View
2
Download
0
Category
Preview:
Citation preview
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Cooperative Wireless Communication
Yohannes Alemseged
yohannes@tugraz.at
Signal Processing and Speech Communication Laboratory
www.spsc.tugraz.at
Graz University of Technology, Austria
Advanced Signal Processing Seminar 24th May. 2006
Cooperative Wireless Communication – p. 1/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Outline
� Motivation
� Cooperative Communication
� Background
� Amplify-And-Forward
� Detect-And-Forward
� Coded Cooperation
� Performance
� Multiple Access
� Challenges
Cooperative Wireless Communication – p. 2/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Motivation
� The mobile wireless channel suffers from fading,
Coping mechanism =⇒ generate diversity by sending
independent copies of the same signal
� Multiple antenna is required for transmit diversity
� Wireless devices are limited by size or hardware complexity
to one antenna
� Wireless antennas are omnidirectional
� In case of Ad-hoc wireless com., no fixed infrastructure
Cooperative Wireless Communication – p. 3/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Cooperative Communication -Background
� Each wireless user is assumedto act as cooperativeagent for another user
� Allow single-antenna mobilesto reap some ofthe benefits of MIMO systems
� Possible improvement both inthroughput and transmit power
Source
Relay
Independent fading paths
Figure 1: Cooperative communication
Cooperative Wireless Communication – p. 4/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Background cont’d
� Basic idea originally by van der Meulen(1969), Cover and El Gamal (1979)
� Capacity of AWGN degraded relaychannel (achievable rate)
� Y1 = X1 + Z1, z1 ∼ N(0, σ1), andY = (X2 + Z2) + Y1, z2 ∼ N(0, σ2)
� C∗ = max0≤α≤1
min{C(P1+P2+2√
αP1P2
N1+N2
)
, C(αP1
N1
)}
where α = 1 − α andC(x) = 1
2log(1 + x).
� Cooperative scheme
� Diversity and fading channel
� Fixed total system resource
A C
B
MultiAccess
Broadcast
Channel 3
Y
X1
X2Y1
� �� � �� �
��z2
� � �
� � � z 1
Figure 2: The relay channel
Cooperative Wireless Communication – p. 5/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Amplify-And-Forward
� How it works
� Amplify and retransmit
� Decision is made at the base station
� Two independently faded versions of same signal,
leading to diversity order of two
� Inter-user channel coefficients are assumed to be known
at the base station to do the optimal decoding
Cooperative Wireless Communication – p. 6/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Amplify-And-Forward cont’d
2
4
Y0,BC
Y0,MA
3
5 =
2
4
h10
h12βh20
3
5 X1 +
2
4
0 1 0
h20β 0 1
3
5
2
6
6
4
Z1
Z0
Z0
3
7
7
5
Source
Relay
Destination
Z0
h10
h12
Z1X2
h20
X1
Figure 3: Amplify-and-Forward [Laneman et al.]
Cooperative Wireless Communication – p. 7/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Amplify-And-Forward cont’d
� High SNR(Diversity-Multiplexingtrade-off)
� Low SNR (Outage Ca-pacity)
Figure 4: Diversity vs. Multiplexing (Laneman et al.)
Cooperative Wireless Communication – p. 8/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Detect-And-Forward
� User detects partner’s bits and then retransmits
the detected bits
� Partners assignment (via pairs)
� CDMA implementation (two users), users’ codes are orthogonal
and channel coherence time is L (L=3)
X1(t) = a1b(1)1 c1(t), a1b
(2)1 c1(t), a1b
(3)1 c1(t)
X2(t) = a2b(1)2 c2(t)︸ ︷︷ ︸
Period 1
, a2b(2)2 c2(t)︸ ︷︷ ︸
Period 2
, a2b(3)2 c2(t)︸ ︷︷ ︸
Period 3
(1)
where aj =√
Pj/Ts, Pj is user j’s power
and TS is symbol period
Cooperative Wireless Communication – p. 9/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Detect-And-Forward contd.
� b̂(i)j is estimate of j’s ith bit
� Power allocation (aij)
� Average power constraint
is maintained
� Cooperation in favorable
interuser channel
X1(t) =[a11b(1)1 c1(t), a12b
(2)1 c1(t),
a13b(2)1 c1(t) + a14b̂
(2)2 c2(t)]
X2(t) =[a21b(1)2 c2(t)︸ ︷︷ ︸
Period 1
, a22b(2)2 c2(t)︸ ︷︷ ︸
Period 2
,
a23b(2)2 c2(t) + a24b̂
(2)1 c1(t)]︸ ︷︷ ︸
Period 3
Cooperative Wireless Communication – p. 10/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Detect-And-Forward cont’d.
� Simplicity and adaptability to channel conditions
� The method fails in case of unsuccessful partner detection
� Hybrid decode-and-forward (to avoid error propagation)
� Cooperative mode for low inst. SNR
� Non cooperative mode for high inst. SNR
Cooperative Wireless Communication – p. 11/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Detect-And-Forward cont’d.
� Users send two new bits per threesymbols, isn’t this counter productive?
η = (1 − υ)CBSC(Q(
√SNR0
1 − υ))
� Negligible loss of throughput at lowSNR0
� Design tradeoff, Lnon−cvs. Lc
� Lc doesn’t have to be constant all thetime
Figure 5: Throughput vs. unused symbolperiods
Cooperative Wireless Communication – p. 12/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Coded Cooperation
� Cooperation =⇒ Channel coding
� Different portions of each user’s code wordare sent via independent fading path
� The divided source data blocks are augmentedby CRC
� eg. original codeword has N1 + N2 bits(puncturing) the original codeword
I First partition, valid codeword with N1 bitsRemaining N2 bits are the puncture bits
I In the first frame each user sends N1 bits andin the second frame partner’s 2nd code partition
Cooperative Wireless Communication – p. 13/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Coded Cooperation cont’d.
User 1
Frame 1 Frame 2
Base Station
Frame 2Frame 1
To Tx
No
Yes
check
RCPC
RCPC
decoder
CRC
User 2
Ownbits
PartnerReceived
CRC
Viterbi
� � � � ��
� � � � � � �
N1
� �
� � � � � ��
N2
� � �
N1
� � � � �
N2
� � � � �
N1
� � � � �
N2
� � � � �
Figure 6: Coded Cooperation [Nosratinia et al.]Cooperative Wireless Communication – p. 14/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Coded Cooperation cont’d.
� Users act independently in the second frame
� both users cooperate
� user 1 cooperate and user 2 doesn’t
� user 2 cooperate and user 1 doesn’t
� both users don’t cooperate
� Level of cooperation (in the eg. N2/N )
� Other channel codes can be employed
(block codes, convolutional codes, ...)
Cooperative Wireless Communication – p. 15/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Performance
� Case studied by[Sendonaris et al.]
� For equitablecomparison, coded baseline system of overall rate1/4 is used
� RCPC rate 1/2 forHybrid anddecode-and-forwardand amplify andforward
� 25% level ofcooperation for codedcooperation
Figure 7: Performance of cooperative signaling methods (-10dB inter-user SNR)
Cooperative Wireless Communication – p. 16/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Multiple Access
� Base station receives original and relayed transmissionsseparately (assumption)
� Separation in time (Different time slot)
� Hardware implication
� Separation in frequency (Different spectrum)
� Use of spreading code [Sendonaris et al.]
� Sufficient isolation is required
Cooperative Wireless Communication – p. 17/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
Challenges
� Loss of rate to the cooperating mobile
� Overall interference in the network
� Cooperation assignment and hand off
� Fairness of the system, "market force and QoS"
� Transmit and receive requirement on the mobiles
� User data has to be encrypted (security)
Cooperative Wireless Communication – p. 18/20
GRAZ UNIVERSITY OF TECHNOLOGY
Signal Processing and Speech Communications Lab
References
1 A. Nosratinia, T. E. Hunter and A. Hedayat, "Cooperative communication in
wireless networks", IEEE Communication Magazine, Vol. 42, Issue 10, Oct.2004, pp. 74-80.
2 T. M. Cover and A. A. E. Gamal, "Capacity Theorems for the Relay Channel",IEEE Trans. Info. Theory, Vol. 25, no 5, Sept. 1979, pp. 572-84.
3 A. Sendonaris, E. Erkip, and B. Aazhang, "User Cooperation Diversity Part I
and Part II," IEEE Trans. Commun., vol. 51, no. 11, Nov. 2003, pp. 1927-48.
4 J.N. Laneman, D.N.C. Tse, and G.W. Wornell , "Cooperative diversity inwireless networks: Efficient protocols and outage behavior" IEEE Transactions
on Information Theory, vol. 50, Issue 12, Dec. 2004, pp. 3062-3080.
Cooperative Wireless Communication – p. 19/20
Recommended