Optimum Power Allocation for Fading Relay Channels

Muhammad Mehboob Fareed and Murat Uysal

Department of Electrical and Computer Engineering
University of Waterloo, Waterloo N2L 3G1, ON, Canada.
mmfareed, muysal@ece.uwaterloo.ca
AbstractOptimum power allocation is a key technique to realize
the full potentials of relay-assisted transmission promised by the recent
information-theoretic results. In this paper, we present a comprehensive
framework for power allocation problem in a single-relay scenario
taking into account the effect of relay location. In particular, we aim
to answer the two fundamental questions: Q1) How should the overall
transmit power be shared between broadcasting and relaying phases?;
Q2) In the relaying phase, how much power should be allocated to
relay-to-destination and source-to-destination links? The power allocation
problem is formulated to minimize a union bound on the bit error rate
(BER) performance assuming amplify-and-forward (AaF) relaying. We
consider three TDMA-based cooperation protocols which correspond
to distributed implementations of MIMO (multi-input-multi-output),
SIMO (single-input-multi-output), and MISO (multi-input-single-output)
schemes. Optimized protocols demonstrate signicant performance gains
over their original versions which assume equal sharing of overall
transmit power between the source and relay terminals as well as
between broadcasting and relaying phases. It is observed that optimized
virtual (distributed) antenna congurations are able to demonstrate a
BER performance as close as 0.4 dB within their counterpart co-located
antenna congurations.
I. INTRODUCTION
User cooperation, also known as cooperative diversity [see 1
and the references there in], exploits the broadcast nature of wire-
less transmission and creates a virtual (distributed) antenna array
through cooperating nodes to extract spatial diversity advantages.
In [1], Nabar et. al. establish a unied framework of TDMA-
based cooperation protocols for single-relay wireless networks. They
quantify achievable performance gains for distributed schemes in
an analogy to conventional co-located multiantenna congurations.
Specically, they consider three protocols named Protocol I, Protocol
II, and Protocol III which correspond to MIMO, SIMO, and MISO
schemes, respectively
1
. In [1], the overall transmit power is supposed
to be uniformly allocated among the source and relay terminals.
The performance of cooperative communication schemes can be
substantially improved by optimally distributing the power among
cooperating nodes. Power allocation methods are formulated for
close-loop cooperative schemes in [2] and [3] where channel state
information (CSI) is assumed to be available at the transmitter side.
In [4], Hasna and Alouini investigate the optimal power allocation
problem for an open-loop transmission scheme (i.e., CSI information
available only at the receiver side), however their results are mainly
restricted to multi-hop systems without diversity advantages. In [5],
Yindi and Hassibi derive an upper bound on the pairwise error
probability (PEP) assuming Protocol III and AaF relaying for a large
number of relays and minimize PEP bound to formulate optimal
power allocation method. Their allocation method suggests that the
source uses half the total power and the relays share the other half
1
The use of term MIMO for Protocol I [1] is somehow misleading because
the diversity order of Protocol I for the considered single-relay scenario, as
shown in the same paper, is limited by 2 achieving the same order as Protocol
II and Protocol III. It might be more appropriate to consider Protocol I as an
enhanced version of Protocol II with additional coding gains (c.f., Section V
for our discussion on performance comparison of protocols).
fairly. For single-relay case, this simply reduces to equal power
allocation. It should be emphasized that this conclusion is a result of
their implicit underlying assumption that relays are located halfway
between source and destination terminals. In [6], Deng et.al. adopt
average signal-to-noise ratio (SNR) and outage performance as the
optimization metrics and investigate the power allocation problem for
Protocol II with AaF relaying. Their proposed method maximizes the
sum and product, respectively, of the SNRs in the direct and relaying
link and results in improved outage probability performance.
In this paper, we present a comprehensive framework for power
allocation problem in open-loop single-relay networks with AaF
relaying. Our work considers Protocols I, II and III of [1] which
effectively cover distributed MIMO, SIMO and SIMO congurations.
Further taking into account the effect of relay location, we attempt
to answer the following fundamental questions:
Q1) How should overall transmit power be shared between broad-
casting and relaying phases?
Q2) How much power should be allocated to relay-to-destination
and source-to-destination links
2
in the relaying phase?
For each considered protocol, we propose optimal power allocation
methods based on the minimization of a union bound on the BER per-
formance. Optimized protocols demonstrate signicant performance
gains over their original versions which assume equal sharing of
overall transmit power between broadcasting and relaying phases and
equal sharing of available power in the relaying phase between relay-
to-destination and source-to-destination links. We further demonstrate
that earlier results in this area such as [5] and [6] are special cases
of our more general results.
Notation: (.)

denote conjugate, (.)

T
transpose, (.)
H
and Hermitian
transpose operations, respectively. E(.) denotes expectation.
II. TRANSMISSION MODEL
We consider a single-relay scenario where terminals operate in
half-duplex mode and are equipped with single transmit and receive
antennas. As illustrated in Fig.1, three nodes source (S), relay (R),
and destination (D) are assumed to be located in a two-dimensional
plane where dSD, dSR, and dRD denote the distances of source-
to-destination (S D), source-to-relay (S R), and relay-to-
destination (RD) links, respectively and is the angle between
lines SR and RD. To incorporate the effect of relay geometry in
our model, we assume that the channel is composed of long-term free-
space path loss and short-term Rayleigh fading. The path loss is c/d

,
where c is a constant that depends on the propagation environment, d
is the propagation distance, and is path loss coefcient. Assuming
the path loss between S D to be unity, the relative gain of S R
and R D links are dened, respectively, as GSR = (dSD/dSR)

and GRD = (dRD/dSR)

, which can be further related to each

other through law of cosines. We assume M-PSK (phase shift keying)
2
The presence of source-to-destination transmission in the relaying phase
depends on the cooperation protocol. Therefore, Q2 is irrelevant in Protocol
II.
1-4244-0264-6/07/$25.00 2007 IEEE
R
S D
T
SD
d
SR
d
RD
d
Fig. 1. Relay-assisted transmission model.
modulation with normalized unit energy for the signals. We consider
AaF relaying where the relay terminal retransmits a scaled version
of the received signal without any attempt to decode it. The received
signals at the destination terminal during the four time slots (two
time slots for Protocol II) can be written as
r = hX+n, (1)
where h is the channel matrix dened as h =
_
hSD hSRhRD

.
hSR, hRD, and hSD denote the zero-mean complex Gaussian fading
coefcients with variances 0.5 per dimension, leading to the Rayleigh
fading channel assumption. In (1), n represents the additive noise
and consists of independent samples of zero-mean complex Gaussian
random variables with variance N0/2 per dimension. X is the
codeword matrix and is given by
X =
_
A0Ex1

A2Ex2

A0Ex2

A2Ex

1
0

A1Ex1 0

A1Ex

2
_
(2)
X =
_
B0Ex 0
0

B1Ex
_
(3)
X =
_
A2Ex2

A2Ex

A1Ex1

A1Ex

2
_
(4)
for Protocol I, Protocol II, and Protocol III, respectively, assuming
that Protocols I and III build upon distributed implementation of
Alamouti code [1]. In (2) and (4), A0, A1 and A2 are dened as A0 =
2KT , A1 = AN1/
_
AD +|hRD|
2
_
, A2 = AN2/
_
AD +|hRD|
2
_
where
AN1 = 2GSRKT , AN2 =
KS (1 + 2GSRKT SNR)
GRD (1 KS) SNR
,
AD =
(1 + 2GSRKT SNR)
2GRD (1 KT ) (1 KS) SNR
with SNR = E/N0. In (3), B0 and B1 are dened as B0 = 2KT
and B1 = BN/
_
BD +|hRD|
2
_
where
BN = 2GSRKT , BD =
1 + 2GSRKT SNR
2GRD(1 KT )SNR
.
In the above, KT and KS (0 KT , KS 1 )
3
are power allocation
parameters. KT is the fraction of power which is reserved for the
source terminals use in the rst time slot, i.e., broadcasting phase.
Remaining fraction 1 KT is used in the second time slot, i.e.
relaying phase, jointly by the source and relay terminals. In the
relaying phaseKS denotes the fraction of available power reserved for
the source terminals use while 1KS is used by the relay terminal.
In our model, the total power used by both source and relay terminals
is proportional to 2E during two time slots.
III. UNION BOUND ON THE BER PERFORMANCE
We consider BER performance as our objective function for power
allocation problem under consideration. A union bound on the BER
3
Only K
T
is relevant for Protocol II optimization.
for coded systems is given by well-known union bound
P
b

1
n

X
p (X)

X=

X
q
_
X

X
_
P
_
X

X
_
, (5)
where p (X) is the probability that codeword X is transmitted,
q(X

X) is the number of information bit errors in choosing
another codeword

X instead of the original one, n is the number
of information bits per transmission and P(X

X) is the PEP. A
Chernoff bound on the (conditional) PEP is given by
P
_
X

X|h
_
e
d
2
(X,

X)
4N
0
)
, (6)
where the Euclidean distance between X and

X is d
2
(X,

X) =
h
H
h
H
with = (X

X)(X

X)
H
. Recalling the denitions
of X in (2), (3), (4) for different protocols and carrying out the ex-
pectation with respect to h, we obtain PEP expressions for Protocols
I, II and III in the following:
1) PEP for Protocol I: The codeword matrix for Protocol I is
given by (2). Replacing this in (6), we have
P
_
X

X|h
_
e
{
SNR
4

1[(A
0
+A
2
)|h
SD
|
2
+A
1
|h
SR
|
2
|h
RD
|
2
]}
,
(7)
with 1 = |x1 x1|
2
+ |x2 x2|
2
. Averaging (7) with respect to
|hSR|
2
and |hSD|
2
which follow an exponential distribution, we
obtain
P
_
X

X

hRD
_
1
_
1 +
1
|hRD|
2
+1
+
1
|hRD|
2
+1
_
(8)
where
1 =
1
_
1 +
SNR
4
AN11
_ _
1 +
SNR
4
A01
_ ,
1 =
_
AD
_
1 +
SNR
4
A01
_
+
SNR
4
AN21

1 +
SNR
4
A0
,
1 =
(AD 1)
2
1 1
, 1 =
(AD 1)
2
(1 1)
,
1 =
AD
1 +A0AN11
SNR
4
.
Carrying out an expectation of (8) with respect to |hRD|
2
, we obtain
P
_
X

X
_
1 (1) =1[1 +1 exp (1) (0, 1)
+1 exp (1) (0, 1)] (9)
where (., .) denotes the incomplete gamma function [7].
2) PEP for Protocol II: The codeword matrix for Protocol II
is given by (3). Replacing this in (6) and averaging the resulting
expression with respect to |hSR|
2
and |hSD|
2
, we have
P
_
X

X|hRD
_
2
_
1 +
2
|hRD|
2
+2
_
, (10)
where
2 =
1
_
1 +
SNR
4
B02
_ _
1 +
SNR
4
BN2
_ ,
2 =
BD
1 +
SNR
4
BN2
, 2 = BD 2, 2 = |x x|
2
By averaging (10) over |hRD|
2
, we obtain the nal form for PEP as
P
_
X

X
_
2 (2) =2 [1 +2 exp (2) (0, 2)] . (11)
Relation to [6]: As discussed in Section I, there has been some
earlier work which investigates power allocation for Protocol II. In
particular, Deng and Haimovich [6] propose two optimal allocation
methods: One depends on the maximization of the sum of the average
SNRs in the direct and relaying link. The other aims to minimize the
outage probability which eventually turns out to be maximization of
product of SNRs in these links. Under the assumptions of [6], (non-
fading R D channel, i.e., hRD = 1), our PEP expression given
by (10) reduces to
P
_
X

X
_

__
1 +
SNRD2
4
__
1 +
SNRR2
4
__
1
, (12)
where the SNRs in direct link and relaying path are, respectively,
dened as SNRD = B0SNR, and SNRR = B1SNR. Expanding
(12), we have
P
_
X

X
_
[1 + (SNRD +SNRR)
2
4
+SNRDSNRR(
2
4
)
2
]
1
. (13)
Under high SNR assumption, the rst term can be further omitted.
It is clear from (13) that, to minimize the resulting PEP, we need to
simultaneously maximize the sum and product of the SNRs in the
direct and relaying links. It is now obvious from above derivation
that criteria discussed in [6] are two special cases of our general
formula. We should further emphasize that our derived PEP in its
original form given by (11) has been derived under fading R D
channel unlike the simplifying non-fading assumption of [6].
3) PEP for Protocol III: The codeword matrix for Protocol III
is given by (4). Replacing this in (6) and averaging the resulting
expression with respect to |hSR|
2
and |hSD|
2
, we have
P
_
X

X

hRD
_
3
_
1 +
3
|hRD|
2
+3
+
3
|hRD|
2
+3
_
,
(14)
where 3, 3, 3, 3 and 3 are dened as
3 =
1
1 +
SNR
4
AN13
,
3 = AD +
SNR
4
AN23, 3 =
AD
1 +
SNR
4
AN13
,
3 =
(AD 3)
2
3 3
, 3 =
(AD 3)
2
3 3
,
and 3 = |x1 x1|
2
+|x2 x2|
2 4
. By averaging (14) over |hRD|
2
,
we obtain the nal form for PEP as
P
_
X,

X
_
3 (1) =3[1 +3 exp (3) (0, 3)
+3 exp (3) (0, 3)]. (15)
Relation to [5]: In [5] Jing and Hassibi derive an upper bound
on PEP for distributed space-time codes. Based on PEP results they
further show that for a xed total transmit power across the entire
network, the optimal power allocation is for the transmitter to expend
half the power and for the relays to collectively expend the other half.
It should be noted that the lack of path loss considerations in their
model, therefore ignoring the geometrical gains of relay location,
has resulted in such a conclusion. The reported result in [5] is a
special case of our derived results when the nodes are placed on the
corners of an equilateral triangle. Under similar assumptions, i.e.,
GSR/GRD = 0dB and = /3, our more general result derived
for Protocol III in (15) reduces to the approximate expression given
in [5 - p.3527].
IV. OPTIMUM POWER ALLOCATION
As noted in Section III, the objective function in our optimization
problem is the union bound on BER. Replacing PEP expressions
given by (9), (11), (15), respectively, for Protocols I, II and III, in
4
Both Protocols I and III are built upon Alamouti code. Therefore,
1
=
3
TABLE I
POWER ALLOCATION PARAMETERS FOR DISTRIBUTED ALAMOUTI CODE.
(a) Protocol I
SNR
G
SR
/G
RD
-30dB 0dB 30dB
[dB] K
T
K
S
K
T
K
S
K
T
K
S
5 0.9535 0.0000 0.6648 0.0000 0.6336 0.0000
10 0.9516 0.0000 0.6501 0.0000 0.6153 0.0000
15 0.9503 0.0000 0.6417 0.0000 0.5812 0.0586
20 0.9493 0.0000 0.6358 0.0000 0.3680 0.3652
25 0.9486 0.0000 0.6315 0.0000 0.3682 0.3583
30 0.9479 0.0000 0.6280 0.0000 0.3608 0.3599
(b) Protocol II
SNR
G
SR
/G
RD
-30dB 0dB 30dB
[dB] K
T
K
T
K
T
5 0.9551 0.6728 0.6466
10 0.9530 0.6580 0.6267
15 0.9517 0.6493 0.6156
20 0.9507 0.6432 0.6081
25 0.9499 0.6385 0.6025
30 0.9492 0.6348 0.5982
(c) Protocol III
SNR
G
SR
/G
RD
-30dB 0dB 30dB
[dB] K
T
K
S
K
T
K
S
K
T
K
S
5 0.9276 0.0722 0.2697 0.7707 0.0236 0.6532
10 0.4765 0.9984 0.2660 0.7565 0.0223 0.6034
15 0.4780 0.9984 0.2631 0.7484 0.0433 0.6066
20 0.4787 0.9984 0.2609 0.7427 0.0464 0.5999
25 0.4792 0.9984 0.2590 0.7384 0.0490 0.5950
30 0.4795 0.9983 0.2575 0.7349 0.0515 0.5911
the BER bound given by (5), we obtain the objective functions to
be used for power allocation. The specic form of BER expressions
depends on the modulation scheme and underlying code. We need to
minimize the resulting BER expressions with respect to the power
allocation parameters KT and KS (0 KT , KS 1). It can
be easily shown that the BER expressions are convex functions
with respect to optimization parameters. Convexity of the functions
under consideration guarantees that local minimum found through
optimization will be indeed a global minimum. Any commercially
available optimization software can be used for this purpose. Here
we use Matlab optimization toolbox command fmincon designed
to nd the minimum of a given constrained nonlinear multivariable
function. It should be also emphasized that this problem needs not to
be solved in real-time for practical systems, because the optimization
does not depend on the instantaneous channel information or the input
data. This means that the optimum power allocation values can be
obtained apriori for given values of operating signal-to-noise ratio
and propagation parameters, and can be stored for use as a lookup
table in practical implementation.
In Table I, we present optimum values of KT and KS for various
values of GSR/GRD which reects the effect of relay location. More
negative this ratio is, more closely relay is placed to destination. On
the other hand, positive values of this ratio indicate that relay is more
close to source. The particular case of GSR/GRD = 0dB indicates
that relay is equidistant from source and destination terminals. For
the rest of our discussion, we assume = , i.e., relay lying on the
line connecting source and destination unless specied otherwise.
For Protocol I, we observe from Table I(a) that
When the relay is close to destination, optimum values of KT
are 0.95, and those of are KS 0. These values indicate that
it is better to spend most of power in broadcasting phase, and
in the relaying phase available power (i.e. 1 KT ) should be
dedicated to the relay terminal.
When relay is equidistant from source and destination, the
optimum value of KT is 2/3 which means that 66% of power
should be spent in the broadcasting phase. The optimum value
of KS is still 0 which indicates that all available power should
be dedicated to the relay terminal in the relaying phase.
When relay is close to source and system is operating in low
SNR region (0-10 dB), optimum values of KT and KS are the
same as in previous case, but in higher SNR region (>10 dB)
the optimum value of KS increases with increasing SNR while
that of KT decreases.
For Protocol II, we observe from Table I(b) that
When relay is close to destination, 95% of power should be
used in broadcasting phase.
When relay is equidistant or close to source, 66% of power is
required by the source to achieve optimum performance. This
fact can be further conrmed for the simple case of non-fading
R D link and GSR = GRD = GSD = 1, by setting the
derivative of PEP given by (10) equal to zero. The minimum
value of BER is readily obtained at KT = 2/3 + 1/6E.
For Protocol III, we observe from Table I(c) that
Optimum value of KS is 1 for negative values of ratio. This
is in contrast with small values of KS observed for Protocol I.
Here, it should be noted that Protocol I is able to guarantee a
diversity order of two even with equal power allocation owing
to the existence of strong S D link in the relaying phase.
However, the diversity order of Protocol III with equal power
allocation reduces to one for scenarios where relay is close to
destination. Such large values of KS in the optimized Protocol
III aim to balance the S D and R D links so that diversity
order of two can be extracted, guaranteeing the full diversity.
For equidistant nodes, optimum value of KT is1/4 and that of
KS is 3/4.
When relay is close to source, KT is too small and KS is 2/3.
In Fig.2, we demonstrate performance gains in power efciency
(as predicted by the derived PEP expressions) achieved by optimum
power allocation (OPA) over equal power allocation (EPA) for a
target BER of 10
3
assuming QPSK modulation. The performance
gains are presented as a function of GSR/GRD. In Fig.2(a) given
for Protocol I, we observe performance improvements of 0.4dB and
0.3dB at GSR/GRD = 0dB and GSR/GRD = 30dB, respectively.
Advantages of OPA are more pronounced for negative values of
GSR/GRD. For example, an improvement of 2.5dB is observed
for GSR/GRD = 30dB. It is clear from this gure that although
power optimization helps in all cases, it is more rewarding in
scenarios where relay is close to destination. In Fig.2(b) given for
Protocol II, we observe performance improvements up to 2.6dB for
negative values of GSR/GRD. For positive values, it is observed
that OPA and EPA performance curves converge to each other. In
Fig.2(c) given for Protocol III, we observe signicant performance
improvements for both negative and positive GSR/GRD values. In
particular, the performance improvements are 8.4dB and 2.9dB at
GSR/GRD = 30dB and GSR/GRD = 30dB, respectively. The
change in characteristic behaviour of Protocol III in comparison to
those of Protocols I and II should be also noted. This is actually not
unexpected; recall that Protocol III is the distributed implementation
of MISO and implements transmit diversity, so it is expected to
perform good when relay is close to source mimicking a virtual
transmit antenna array. Protocol II is the distributed implementation
of SIMO and implements receive diversity, so it is expected to
80 60 40 20 0 20 40 60 80
14
16
18
20
22
Fig.2.a Protocol I


80 60 40 20 0 20 40 60 80
12
14
16
18
20
S
N
R

R
e
q
u
i
r
e
d

f
o
r

B
E
R

=

1
0

3
































80 60 40 20 0 20 40 60 80
20
25
30
35
40
G
SR
/G
RD
[dB]


EPA
OPA
EPA
OPA
EPA
OPA
Fig.2.b Protocol II
Fig.2.c Protocol III
Fig. 2. SNR required to achieve BER of 10
3
for Protocol I, II and III.
perform good when relay is close to destination mimicking a virtual
receive antenna array. Protocol I is a combination of both Protocol II
and Protocol III. It is observed from our results that the advantages
of receive diversity are dominating in this hybrid version.
V. SIMULATION RESULTS
To further conrm the performance gains of OPA promised by
the PEP expressions, we have conducted an extensive Monte Carlo
simulation study to compare the BER performance of the considered
protocols with EPA and OPA. Our simulation results are presented
in Fig. 3 where we assume QPSK modulation and = . In Fig. 3
illustrated for Protocol I, we observe performance improvements of
2.5dB, 0.4dB, and 0.29dB at a target BER of 10
3
for GSR/GRD =
30dB, 0dB and 30dB respectively. These are similar to performance
gains predicted for Protocol I through our PEP expressions. Similar
conrmation holds for the other two protocols, but not included here
due to space limitations.
Figs. 4 and 5 present a performance comparison of three protocols
with EPA and OPA assuming GSR/GRD = 30dB. As benchmarks,
we include the performance of non-cooperative direct transmission,
(i.e., no relaying), Alamouti code, and maximal ratio combining
(MRC) with two co-located antennas. It should be noted that the
inclusion of co-located antenna scenarios help us to demonstrate how
close the virtual antenna implementations can come to their co-
located counterparts. The performance of Protocol II and Protocol
III, which are distributed SIMO and MISO congurations, are lower
bounded by that of MRC and Alamouti code, respectively. To make a
fair comparison between cooperative and direct transmission schemes
0 5 10 15 20 25
10
6
10
5
10
4
10
3
10
2
10
1
10
0
SNR [dB]
B
E
R
EPA G
SR
/G
RD
= 30 dB
OPA G
SR
/G
RD
= 30 dB
EPA G
SR
/G
RD
= 0 dB
OPA G
SR
/G
RD
= 0 dB
EPA G
SR
/G
RD
= 30 dB
OPA G
SR
/G
RD
= 30 dB
Fig. 3. Simulated BER performance of Protocol I for different values of
G
SR
/G
RD
.
which achieve rates of 1/2 and 1 respectively, direct transmission and
co-located antenna scenarios are simulated with BPSK. We observe
that Protocol I and Protocol II have a similar performance and
outperform Protocol III whose diversity is limited to one for the
considered GSR/GRD = 30dB. Suffering severely from the low
SNR in source-to-relay link, Protocol III is even outperformed by
direct transmission under the same throughput constraint and is far
inferior to its co-located counterpart, i.e., Alamouti scheme. From
Fig.5, we observe that optimized version of Protocol III achieves
a diversity order of two and outperforms direct transmission after
SNR = 8dB. Unlike Protocol III, Protocols I and II guarantee
full diversity under EPA assumption, however their performance is
still 3dB away from the MRC performance. Under OPA assumption,
Protocol II is able to operate just 0.4dB away from the MRC bound.
Another interesting observation from Figs. 4 and 5 is that Protocol I
and II provide nearly identical performance in both EPA and OPA.
It should be emphasized that this observation is made considering
the use of Protocol I in conjunction with the underlying Alamouti
code. With a more powerful space-time code such as Golden code
[8], it is expected that Protocol I will be able to outperform Protocol
II providing additional coding gains.
VI. CONCLUSION
In this paper, we have investigated optimum power allocation
methods for a single-relay scenario with AaF relaying for Protocols
I, II, III of [1] which correspond to distributed implementations of
MIMO, MISO and SIMO communications. For each cooperation
protocol, we have derived union bounds on BER performance which
are then used to optimally allocate the power among cooperating
nodes in broadcasting and relaying phases. In comparison to their
original counterparts, optimized protocols demonstrate signicant
performance gains depending on the relay geometry. Our results
further provide a detailed comparison among Protocols I, II and
III which give insight into the performance of these protocols
incorporating the effects of relay location and power allocation.
REFERENCES
[1] R. U. Nabar, H. B olcskei, and F. W. Kneub uhler, Fading relay channels:
Performance limits and space-time signal design, IEEE J. Sel. Areas
Commun., vol. 22, no. 6, pp. 1099-1109, August 2004.
0 5 10 15 20 25
10
6
10
5
10
4
10
3
10
2
10
1
10
0
SNR [dB]
B
E
R
Protocol I
Protocol II
Protocol III
Noncooperative
Alamouti
MRC
Fig. 4. The performance comparison of Protocol I, Protocol II and Protocol
III with equal power allocation (G
SR
/G
RD
= 30dB).
0 5 10 15 20 25
10
6
10
5
10
4
10
3
10
2
10
1
10
0
SNR [dB]
B
E
R
Protocol I
Protocol II
Protocol III
Noncooperative
Colocated Alamouti
Colocated MRC
Fig. 5. The performance comparison of optimized versions of Protocol I,
Protocol II and Protocol III (G
SR
/G
RD
= 30dB).
[2] A. Host-Madsen and J. Zhang, Capacity Bounds and Power allocation
in Wireless Relay Channel, IEEE Trans. Inf. Theory, vol. 51, no. 6, pp.
2020-2040, June 2005.
[3] N. Ahmed and B. Aazhang, Outage minimization with limited feedback
for the fading relay channel, IEEE Trans. Commun., vol. 54, no. 4, pp.
659-669 April 2006
[4] M. O. Hasna and M.-S. Alouini, Optimal power allocation for relayed
transmissions over Rayleigh-fading channels, IEEE Trans. Wireless
Commun., vol. 3, no. 6, pp. 1999-2004, November 2004.
[5] Y. Jing and B. Hassibi, Distributed space-time coding in wireless relay
networks, IEEE Trans. Wireless Commun., vol. 5, no. 12, pp. 3524-
3536, December 2006.
[6] X. Deng, and A. M. Haimovich, Power Allocation for Cooperative
Relaying in Wireless Networks, IEEE Commun. Lett., vol. 9, no. 11,
pp. 994-996, November 2005.
[7] I. S. Gradshteyn and I. M. Ryzhik, Table of Integrals, Series and
Products, Academic Press, 5th edition, 1994.
[8] J.-C. Belore, G. Rekaya, and E. Viterbo, The Golden code: A 2 2
full-rate space-time code with non-vanishing determinants, IEEE Trans.
Inf. Theory, vol. 51, no. 4, pp. 1432-1436, April 2005.