability. Most errors are captured and corrected
by the lightweight HARQ protocol. Only residu-
al HARQ errors are detected and resolved by
the more expensive (in terms of latency and
overhead) ARQ retransmissions.
Finally, it should be noted that services that
can sustain error rates on the order of 10
–3
to
10
–2
, while gaining from reduced delays, can be
mapped to a radio bearer running the RLC pro-
tocol in unacknowledged mode (UM), that is,
without the second ARQ layer. In that case,
residual errors on the MAC layer are not recov-
ered, but packet losses propagate to higher lay-
ers. RLC UM generally is assumed to be used
for VoIP and real-time gaming traffic.
RANDOM ACCESS
To keep transmissions from different UEs
orthogonal, uplink transmissions in LTE are
aligned with the frame timing at the eNB. When
timing is not aligned yet or alignment was lost
due to a period of inactivity during which time
alignment was not maintained by the eNB, a
random access (RA) procedure is performed to
acquire time alignment.
The RA procedure establishes uplink-time
alignment by means of a four-phase contention-
based procedure outlined in the following and
shown in Fig. 5:
1. RA Preamble — The UE randomly selects
an RA preamble sequence from the set of
sequences available in the cell and trans-
mits it on an RA channel. A guard period
is applied to the RA preamble transmission
to avoid creating interference in adjacent
subframes. To minimize non-orthogonal
transmissions and thereby improve resource
efficiency, unsynchronized and unscheduled
transmissions, like the first step in the RA
procedure, do not carry data.
2. RA Response — The eNB detects the
preamble transmission, estimates the uplink
transmission timing of the UE, and
responds with an RA response providing
the UE with the correct timing-advance
value to be used for subsequent transmis-
sions and with a first grant for an uplink

Figure 4 HARQ and ARQ retransmissions on MAC and RLC layer.
Uplink L1
Sliding window ARQ
UL ARQ
transmitter
RLC
Stop and wait HARQ MAC
RLC STATUS
BLER ~10
-4
-10
-3
HARQ ACK/NACK
error rate=10
-4
-10
-3
Transport block
BLER e.g.10
-1
RLC STATUS
as DL HARQ data
RLC PDU
BLER ~10
-4
-10
-3
RLC SDU
BLER ~10
-6
UL HARQ
transmitter
DL HARQ
receiver
UL ARQ
receiver
DL HARQ
transmitter
UL HARQ
receiver
Downlink L1

Figure 5. Contention-based random access procedure.
eNB
2 4
1 3
R
A

p
r
e
a
m
b
l
e
R
A

m
e
s
s
a
g
e
(
U
E

i
d
e
n
t
i
t
y
,

B
S
R
,

e
t
c
.
)
F
u
r
t
h
e
r

u
p
l
i
n
k
/
d
o
w
n
l
i
n
k
t
r
a
n
s
m
i
s
s
i
o
n
s
R
A

r
e
s
p
o
n
s
e
(
t
i
m
i
n
g

a
d
v
a
n
c
e
,

U
L

g
r
a
n
t
,

e
t
c
.
)
R
A

c
o
n
t
e
n
t
i
o
n

r
e
s
o
l
u
t
i
o
n
(
U
L

g
r
a
n
t
,

D
L

a
s
s
i
g
n
m
e
n
t
)
UE
MEYER LAYOUT 3/25/09 2:18 PM Page 55

Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
Canaux logiques, de transports et physiques
686 Wireless Communications
PCCCH
BCCH
CCCH
DCCH
DTCH
MCCH
MTCH
CCCH
DCCH
DTCH
ULSCH
RACH
PUSCH
PRACH
PUCCH
PCCH
BCH
DLSCH
MCH
PBCH
PDSCH
PMCH
PDCCH
PHICH
PCFICH
Logical Transport Physical
Downlink Channels
Logical Transport Physical
Uplink Channels
Figure 27.14 Mapping between logical, transport, and physical channels.
◦ Physical Random Access CHannel (PRACH): it is used for the random access, i.e., MS com-
municating to the BS before a connection with scheduling has been established.
Figure 27.14 summarizes the mapping between the channels.
27.4.2 Synchronization Signals
The SS carries information about the timing of the cell, as well as the cell ID. LTE actually provides
two SSs, the Primary Synchronization Signal (PSS) and the Secondary Synchronization Signal
(SSS). In contrast to other systems, these signals are not called “channels,” but perform similar
functions as, e.g., the synchronization channels in WCDMA. To understand the functionality of the
SS, keep in mind that there are 504 cell IDs defined for LTE, which are divided into 168 ID groups.
The PSS is transmitted in the last symbol of the first slot of subframes 0 and 5 of every frame,
extending over 72 subcarriers. The waveform transmitted during that slot is one of three allowed
Zadoff–Chu sequences of length 63, extended with 5 zeroes at the lower and 5 zeroes at the upper
band edge. Which of the three sequences is transmitted depends on the cell ID within a group
(but note that the PSS does not provide the group ID; that is transmitted in the SSS). The MS can
obtain the fine-resolution timing from the PSS (within one symbol, because the signal lasts only
one symbol, and even within one OFDM sample, by correlation with the Zadoff–Chu sequence).
However, there is still a timing ambiguity by multiples of 5 ms, i.e., the periodicity of the PSS
signal; therefore, frame timing is not available.
The SSS signal is transmitted in the symbol directly before every PSS signal. The SSS carries
information about the cell ID group: the signal also extends over 72 subcarriers. It is an interleaved
combination of two m-sequences (see Section 18.2.6) of length 31. The resulting length-62 sequence
is extended with zeroes at the band edges, just like for the PSS. Only 168 sequences are valid, and
signify the cell ID group. In contrast to PSS, the signal transmitted in the first slot of a frame is
different from the signal in the second slot: while it contains the same m-sequences, the interleaving
in the second slot is different from that in the first slot. This allows the RX to acquire frame timing
(i.e., ambiguity of timing is multiples of 10 ms), as well as the cell ID, from a single observation
of an SSS.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
Couche PHY
3GPP Long-Term Evolution 671
Codewords Layers Antenna ports
Layer
mapper
Precoding
Resource element
mapper
Resource element
mapper
OFDM signal
generation
OFDM signal
generation
Scrambling
Modulation
mapper
Modulation
mapper
Scrambling
Figure 27.3 Overview of the physical layer procedure.
From [3GPP LTE] © 2009. 3GPP
TM
TSs and TRs are the property of ARIS, ATIS, CCSA, ETSI, TTA and TTC
who jointly own the copyright in them. They are subject to further modifications and are therefore provided
to you “as is” for information purposes only. Further use is strictly prohibited.
16-QAM
QPSK
1011
1010
1110 1100 0100 0110
0111 0101 1101 1111
1001
1000
0001 0011
10 00
01 11
0010 0000
Figure 27.4 Mapping of bit combinations onto symbols.
From [3GPP LTE] © 2009. 3GPP
TM
TSs and TRs are the property of ARIS, ATIS, CCSA, ETSI, TTA and TTC
who jointly own the copyright in them. They are subject to further modifications and are therefore provided
to you “as is” for information purposes only. Further use is strictly prohibited.
• Modulation of scrambled bits to generate complex-valued modulation symbols: the modulation
formats used for data transmission are Quadrature-Phase Shift Keying (QPSK), 16-Quadrature
Amplitude Modulation (QAM), and 64-QAM, with Gray mapping (i.e., signal points located
next to each other are distinguished only by 1 bit); see Figure 27.4. The choice of the modulation
format depends on the quality of the propagation channel: for better signal-to-interference- and
Signal-to-Noise Ratio (SNR), higher order modulation can be employed.
• Mapping of the modulation symbols onto the transmission layers: LTE foresees multiple layers
(roughly equivalent to “spatial streams” of Section 20.2) for the transmission with multiple
antennas; see also Section 27.3.7.
• Precoding of the symbols on each layer for transmission on the antenna ports: this step is also
related to multiple-antenna transmission.
• Mapping of symbols to Resource Elements (REs): assign which symbols are to be transmitted
in which time/frequency resource (i.e., time and subcarrier). In the case of multiple transmit
antennas, this mapping is done at each antenna port separately.
• Generating the time domain OFDM signal : (again, for each antenna port separately).
For the uplink, the steps are almost identical, except that:
• The assignment of symbols to time/frequency resources is different; only contiguous subcarriers
can be used by one MS.
• Data scrambling is done with sequences that depend on the MS.
• The signals are DFT encoded before being sent to the OFDM modulator.
Principaux ´ el ´ ements
Codage canal convolutionel et turbo-codes,
Scrambling,
Modulation MQAM, M=(4,16,64),
M´ edium Access Control (MAC) : H-ARQ, mapping entre canaux
logiques et physiques, multiplexing
MIMO,
Mapping des symboles sur les Resource Elements (assignation
des pav´ es temps-fr ´ equence).
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
Couches PHY : trame FDD
672 Wireless Communications
27.3 Physical Layer
27.3.1 Frames, Slots, and Symbols
In LTE, the time axis is divided into entities that play an important role in the transmission of
different channels. These time entities have the following hierarchy (see Figure 27.5):
• The fundamental time unit of LTE transmission is a radio frame, which has a duration of 10 ms.
• Each radio frame is divided into 10 subframes (each being 1ms long). Subframes are the funda-
mental time unit for most LTE processing, like scheduling.
• Each subframe consists of two slots, which are each 0.5 ms long.
• Each slot consists of 7 (or 6) symbols.
Duration of the different units is often given in terms of the sampling time T
s
= 1/30,720,000 s.
Note that this “sampling time” is a bookkeeping unit; RXs are not obligated to actually sample
at the corresponding rate. In particular, for bandwidths <15 MHz, a larger sampling time (lower
sampling frequency) is feasible.
One slot, T
slot
= 15360 T
s
= 0.5 ms
One radio frame, T
f
= 307200 T
s
= 10 ms
#0 #1 #2
One subframe
Slot = 0.5 ms
Normal cyclic prefix mode
Extended cyclic prefix mode (e.g., MB-SFN operation)
T
cp
= 160 T
s
T
cp
= 512 T
s

66.7 µs

= 2048 T
s
144 T
s
CP
CP
CP CP CP CP CP CP
CP
LB #0
LB #0
LB #1 LB #2 LB #3 LB #4 LB #5
LB #5
LB #6
#3 #18 #19
Figure 27.5 Structure of one slot in LTE.
From [3GPP LTE]. © 2009. 3GPP
TM
TSs and TRs are the property of ARIS, ATIS, CCSA, ETSI, TTA and TTC
who jointly own the copyright in them. They are subject to further modifications and are therefore provided
to you “as is” for information purposes only. Further use is strictly prohibited.
Let us now turn to the details of a symbol. Since the modulation format is OFDM (regular OFDM
for the downlink, and DFT-precoded OFDM in the uplink), multiple subcarriers are present. The
regular spacing between the subcarriers is f = 15 kHz. The 15-kHz spacing of the subcarriers
leads to an OFDM symbol duration (without cyclic prefix) of 67 μs = 2048 T
s
. One subcarrier, for
the duration of 1 OFDM symbol, is called a resource element . We can fit 6 or 7 OFDM symbols
into one slot, depending on the duration of the cyclic prefix. In the “regular” case, the duration
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
Long Term Evolution
Couches PHY : grille d’allocation
3GPP Long-Term Evolution 673
of the cyclic prefix is 160 T
s
in the first OFDM symbol and 144 T
s
for the subsequent symbols.
A long cyclic prefix is 512 T
s
, so that only a total of 6 OFDM symbols fit into one slot. Such a
long cyclic prefix is used in environments with large delay spread and/or for MB-SFN. To simplify
the notation, the description in the remainder of this chapter assumes “normal-length” cyclic prefix
unless otherwise stated.
Time/frequency resources are assigned to different users as integer multiples of a Resource Block
(RB) (Figure 27.6). More precisely, an RB is 12 subcarriers (180 kHz) over the duration of one
slot.
3
For the uplink, only contiguous RBs can be assigned to one MS. Furthermore, the number
of RBs has to be decomposable into factors of 2, 3, and 5; this is done to ensure an efficient
implementation: with that prescription, any of the necessary DFTs can be composed of radix-2,
radix-3, and radix-5 butterfly structures.
One uplink slot T
slot
One downlink slot T
slot
Resource block
Resource element (k, l ) Resource element (k, l )
k = 0 k = 0
l = 0
Uplink resources
Downlink resources
l = 0
N
symb
SC-FDMA symbols
UL
N
symb
OFDM symbols
DL
k = N
RB
N
sc
UL RB
– 1
k = N
RB
N
sc
DL RB
– 1
N
symb
× N
sc
UL RB
resource elements
Resource block
N
symb
× N
sc
DL RB
resource elements
N

R
B


×

N

R
U
L
R
B
s
u
b
c
a
r
r
i
e
r
s
N

R
B


×

N

R
U
L
R
B
s
u
b
c
a
r
r
i
e
r
s
N

C
C
R
.
R
s
u
b
c
a
r
r
i
e
r
s
N

C
C
R
.
R
s
u
b
c
a
r
r
i
e
r
s
l = N
symb
– 1
UL
l = N
symb
– 1
DL
Figure 27.6 Resource blocks for uplink and downlink. In this figure: SC-FDMA, Single-Carrier Frequency
Division Multiple Access.
From [3GPP LTE]. © 2009. 3GPP
TM
TSs and TRs are the property of ARIS, ATIS, CCSA, ETSI, TTA and TTC who jointly own the copyright
in them. They are subject to further modifications and are therefore provided to you “as is” for information purposes only. Further use is strictly
prohibited.
In the TDD case, subframes can be assigned flexibly to uplink and downlink, with the exception
of subframes 0 and 5, which are always used for the downlink, and subframe 2, which is always
used for the uplink. For every transition from downlink to uplink, there is a guard interval, to avoid
collisions between the packets “on the air” (see Section 17.5). Consequently, there are subframes
that contain three distinct parts: a Downlink Pilot Time Slot (DwPTS), Uplink Pilot Time Slot
(UpPTS), and a guard interval between them. Note that a guard interval is not necessary for a
transition from uplink to downlink.
3
A reduced subcarrier spacing of 7.5 kHz is also foreseen, together with an extended cyclic prefix. In that case,
only 3 OFDM symbols fit into one slot; but 24 subcarriers are contained in one RB.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
G´ en´ eralit ´ es
Acc` es multiple bas´ es CDMA : DS-CDMA (Direct-Sequence),
duplexage : FDD le plus d´ eploy´ e,
FDD : 2 bandes de 5 Mhz (DOWN et UP)
TDD : une seule bande de 5 Mhz
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
G´ en´ eralit ´ es
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
CDMA et ´ etalement de spectre
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
CDMA et ´ etalement de spectre
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
CDMA et ´ etalement de spectre
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Gain de traitement et d’ ´ etalement
Gain d’ ´ etalement :
SF =
T
S
T
c
B
spr
B
S
Gain de traitement :
G
p
=
B
spr
B
inf
∼ SF + G
c
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
DS-CDMA : codes d’ ´ etalement
Codes de Hadamard :
H
2M
=

H
M
H
M
H
M
−H
M

⇒ H
1
= [+1], H
2M
=

+1 +1
+1 −1

Codes pseudo- al ´ eatoires :
m-s´ equences Codes de Gold
LFSR de taille n XOR de deux m-s´ equ.
Si p´ eriode de taille 2
n
−1, + d´ ecalages possibles
m-s´ equence (maximal length sequ.)
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
DS-CDMA : codes ´ etalement
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
DS-CDMA : acc` es multiple
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Codage de canal
WCDMA/UMTS 649
Input
Input
D
D D D D D D D D
D D D D D D D
(a) Rate 1/2 convolutional encoder
Output 0
Output 1
(b) Rate 1/3 convolutional encoder
Output 0
Output 1
Output 2
Figure 26.8 Structure of convolutional encoders.
Reproduced from [UMTS 1999] © 1999. 3GPP TSs and TRs are the property of ARIB, ATIS, ETSI, CCSA, TTA, and TTC who jointly own the
copyright in them. They are subject to further modifications and are therefore provided “as is” for information purposes only. Further use strictly
prohibited.
x(t )
D
D D D
D D
Y(t )
X(t )
+
+ +
+
+ +
+
+
Interleaver
Y' (t )
X' (t )
Figure 26.9 Structure of a turbo encoder.
Reproduced from [UMTS 1999] © 1999. 3GPP TSs and TRs are the property of ARIB, ATIS, ETSI, CCSA, TTA, and TTC who jointly own the
copyright in them. They are subject to further modifications and are therefore provided “as is” for information purposes only. Further use strictly
prohibited.
26.6 Spreading and Modulation
26.6.1 Frame Structure, Spreading Codes, and Walsh–Hadamard Codes
WCDMA relies on CDMA for multiple access. However, transmission timing is still based on a
hierarchical timeslot structure similar to GSM’s: frames of duration T
f
= 10 ms are divided into
WCDMA/UMTS 649
Input
Input
D
D D D D D D D D
D D D D D D D
(a) Rate 1/2 convolutional encoder
Output 0
Output 1
(b) Rate 1/3 convolutional encoder
Output 0
Output 1
Output 2
Figure 26.8 Structure of convolutional encoders.
Reproduced from [UMTS 1999] © 1999. 3GPP TSs and TRs are the property of ARIB, ATIS, ETSI, CCSA, TTA, and TTC who jointly own the
copyright in them. They are subject to further modifications and are therefore provided “as is” for information purposes only. Further use strictly
prohibited.
x(t )
D
D D D
D D
Y(t )
X(t )
+
+ +
+
+ +
+
+
Interleaver
Y' (t )
X' (t )
Figure 26.9 Structure of a turbo encoder.
Reproduced from [UMTS 1999] © 1999. 3GPP TSs and TRs are the property of ARIB, ATIS, ETSI, CCSA, TTA, and TTC who jointly own the
copyright in them. They are subject to further modifications and are therefore provided “as is” for information purposes only. Further use strictly
prohibited.
26.6 Spreading and Modulation
26.6.1 Frame Structure, Spreading Codes, and Walsh–Hadamard Codes
WCDMA relies on CDMA for multiple access. However, transmission timing is still based on a
hierarchical timeslot structure similar to GSM’s: frames of duration T
f
= 10 ms are divided into
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
DS-CDMA : contr ˆ ole de puissance
Motivations
Effet “Near-Far” : compenser en liaison montante
l’affaiblissement des utilisateurs trop loin pour ´ eviter
“l’ ´ eblouissement”,
Compensation des v´ enouissments temporels,
Maintenir une QoS au r ´ ecepteur
Modes
Boucle ouverte : pour le mobile, d´ eterminer son niveau de
puissance ` a l’aide de canaux d´ edi ´ es, surtout pour
´ evanouissements long terme (path-loss et shadowing),
Boucle ferm´ ee : le r ´ ecepteur calcule son niveau de puissance
requis et envoie les consignes ` a l’ ´ emetteur, pour
´ evanouissements rapides.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
UTRA-FDD : architecture en couches g´ en´ erale
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
UTRA-FDD : stack protolaire
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Canaux de transports et canaux logiques
Canaux logiques
Canaux de contr ˆ oles : Broadcast Control CHannel (BCCH),
Paging Control CHannel (PCCH), Common Control CHannel
(CCCH), Dedicated Control channel (DCCH),
Canaux de traffic : Dedicated Traffic Channel (DTCH), Common
Traffic Channel (CTCH).
Canaux de transports
Canaux communs : Broadcast CHannel (BCH), Paging CHannel
(PCH), Random Access CHannel (RACH), Forward Access
CHannel (FACH), Common Packet CHannel (CPCH), Downlink
Shared CHannel (DSCH),
Canaux d´ edi ´ es : Dedicated (transport) channel (DCH).
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Canaux physiques
Canaux physiques
Voie montante :
Dedicated Physical Control CHannel (DPCCH) :bits pilotes,
Transmit Power Control (TPC), Feed Back Information (FBI),
Transport Format Combination Indicator (TFCI),
Dedicated Physical Data CHannel (DPDCH) : data,
Physical RACH (PRACH) et Physical Common Packet CHannel
(PCPCH).
Voie descendante :
DPCCH et DPDCH,
Canaux de contr ˆ ole propres : Primary Common Control Physical
CHannel (P-CCPCH), Secondary Common Control Physical
CHannel (S-CCPCH), Synchronization CHannel (SCH), Common
PIlot CHannel (CPICH),
autres : Physical Downlink Shared CHannel (PDSCH), Acquisition
Indication CHannel (AICH), Page Indication CHannel (PICH).
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Mapping canaux de transports et physiques
646 Wireless Communications
Transport Channels Physical Channels
DCH
RACH
CPCH
BCH
FACH
PCH
DSCH
Dedicated Physical Data CHannel (DPDCH)
Dedicated Physical Control CHannel (DPCCH)
Physical Random Access CHannel (PRACH)
Physical Common Packet CHannel (PCPCH)
Common Pilot CHannel (CPICH)
Primary Common Control Physical CHannel (P-CCPCH)
Secondary Common Control Physical CHannel (S-CCPCH)
Synchronisation CHannel (SCH)
Physical Downlink Shared CHannel (PDSCH)
Acquisition Indication CHannel (AICH)
Page Indication Channel (PICH)
Figure 26.6 Matching of physical and logical channels.
Reproduced from [UMTS 1999] © 1999. 3GPP TSs and TRs are the property of ARIB, ATIS, ETSI, CCSA, TTA, and TTC who jointly own the
copyright in them. They are subject to further modifications and are therefore provided “as is” for information purposes only. Further use
strictly prohibited.
Excited Linear Prediction (ACELP) (see Chapter 15). WCDMA-AMR contains eight different
encoding modes, with source rates ranging from 4.75 to 12.2 kbit/s, as well as a “background
noise” mode.
4
26.5.2 Multiplexing and Interleaving
Multiplexing, coding, and interleaving are very complicated procedures that allow a high degree of
flexibility. A data stream coming from upper layers has to be processed before it can be transmitted
via the transport channels on the air interface. Transport channels are processed in blocks of 10-,
20-, 40- or 80-ms duration. We first discuss multiplexing and coding in the uplink. The block
diagram in Figure 26.7a illustrates the order of the processes involved in multiplexing and coding:
• When processing a transport block, the first step is to append a Cyclic Redundancy Check (CRC)
field. This field, which can be 8, 12, 16, or 24 bits long, is used for the purpose of error detection. It
is calculated for each block of data for one transmission time interval from the code polynomials:
G(D) = D
8
+D
7
+D
4
+D
3
+D +1 for 8-bit CRC (26.1)
G(D) = D
12
+D
11
+D
3
+D
2
+D +1 for 12-bit CRC (26.2)
G(D) = D
16
+D
12
+D
5
+1 for 16-bit CRC (26.3)
G(D) = D
24
+D
23
+D
6
+D
5
+D +1 for 24-bit CRC (26.4)
and attached at the end of the block.
4
A later standardized “wideband” AMR that results in even better speech quality uses nine modes with rates
between 6.6 and 23.85 kbit/s.
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Mapping canaux de transports et physiques
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Chaˆıne UTRA-FDD, chaine globale
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Chaˆıne UTRA-FDD, voie montante
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Chaˆıne UTRA-FDD, voie montante
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Chaˆıne UTRA-FDD, voie montante
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Chaˆıne UTRA-FDD, voie montante
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Chaˆıne UTRA-FDD, voie montante
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Chaˆıne UTRA-FDD, voie descendante
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Chaˆıne UTRA-FDD, voie descendante
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Chaˆıne UTRA-FDD, voie descendante
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Chaˆıne UTRA-FDD, voie descendante
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : W-CDMA
Chaˆıne UTRA-FDD, voie descendante
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : HSXPA
Chaˆıne HXPA et canaux
Principles of HSDPA 361
14.1.1. HS-DSCH transport channel
The coding chain of the HS-DSCH is depicted in Figure 14.2 [TS 25.212, R5].
After receiving a transport block from MAC, the physical layer adds CRC bits, code
block segmentation, turbo encoding, rate matching, interleaving and constellation
rearrangement before the data is sent to the physical channel(s). The physical layer
spreads and scrambles the data and then maps it to one or more QPSK or 16-QAM
constellation. Bits are taken in groups of 4 and used to obtain the appropriate QAM
symbol [TS 25.213, R5]. Before this symbol lookup, a bit rearrangement is
performed. There are four possible rearrangements – this gives the system the
chance to transmit bit streams in such a way that all bits experience the same
average level of error after combining successive retransmissions in the receiver
side.
Physical channel
segmentation
L1 H-ARQ
Constellation
re-arrangement
(16-QAM)
Interleaving
Physical channel
mapping
HS-PDSCH 1
HS-PDSCHN
CRC
attachment
Bit
scrambling
Code block
segmentation
One transport block
in each TTI
Turbo
encoder
Upper layers Physical layer
Physical channel
segmentation
L1 H-ARQ
Constellation
(16-QAM)
Interleaving
Physical channel
mapping
HS-PDSCH 1
HS-PDSCHN
CRC
attachment
Bit
scrambling
Code block
segmentation
One transport block
in each TTI
Turbo
encoder
Upper layers Physical layer
rearrangement

Figure 14.2. Operations associated to HS-DSCH in the transmitter side
As the DSCH, the HS-DSCH can be shared in time by several users attached to a
Node B. However, the HS-DSCH distinguishes itself from DSCH by the following:
– the Transmission Time Interval (TTI) is always 2 ms (mapped to a radio sub-
frame of 3 slots). This enables short transmit delays between packets while the
channel is shared by multiple users. It also enables better tracking of the time
varying radio conditions and fast multiple retransmissions in the case of receive
errors;
– the number of transport blocks and the number of HS-DSCHs per TTI is
always one;
– only one interleaving operation is applied over 2 ms time periods;
– the adoption of the Hybrid Automatic Repeat reQuest (H-ARQ) mechanism in
the physical layer;

360 UMTS
transfer in the downlink: the DCH, the FACH, the DSCH and the HS-DSCH. The
key characteristics of these channels and those of their corresponding physical
channels are given in Table 14.1. On the other hand, Figure 14.1 summarizes the
transport-channel to physical-channel mapping within Release 5.
Downlink
Physical
channels
Transport
channels
Uplink
DCH
DPDCH
DSCH FACH PCH BCH
PDSCH S-CCPCH P-CCPCH
RACH CPCH
PRACH PCPCH
DCH
DPDCH
HS-DSCH
HS-PDSCH
HS-SCCH
HS-DPCCH
DPCCH
CPICH
SCH
AICH
AP-AICH
PICH
CSICH
CD/CA-ICH
DPCCH
HSDPA
Downlink
Physical
channels
Transport
channels
Uplink
DCH
DPDCH
DSCH FACH PCH BCH
PDSCH S-CCPCH P-CCPCH
RACH CPCH
PRACH PCPCH
DCH
DPDCH
HS-DSCH
HS-PDSCH
HS-SCCH
HS-DPCCH
DPCCH
CPICH
SCH
AICH
AP-AICH
PICH
CSICH
CD/CA-ICH
DPCCH
HSDPA

Figure 14.1. Transport-channel to physical-channel mapping in UTRA/FDD Release 5
Downlink transport channels
DCH FACH DSCH HS-DSCH
TTI (ms) 10, 20, 40, 80 10, 20, 40, 80 10, 20, 40, 80 2
Type of channel
coding
turbo coding
convol. coding
turbo coding
convol. coding
turbo coding
convol. coding
turbo coding
Code rates 1/2, 1/3 1/2, 1/3 1/2, 1/3 1/3
CRC size 0, 8, 12, 16, 24 0, 8, 12, 16, 24 0, 8, 12, 16, 24 24
HARQ yes, in RLC yes, in RLC yes, in RLC yes, in phy layer
Downlink physical channels
DPCH S-CCPCH PDSCH HS-PDSCH
Spreading factor variable, 4-512 variable, 4-256 variable, 4-256 fixed, 16
Codes per user 8 codes (max.) 1 code 8 codes (max.) 15 codes (max.)
Modulation QPSK QPSK QPSK QPSK, 16-QAM
Power control fast slow fast slow
TX diversity
OL STTD and
CL modes 1 and 2
OL STTD
OL STTD and
CL modes 1 and 2
OL STTD and
CL mode 1
Soft handover/
macrodiversity
yes no
associated
DPCH only
associated
DPCH only
Downlink transport channels
DCH FACH DSCH HS-DSCH
TTI (ms) 10, 20, 40, 80 10, 20, 40, 80 10, 20, 40, 80 2
Type of channel
coding
turbo coding
convol. coding
turbo coding
convol. coding
turbo coding
convol. coding
turbo coding
Code rates 1/2, 1/3 1/2, 1/3 1/2, 1/3 1/3
CRC size 0, 8, 12, 16, 24 0, 8, 12, 16, 24 0, 8, 12, 16, 24 24
HARQ yes, in RLC yes, in RLC yes, in RLC yes, in phy layer
Downlink physical channels
DPCH S-CCPCH PDSCH HS-PDSCH
Spreading factor variable, 4-512 variable, 4-256 variable, 4-256 fixed, 16
Codes per user 8 codes (max.) 1 code 8 codes (max.) 15 codes (max.)
Modulation QPSK QPSK QPSK QPSK, 16-QAM
Power control fast slow fast slow
TX diversity
OL STTD and
CL modes 1 and 2
OL STTD
OL STTD and
CL modes 1 and 2
OL STTD and
CL mode 1
Soft handover/
macrodiversity
yes no
associated
DPCH only
associated
DPCH only

Table 14.1. Key parameters of downlink transport/physical data channels in Release 5
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : HSXPA
Chaˆıne HSXPA : structure
362 UMTS
– only turbo coding is used for channel coding based on the Release 99 1/3 turbo
encoder scheme, though other coding rates can be obtained with H-ARQ;
– support for 16-level Quadrature Amplitude Modulation (16-QAM) in addition
to QPSK modulation.
14.1.2. Mapping of HS-DSCH onto HS-PDSCH physical channels
The HS-DSCH is mapped onto one or several High Speed Physical Downlink
Shared Channels (HS-PDSCHs). The structure of an HS-PDSCH is composed of
sub-frames of 3 slots (7,690 chips) each. Channelization coding and scrambling are
applied as shown in Figure 14.3.
Σ
cell scrambling
code (complex)
gain m
HS-PDSCHm
I branch
Q branch
j
QPSK symbols
s
e
r
i
a
l

t
o

p
a
r
a
l
l
e
l
c
16,m
HS-PDSCHk
j
s
e
r
i
a
l

t
o

p
a
r
a
l
l
e
l
c
16,k
S
dl,n
S
dl,n
RRC
filter
Re{.}
Im{.}
RRC
filter
DAC
DAC
Σ
cell scrambling
code (complex)
gain m
HS-PDSCHm
I branch
Q branch
j
QPSK symbols
s
e
r
i
a
l

t
o

p
a
r
a
l
l
e
l
c
16,m
HS-PDSCHk
j
s
e
r
i
a
l

t
o

p
a
r
a
l
l
e
l
c
16,k
S
dl,n
S
dl,n
RRC
filter
Re{.}
Im{.}
RRC
filter
DAC
DAC
PA
16-QAM symbols
channelization
code
I branch
Q branch
gain k
cos(ωt)
-sin(ωt)

Figure 14.3. Simplified transmission chain in Node B for HS-PDSCH
The channelization codes have a fixed spreading factor, SF = 16. This enables a
maximum of 15 parallel codes for user traffic and signaling while leaving one for
other required control and data bearers. Although the available HS-PDSCHs are
primarily shared in the time domain, it is also possible to share the code resources
using code multiplexing, in which case several users share the code resources within
the same TTI, as shown in Figure 14.4. Multi-code transmissions are allowed:
multiple channelization codes (HS-PDSCHs) can be assigned to the UE in the same
TTI, depending on the UE capability. The same scrambling code sequence is applied
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : HSXPA
Chaˆıne HSXPA : trame
364 UMTS
Sub-frame 1 Sub-frame 0 Sub-frame i Sub-frame 4
2 ms (7,680 chips)
HARQ-ACK Channel-Quality Indication (CQI)
5,120 chips
10 ms
Slot 1 Slot 0 Slot 2
Data
M × 160 bits (M = 2 for QPSK and M = 4 for 16-QAM)
Sub-frame (2 ms)
• SF = 16 (fixed)
• Bit rate = 480-960 kbps
• QPSK or 16 -QAM
Slot 1 Slot 0 Slot 2
Data
40 bits
2,560 chips
Sub-frame (2 ms)
a) HS-PDSCH
• SF = 128 (fixed)
• Bit rate = 60 kbps
• QPSK or 16 -QAM
b) HS-SCCH
• SF = 256 (fixed)
• Bit rate = 15 kbps
• QPSK
c) HS-DPCCH
Sub-frame 1 Sub-frame 0 Sub-frame i Sub-frame 4
HARQ-ACK Channel-Quality Indication (CQI)
10 ms
Slot 1 Slot 0 Slot 2
Data
M × 160 bits (M = 2 for QPSK and M = 4 for 16-QAM)
Sub-frame (2 ms)

• Bit rate = 480-960 kbps
• QPSK or 16 -QAM
Slot 1 Slot 0 Slot 2
Data
40 bits
Sub-frame (2 ms)
a) HS-PDSCH

• Bit rate = 60 kbps
• QPSK or 16 -QAM
b) HS-SCCH

• Bit rate = 60 kbps
• QPSK or 16 -QAM
b) HS-SCCH

• Bit rate = 15 kbps
• QPSK
c) HS-DPCCH

• Bit rate = 15 kbps
• QPSK
c) HS-DPCCH
2,560 chips
2,560 chips

Figure 14.5. Sub-frame structure of HS-PDSCH, HS-SCCH and HS-DPCCH
HS-SCCH structure
The HS-SCCH is a fixed rate (60 kbps, SF = 128) downlink physical channel used
to carry the downlink signaling related to the HS-DSCH transmission (see Figure
14.5b). The power of the HS-SCCH is controlled by Node B. This channel conveys:
– the set of channelization codes for the current HS-PDSCH(s);
– the modulation scheme that is being used (QPSK or 16-QAM);
– the size of the transport block;

– the H-ARQ process identifier;

– the Redundancy Version (RV) and constellation re-arrangement parameters;
– the UE identity, i.e. the HS-DSCH Radio Network Identifier (H-RNTI).
The above information provides timing and coding information, thus enabling
the UE to listen to the HS-DSCH at the correct time and using the correct codes to
enable successful decoding of UE data. The UE using an active HS-DSCH must be
capable of receiving up to four parallel HS-SCCHs in order to determine if data is
being transmitted to the UE in the next time period [TS 25.214, R5].
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : HSXPA
Chaˆıne HSXPA : structure canaux
Principles of HSDPA 365
HS-DPCCH structure
The structure of the HS-DPCCH is presented in Figure 14.5c. It is a fixed rate
(15 kbps, SF = 256) uplink channel that carries the acknowledgements (positive
(ACK) or negative (NACK)) of the packet received on HS-PDSCH and also the
Channel Quality Indication (CQI). The CQI are estimated and then transmitted by
the UE in steps of 2.0 ms according to the network configuration and a repetition
scheme can be applied. The HS-DPCCH is always accompanied by a DPCCH in the
uplink. The power of the HS-DPCCH is defined as an offset compared to this
channel.
Associated DPCH
On top of the HS-PDSCH, the HS-SCCH and the HS-DPCCH, every UE has an
associated Release 99 dedicated physical channel (DPCH) in both the uplink and
downlink directions (see Figure 14.6). The downlink associated channel carries the
signal radio bearer for layer 3 signaling as well as power control commands for the
uplink channel, whereas the uplink channel is used as feedback channel, carrying for
instance the TCP and FBI bits. Other services such as speech can also be carried on
the DPCH.
Σ
UE scrambling
code (complex)
HS-DPCCH
I + jQ
Q
c
256,0
DPCCH
c
256,k
j
β
HS
β
C
c
SF/4,n
DPDCH
β
D
S
DPCH,m
I
to RRC
filter
a) Spreading of DPCH/HS-DPCCH in the uplink
Σ
UE scrambling
code (complex)
HS-DPCCH
I + jQ
Q
c
256,0
DPCCH DPCCH
c
256,k
j
β
HS
β
C
c
SF/4,n
DPDCH DPDCH
β
D
S
DPCH,m
I
to RRC
filter
a) Spreading of DPCH/HS-DPCCH in the uplink

HS-DPCCH
HS-SCCH #1
H
S-SCCH #2
H
S-SCCH #3
HS-SCCH
#4
ACK/NACK, CQI
UL D
PCH (R99)
D
L DPCH (R99)
UE
Node B
Modulation, FEC, code
allocation, UE-ID
H
S
-PD
SC
H
b) Interaction of DPCH/HS-DPCCH with HS-SCCH/HS-PDSCH
Data
HS-DPCCH
HS-SCCH #1
H
S-SCCH #2
H
S-SCCH #3
HS-SCCH
#4
ACK/NACK, CQI
UL D
PCH (R99)
D
L DPCH (R99)
UE
Node B
Modulation, FEC, code
allocation, UE-ID
H
S
-PD
SC
H
b) Interaction of DPCH/HS-DPCCH with HS-SCCH/HS-PDSCH
Data

Figure 14.6. Spreading of DPCH/HS-DPCCH and interaction with HS-SCCH/HS-PDSCH
Canaux de propagation et mod´ elisation Syst ` eme multi-utilisateurs : acc` es multiple et duplexage GSM : un exemple de syst ` eme bas´ e FDMA/TDMA Acc` es multiples par R´ epartition en fr ´ equences : OFDMA et SC-FDMA M´ ecanismes d’adaptation et de diversit ´ e 3GPP-LTE UMTS : W-CDMA et HXPA
UMTS : HSXPA
Chaˆıne HSXPA : fonctionalit ´ es avanc´ ees (AMC +HARQ)
Principles of HSDPA 367
radio conditions to maintain the error rate. Table 14.2 gives examples of user data
rates from different combinations of modulation and coding rates.
Modulation
Effective
code rate
Data rate
(1 HS-PDSCH)
Data rate
(5 HS-PDSCHs)
Data rate
(15 HS-PDSCHs)
QPSK 1/4 120 kbps 0.6 Mbps 1.8 Mbps
QPSK 1/2 240 kbps 1.2 Mbps 3.6 Mbps
QPSK 3/4 360 kbps 1.8 Mbps 5.4 Mbps
16-QAM 1/2 480 kbps 2.4 Mbps 7.2 Mbps
16-QAM 5/8 600 kbps 3.0 Mbps 9.0 Mbps
16-QAM 3/4 720 kbps 3.6 Mbps 10.8 Mbps
Modulation
Effective
code rate
Data rate
(1 HS-PDSCH)
Data rate
(5 HS-PDSCHs)
Data rate
(15 HS-PDSCHs)
QPSK 1/4 120 kbps 0.6 Mbps 1.8 Mbps
QPSK 1/2 240 kbps 1.2 Mbps 3.6 Mbps
QPSK 3/4 360 kbps 1.8 Mbps 5.4 Mbps
16-QAM 1/2 480 kbps 2.4 Mbps 7.2 Mbps
16-QAM 5/8 600 kbps 3.0 Mbps 9.0 Mbps
16-QAM 3/4 720 kbps 3.6 Mbps 10.8 Mbps

Table 14.2. User data rates on top of Layer 2 obtained from different coding
rates and modulation schemes (including overhead)
14.3. Hybrid Automatic Repeat Request (H-ARQ)
In a typical ARQ mechanism, the receiver sends an acknowledge (ACK)
message to the sending station when a data block has been successfully received,
e.g. by CRC checksum comparison. When the checksum calculated by the receiver
does not match the checksum included within the transmitted data block, the
receiver will send a negative acknowledge (NACK) to the sender and discard the
erroneous block. This so-called Stop and Wait (SAW) method is not very efficient
for two reasons. First, the sender may retransmit the erroneous block and second, the
transmitter is inactive until it gets a response. Hybrid ARQ is a combination of ARQ
and forward error correction (FEC) aiming at minimizing retransmissions: the
erroneous blocks are kept and are used for a combined detection with
retransmission. Moreover, in order to avoid waiting times, N parallel SAW-ARQ
processes are alternatively used within the same channel in separate TTIs. The delay
between the original and the first retransmission is in the order of 12 ms.
The H-ARQ mechanism employed in the HSDPA concept is located in the
physical layer, thus enabling fast retransmissions. Two retransmission strategies are
used: chase combining and incremental redundancy (IR). With chase combining, the
UE sends a NACK to the sending Node B when detecting a block with errors.
Rather than discarding the erroneous block, it will be stored. In the case where the
retransmitted block is also received in error, the previous block and the current block
are combined weighted by the SNR estimation. Each time a block is resent, the same
coding scheme is used.
Principles of HSDPA 369
14.4. H-ARQ process example
Figure 14.9 illustrates an example of H-ARQ process:
– Node B schedules data on the HS-DSCH to two users UE1 and UE2 according
to packet prioritization and resource availability;
– prior to sending data on the HS-DSCH, Node B transmits the HS-SCCH two
slots in advance of the HS-DSCH;
– the UEs monitors the set of HS-SCCH signaled by the network on every TTI (2
ms). In decoding the UE identity field within the first slot of the HS-SCCH, the UEs
know to whom the data block on the HS-DSCH is destined. If it turns out to be for
UE1, it decodes the remaining information in the HS-SCCH by giving the
parameters required to decode the HS-DSCH: modulation scheme, multi-code set,
H-ARQ process control, etc.;
– the H-ARQ starts after decoding the HS-SCCH and receiving the first transport
block 1. in order for the serving Node B to know if the block was detected with
errors or not, the UE1 sends a CRC-based ACK/NACK response on the HS-
DPCCH. Whilst waiting for the UE1 feedback, Node B takes the opportunity to send
blocks 2 and 3 so that several H-ARQ processes are active in parallel;
– from the procedure described in [TS 25.214, R5], the UE1 estimates and then
sends a report on the channel quality (CQI) to Node B by choosing a transport
format (modulation, code rate and Tx power offset) such that a target BLER is met.
For the next transport block sent, Node B determines the transport format according
to the recommended transport format and possibly on power control commands of
the associated DPCH;
– the UE keeps on monitoring exclusively the HS-SCCH used in the
immediately preceding sub-frame, looking for an indication that there is about to be
some data destined to it.
HS-PDSCH
HS-DPCCH
UE1
UE1
UE1, block 1 UE1, block 2
UE1, block 3 UE2, block 1 UE1, block 1 UE1, block 4 UE2, block 2
UE1 UE1 UE2 UE1 UE1 UE2
NACK + CQI ACK + CQI NACK + CQI ACK + CQI NACK + CQI ACK + CQI
UE1, block 3
ACK + CQI
UE1 HS-SCCH
HS-DPCCH
UE2
HS-PDSCH
HS-DPCCH
UE1
UE1
UE1, block 1 UE1, block 2
UE1, block 3 UE2, block 1 UE1, block 1 UE1, block 4 UE2, block 2
UE1 UE1 UE2 UE1 UE1 UE2
NACK + CQI ACK + CQI NACK + CQI ACK + CQI NACK + CQI ACK + CQI
UE1, block 3
ACK + CQI
UE1 HS-SCCH
HS-DPCCH
UE2

Figure 14.9. Example of H-ARQ process with two users