Você está na página 1de 8

Journal of Information & Computational Science 9: 16 (2012) 50135020

Available at http://www.joics.com

Subcarrier and Power Allocation in OFDM-based


Relaying System
Yaru FU, Qi ZHU
Key Lab on Wideband Wireless Communications and Sensor Network Technology of Ministry of
Education, Jiangsu Key Laboratory of Wireless Communications, Nanjing University of Posts and
Telecommunications, Nanjing 210003, China

Abstract
Subcarrier and power allocation in orthogonal frequency division multiplexing (OFDM) relaying system
has been regarded as a powerful method to improve system performance. A dual-hop OFDM relaying
system employing decode-and-forward (DF) relaying protocol is considered and a low-complexity subcarrier and power allocation (PA) scheme with four steps is proposed in this paper. First, the dual-hop
subcarriers are matched according to their actual strength, then we allocate the total transmit power
between source and relay with respect to their average channel attenuation of first and second hop.
Afterwards, we reformulate the original max-min problem into a standard maximization problem and
give the optimal PA at relay for a given source PA to maximize the instantaneous rate. Finally, the
sum transmit power of each given subcarrier pair is reallocated to further enhance the system capacity.
Simulation results show that compared with the uniform power allocation algorithms, the proposed
power allocation algorithms can achieve higher system capacity.
Keywords: Relay; Decode-and-forward (DF); Power Allocation; Subcarrier Matching

Introduction

Cooperative relaying strategies have become a major topic in the wireless research community
because they can acquire a large capacity increase with the lowest cost. According to the dierence
of the received signal, two main relay strategies have been identied: amplify-and-forward (AF)
and decode-and-forward (DF). AF means that the received signal is multiplied by a factor and
then retransmitted by the relay without performing any decoding. In contrast, the signal is
decoded at the relay and re-encoded for retransmission in the DF strategy [1] [2]. The power
allocation analysis in terms of AF and DF system was done by Zhang Qi et al. [3].

This work is supported by National Natural Science Foundation of China (61171094)National Science & Technology Key Project (2011ZX03001-006-02, 2011ZX03005-004-03) and Key Project of Jiangsu Provincial Natural
Science Foundation (BK2011027).

Corresponding author.
Email address: zhuqi@njupt.edu.cn (Qi ZHU).

15487741/ Copyright
December 1, 2012

2012 Binary Information Press

5014

Y. Fu et al. /Journal of Information & Computational Science 9: 16 (2012) 50135020

Orthogonal frequency division multiplexing (OFDM) is a mature technique to mitigate the


problems of frequency selective fading and inter-symbol interference by dividing the total bandwidth into several orthogonal subcarriers. Relaying for OFDM systems is considered theoretically
in [4], after that many papers have studied the power distribution methods in single hop OFDM
system by the water-lling approach. Yu Guan-ding gives a joint method by applying the Hungarian algorithms for subcarriers matching and water-lling approach for power allocation with very
high complexity which is prohibitive in real-time wireless communication system [5]. An iterative
optimization analysis of non-cooperative relaying systems is done by Ingmar Hammerstr
om [6],
which needs to prove the convergence of iteration.
In recent years more and more scholars focus on research of multi-hop multi-relay multidestination OFDM based relaying system. An optimal resource allocation scheme for wireless
multi-destination relay systems with OFDM modulation is proposed by N. Zhou et al. [7]. In [8],
end-to-end resource allocation in linear multi-hop networks is investigated. In [9], the relay-node
selection as well as power allocation is investigated for a single source, multi-relay and single destination system. All these complex systems can simply be regarded as combination of a number
of subsystems, each consisting of one source, one relay and one destination, and power allocation
of the system can be attributed to the single node power allocation. In this paper we focus on
the research of dual-hop OFDM-based relaying system.
The remainder of this paper is organized as follows. In next section the relaying OFDM system
model is introduced. Section III introduces our subcarrier matching and power allocation strategy.
Performance results and analysis are presented in section IV and conclusions are drawn in the
last section.

System Model

An OFDM based two-hop relaying system is considered where the source communicates with
the destination using a single regenerative (DF) relay. All nodes hold only one antenna. And
we assume that there is no direct transmission from source to relay that is the destination can
receive signals only from the relay because of high shadowing between source and destination.
The TDD transmission protocol is adopted which means that the communication between the
source and the destination covers two time slots. The source transmits an OFDM symbol through
the source-relay channel during the rst time slot. During the second time slot, the relay decodes
and re-encodes the signal, and retransmits it toward the destination over the relay-destination
channel. The destination decodes the signal based on the received signal from the relay. It is
assumed that all of the channel state information (CSI) at the transmitter is known. The block
diagram of joint subcarrier matching and power allocation is shown as Fig. 1 [1].
In the rst time slot, source node sends symbol with power psk1 to the relay node over the
subcarrier k1 . Meanwhile the relay retransmits the symbol after decoding and re-encoding with
prk2 over subcarrier k2 in the second time slot. The channel capacity Ck1 k2 of the communication
between source and destination with the half-duplex relay on the subcarrier k1 , k2 is therefore
given by
Ck1 k2 = min {Csk1 , Crk2 }
(1)
(
)
psk1 hsk1
W
log2 1 +
(2)
Csk1 =
2N
N0 W/N

Y. Fu et al. /Journal of Information & Computational Science 9: 16 (2012) 50135020

5015

Fig. 1: Block diagram of joint subcarrier matching and power allocation

Crk2

(
)
W
prk2 hrk2
=
log2 1 +
2N
N0 W/N

(3)

where the factor 1/2 is due to the two time slots used in the transmission, W is the available
bandwidth of the system, N is the number of the subcarriers, hsk1 and hrk2 is the corresponding
channel gain of the k1 th and k2 th subcarrier between source and relay, relay and destination
respectively. Csk1 and Crk2 is the channel capacity of the rst and the second hop. The noise
variance is denoted by N0 .

3
3.1

System Problem Formulation


Subcarriers matching method

In the previous sections, we have assumed that the symbols of the source transmitted over the
k1 th subcarrier are encoded by the relay and retransmitted over the k2 th subcarrier. A
higher performance will be achieved if the subcarriers of the two hops are matched according to
their actual strength.i.e, the source to relay channel with the highest channel gain is matched
with the relay to destination channel with the highest channel gain [10].

3.2

Optimization problem

As we know that the bits transmitted at the source can be reallocated to the subcarriers at the
relay in arbitrary way theoretically. But for simplication, we constraint that the bits transmitted
on a subcarrier over the source to relay channel can be reallocated to only one subcarrier over
the relay to destination channel. In order to explain this clearly the coecient is introduced,
which can only be either 1 or 0. For example k1 k2 = 1 represents that the subcarrier k1 over
the source to relay channel is matched to the subcarrier k2 over the relay to destination channel.
Naturally the optimization problem for the joint subcarrier matching and power allocation of this
paper can be given by [1]

max

k1 k2 ,psk1 ,prk2

k1 =1

Ck1 k2 =

max

k1 k2 ,psk1 ,prk2

k1 = 1

{
min Csk1 , k1 k2

k2 =1

}
Crk2

(4)

5016

Y. Fu et al. /Journal of Information & Computational Science 9: 16 (2012) 50135020

Subject to:

k1 =1
N

k2 =1
N

psk1 = P 1
prk2 = P2
psk1 +

k1 =1

prk2 = P

(5)

k2 =1

psk1 , prk2 0k1 , k2


N

k1 k2 = 1
k2 =1

k1 k2 = {0, 1} , k1 , k2
where P is the total system power constraint, P 1 , P2 is the source and relay power constraints.
Motivated by the fractional power allocation in [11], we allocate the total transmit power P
between source and relay by
{
B
P1 = A+B
P
(6)
A
P2 = A+B P
where

N
N
1
1
2
|hsk1 | , B =
|hrk2 |2
A=
N k =1
N k =1
1

3.3

(7)

The proposed power allocation scheme

From the subcarriers matching method discussed above the corresponding relationship k1 k2 between subcarriers of two hops is easily to get. For the sake of nding the optimal psk1 , prk2 which
maximize the throughput of the system, we rstly give out the equivalent channel gain of each
subcarrier pair based on k1 k2 , then reformulate the original max-min problem into a standard
maximization problem. The proposed PA scheme can be obtained by using optimization method
and the Karush-Kuhn-Tucker (KKT) conditions.
1) The equivalent channel gain
In order to get the optimal power allocation of the relay subcarriersthe knowledge of the equivalent channel gain of the paired subcarrier is necessary. And it is obvious that only when both
capacities are equated can the end-to-end capacity be maximized, that is Csk1 = Crk2 , so that
(
)
(
)
W
psk1 hsk1
W
prk2 hrk2
log2 1 +
log2 1 +
=
(8)
2N
N0 W/N
2N
N0 W/N
We assume psk1 + prk2 = pk1
C k1 k2

(
)
W
psk1 hsk1
W
=
log2 1 +
log2 1 +
=
2N
N0 W/N
2N

pk1 hrk2
h
hsk1 +hrk2 sk1

N0 W/N

(
)
W
p
h
k1
k
k
1
2
=
log2 1 +
2N
N0 W/N

The equivalent channel gain of the given matched subcarrier pair k1 k2 is


hk1 k2 =

hk1 hk2
hk1 + hk2

(9)

Y. Fu et al. /Journal of Information & Computational Science 9: 16 (2012) 50135020

The channel capacity of the subcarrier pair k1 k2 is


(
(
))
hk1 hk2
W
psk1 + prk2
Ck1 k2 =
log2 1 +
2N
hk1 + hk2
N0 W/N

5017

(10)

Then the original max-min problem (4) can be reformulated by


max

k1 k2 ,psk1 ,prk2

k1 =1

Ck1 k2

(
(
))
N

hk1 hk2
psk1 + prk2
W
log2 1 +
=
max
k1 k2 ,psk1 ,prk2
2N
hk1 + hk2
N0 W/N
k1 = 1

(11)

2) Power allocation of the source and relay node


We assume that the power allocation of source node is P1
(
)T
P1 P1 P1
P1
T
P1 = (ps1 , ps2 , ps3 , . . . . . . , psN ) =
, ,
,
N N N
N

(12)

Then the power allocation of relay by using the induced optimization method can be denoted
by the vector P2
P2 = (pr1 , pr2 , pr3 , . . . . . . , prN )T
(13)
By using Lagrange multiplier method, we can set up the Lagrange function as
( N
)
(
))
(
N

hk1 hk2
psk1 + prk2
prk2 P2 + P2
L (prk2 , , ) =
log 1 +

hk1 + hk2
N0 W/N
k =1
k =1

(14)

The derivative of the Lagrange with respect to prk2 is given by


L (prk2 , , )
=
prk2
1+

hk1 hk2
1
hk1 +hk2 N0 W/N
hk1 hk2
hk1 +hk2

psk1 +prk2
N0 W/N

) +

Using the Karush-Kuhn-Tucker (KKT) conditions we get the solution of the optimization problem as
hk1 hk2
1

hk1 +hk2 N0 W/N


) psk1
prk2 = h h (
(15)
k1 k2
1

hk +hk
N0 W/N
2

If let

hk1 hk2
1
hk1 +hk2 N0 W/N

= Hk1 hence prk2 =

1
Hk1

psk1

The power cannot be negative (prk2 0) so


[
]+
1
1
prk2 =

psk1
Hk1

(16)

where [X]+ = max(X, 0) and is the Lagrangian multiplier chosen to satisfy the overall transmit
power constranint of the relay node.
3) Power reallocation of the given subcarrier pairs

5018

Y. Fu et al. /Journal of Information & Computational Science 9: 16 (2012) 50135020

From the analysis above we can get the power allocation vectors P1 and P2 of the source node
and the relay node, then the sum transmit power values of each given subcarrier pair is known
separately, after this we use the optimal reallocation to further optimize the system performance.
Denoting the sum of the transmit power on the ith subcarrier pair by pi , we can get psi +pri = pi .
The nal power allocation values can be formulated as
{

psi =
pri =

hri
p
hsi +hri i
hsi
p
hsi +hri i

(17)

Simulation Results and Analysis

In our simulations the number of available subcarrier N = 128. We assume that all three nodes
are located in a line. The distance between source and destination is denoted by R (R = 1000m).
We use the Okumura-Hata model: L (d) = 137.74 + 35.22 lg d to measure path loss where d(km)
is the distance between two nodes. The distance between source and destination, source and
relay is denoted by Dsd and Dsr , respectively.
We consider frequency-selective channels which
P 1
can be dened in the time domain by h (t) = p=0 (p) (t pT /N ), where (p) is the complex
amplitude of the p - th path, P the number of channel taps and T the OFDM symbol interval.
We assume that all taps are subject to Rayleigh fading (p) CN (0, 1/P ). The additive white
Gaussian noiseAWGNvariance of each subcarrier is the same as 2 = 128dBm.

Fig. 2: P1 = P2 = P/2 without subcarrier Fig. 3: P1 = PB/(A+B), P2 = PA/(A+B)


matching
and without subcarrier matching
Fig.2 shows the case that source and relay have the same transmit power constraint P1 = P2 =
P/2, P = 30dBm. Uniform source and uniform relay means both subcarriers of source and relay
node has the same allocated power psk1 = prk1 = P/2/N . Uniform sourceuniform relay and
reallocation separate from the former method by using power redistribution (17). The uniform
source optimal relay and reallocation is the proposed power allocation of this paper. Fig.3 is
shown for the case that source and relay have the dierent transmit power constraint by using (6)
which can help enhance the system capacity to some extent. The system capacity of our proposed
scheme is the highest of all, because the power allocation of the second hop is the optimal and
the subsequent power redistribution can further increase the system capacity.

Y. Fu et al. /Journal of Information & Computational Science 9: 16 (2012) 50135020

5019

Fig. 4: P1 = P2 = P/2 and with subcarrier Fig. 5: P1 = PB/(A+B), P2 = PA/(A+B)


matching
and with subcarrier matching
Fig.4, is the performance of proposed algorithm with subcarrier matching and power constraint
of P1 + P2 = P. Fig.5 is the performance of the case with matching of subcarriers and also using
power allocation between source and relay node by (6).
From all the gs we can see that the scheme with subcarrier matching can achieve better system
capacity than these without matching, and all power allocation and subcarrier matching schemes
achieve their maximum average rate for the case, where the relay is located in the middle between
source and destination which can be explained by (8). For Dsr /Dsd < 0.5 the proposed PA scheme
achieves a higher average system capacity than the uniform PA at the relay. This is due to the
fact that in this area the attenuation between relay and destination is larger than the attenuation
between source and relay. Therefore, it is more important to allocate more power to the good
subcarriers of the second hop. But for Dsr /Dsd > 0.5 this method can only achieves a lower
increasing than the uniform PA at the relay. But the power redistribution can further increase
system capacity to some extent.

Fig. 6: The proposed method compared with uniform power allocation


Fig.6 shows the average system capacity versus the ratio of the distance between source node to
relay node and the source node to destination node while the total system power constraint P is
dierent from 32dbm to 34dbm. For every power constraint P, the performance of the proposed
method is better than these without power allocation this is because the system capacity is limited
by the transmit power of each node and the scheme we used in this paper can make full use of the

5020

Y. Fu et al. /Journal of Information & Computational Science 9: 16 (2012) 50135020

power resource eciently. The simulation result also shows that the proposed scheme achieves
higher total system capacity when the total power of system is small.

Conclusion

Both subcarrier matching and power allocation can improve the capacity of the OFDM based relay
system. In this paper, we propose a low-complexity subcarrier matching and power allocation
scheme to maximize channel capacity. We focus the optimal on the second hop of the system so
when the attenuation between relay and destination is larger which is mean that the relay node
is near to the base station the mentioned power allocation and carriers matching method can
achieve larger rate gains. Hence, it will be useful for the cell-edged users in OFDM based relay
cellular networks in practice.

References
[1]

W. Wang, S. Yan, and S. Yang. Optimally joint subcarrier matching and power allocation in
OFDM multihop system. EURASIP Journal on Advances in Signal Processing, 2008: 1-8.

[2]

J. HU, X. CHEN. Performance Analysis of Amplify-and-Forward Non-Fixed Relays Cooperative


Network with Relay Selection. Journal of Computational Information Systems, 2010, 6 (9): 29012908.

[3]

Z. Qi, J.Zhang, and C. Shao. Power allocation for regenerative relay channel with Rayleigh fading.
Milan, Italy: IEEE Press, 2004: 1167-1171.

[4]

G. Li and J. Liu. On the Capacity of the Broadband Relay Networks. Asilomar, USA: Signals,
Systems and Computers, 2004: 1318-1322.

[5]

G. Yu, Z. Zhang, and Y. Chen. Power allocation for non-regenerative OFDM relaying channels
[C]Wireless Communications Networking and Mobile Computing, 2005. [S. l.]: IEEE, 2005: 185188.

[6]

I.Hammerstr
om and A. Wittneben. On the optimal power allocation for nonregenerative OFDM
relay links. IEEE 2006 International Conference Communication, 10: 4463-4468.

[7]

N. Zhou, X. Zhu, Y. Huang, H. Lin. Optimal resource allocation for orthogonal frequency division
multiplexing-based multi-destination relay systems. IET Communications, 2011, 5(14): 2075-2081.

[8]

X. Zhang, W. Jiao, and M. Tao. End-to-End resource allocation in OFDM based linear multi-hop
networks. IEEE Transaction Communication, 2009, 57(10): 3034-3044.

[9]

Reza Hoshyar, Mehrdad Shariat, Rahim Tafazolli. Subcarrier and power allocation with multiple
power constraints in OFDMA systems. IEEE Communication letters, 2010, 14(7): 644-646.

[10] X. Zhang, Y. Gong. Adaptive power allocation for multi-hop OFDM relaying systems. Singapore:
Information, Communications & Signal Processing, 2007: 1-5.
[11] M. O. Hasna and M-S. Alouini. Optimal power allocation for re-layed transmissions over rayleighfading channels. IEEE Transaction Wireless Communication, 2004, 3: 1999-2004.

Você também pode gostar