[ieee 2013 13th mediterranean microwave symposium (mms) - saida (2013.9.2-2013.9.5)] 2013 13th...
TRANSCRIPT
Improving Relay Selection in Cellular Networks
Michel Nahas, Zouhair El-Bazzal
School of Engineering
Lebanese International University
Beirut, Lebanon
{michel.nahas, zouhair.bazzal}@liu.edu.lb
Nadine Akkari
Faculty of Computing and Information Technology
King Abdulaziz University
Jeddah, Saudi Arabia
Abstract—To extend and improve the cell coverage, an
efficient solution is to place several relay antennas in the cell.
Several protocols for relay selection were proposed in the past
years. However, the asynchronism between the received signals
due to the different locations of the relays was not addressed. In
this paper, we propose to select the relays based on several
important criteria. The SNR (Signal to Noise Ratio) between the
relays and the mobile is first taken into consideration. Then, the
delay differences between the different signals arriving to a mobile
are computed. This proposed protocol will allow the receiver to
choose the relays that verify a certain SNR threshold and delay
conditions. Thus, the best relays in terms of SNR and delay are
selected for the communication between the base station and a
mobile. Moreover, the overall enhancement of the cooperative
system performance, in terms of signaling time and error rates, is
shown by numerical calculations and simulations.
I. INTRODUCTION
Cooperative communication uses multiple relays to
retransmit the source signal to the destination, which improves
both area coverage and quality of services [1]. This
cooperation forms a virtual MIMO (Multiple Input Multiple
Output) system, where the relays are located in different
geographical places to achieve a spatial diversity [2].
In a cellular system, relay antennas are distributed inside the
cell to help the Base Station (BS) to communicate with the
mobile nodes [3]. This low-cost solution extends the coverage
and improves the throughput and reliability of cellular
networks [4]. However, the spatial distribution of relays would
generate an asynchronous system due to the difference in local
oscillators, and different propagation and/or processing delays
of the relays [5].
Therefore, to enhance the communication performance in
cooperative cellular networks, a good selection of relays
should be carried out. This selection has to take into
consideration many parameters such as channel coefficients,
the position of the receiver relatively to the relays and the time
delay between the received signals. An efficient relay
selection protocol would choose relays that will enhance the
cellular communication performance [6] [7].
Several relay selection protocols exist in the literature. In
[8] the concept of BRS (Basic Relay Selection) protocol was
introduced. The authors also proposed several modifications to
this basic protocol, such as RSOD (Relay Selection On
Demand). This latter protocol decreases the energy
consumption, by allowing the receiver to decide whether
diversity is needed or not. Further saving in energy
consumption was achieved by RSER (Relay Selection with
Early Retreat) [8]. RSER protocol proposed that relays having
low Signal to Noise Ratio (SNR) values will retreat and will
not continue the selection process.
However, all the protocols presented above ignored the
asynchronism issue to simplify the relay selection procedure.
Recent studies have studied the effect of asynchronism in
relay networks and its influence on the performance gain of
the system [9] [10]. These latter works have proposed relay
selection protocols for asynchronous cooperative networks at
the expense of the diversity. In fact, to eliminate the effect of
asynchronism between several signals forwarded by the
relays, only one relay is chosen at the destination. Therefore,
relay selection protocols in [9] and [10] solved the problem of
asynchronism in cooperative networks but they only achieved
a diversity order of one.
Recently, an asynchronous relay selection protocol based
on the RSER protocol was proposed in [11] for distributed
networks (Ad hoc, Wireless sensors networks…). This latter
protocol eliminated successfully the effect of asynchronism,
while ensuring a high diversity order.
In this paper, we are going to design a protocol that selects the
relays in an asynchronous centralized network like cooperative
cellular networks. This protocol will complete the selection
process in few steps by enforcing several constraints
concerning the SNR and the delay of the different received
signal at the destination. After the process is completed, only
the relays that respond to the SNR and delay criteria are
selected. Moreover, the centralized relay selection protocol
will reduce the energy consumption and signal and complexity
by implementing an early retreat strategy for some relays.
The paper is organized as follows. In Section II, we present
the system model of the cooperative cellular network. The
new relay selection protocol for asynchronous cellular
communication is described in Section III. The derivation of
the signaling complexity of the selection process is proposed
in Section IV along with some simulation results. We
conclude in Section V.
II. SYSTEM MODEL
Let us consider a circular cell with a base station at its
center and N fixed-position relay nodes (see Fig. 1). The
relays can enhance the communication between the BS and the
mobile M in the cell. To achieve a diversity order d, the
number of relays participating successfully in the
communication with the BS must be equal to (d – 1) [5].
However, these relays should be selected carefully.
978-1-4673-5820-0/13/$31.00 ©2013 IEEE
Fig. 1. Cell with some relay nodes.
Actually, due to the different location of the relays, and
difference in process and propagation delays, each BS-relay-
mobile path will have a different time delay. Therefore, the
selection process should choose the best (d – 1) relays for the
in terms of SNR and relative time delays between the different
arriving signals.
We assume that the BS, relay nodes and mobile has each
one antenna. Therefore, in order to have diversity in the cell, a
Space Time Code (STC) should be implemented across the
transmitter and the relays. Several optimal STCs were
designed for different antennas configuration like Alamouti
code [12], Golden code [13] and TAST codes (Threaded
Algebraic Space-Time)… But due to asynchronism between
received signals, diversity will be lost for some values of
delays. It has been shown in [5] that a delay of one symbol
period � would destruct the STC structure and leads to a drop
in the diversity order. Thus, a critical delay difference between
any two received signals �������� � should be avoided.
III. CENTRALIZED RELAY SELECTION PROTOCOL
In this section, we will describe the relay selection protocol
designed to select relays having good SNR (higher
than �������), while avoiding critical delays ��������.
This protocol, that we will name Centralized Relay Selection
(CRS) protocol is conceived to eliminate the signals arriving
at the destination after a certain delay �������.
Without loss of generality, let us assume a downlink
communication system where the BS wants to communicate
with a mobile M inside the cell boundaries. Each of the N
relay available within the cell, has a different identifier ID.
First, the BS broadcasts a RTS (request-to-send) in the cell.
The RTS signal is captured by the mobile M and all the relays,
as shown in Fig. 2.
Then, the mobile will reply with a CTS (clear-to-send). This
message will be received by the BS and the neighboring relays
(Fig. 3). CTS signal may not reach the distant relays which
will no longer participate in this process.
Fig. 2. RTS transmitted by the Base Station.
The remaining relays calculate their respective SNR values by
using both RTS and CTS messages as in RSER protocol [8].
Then, each relay compares its calculated SNR to the SNR
threshold value. The relays with �������, which
means that they have bad channel gains, will retreat from the
relay selection procedure. Thus, the relays with bad channel
conditions are going to retreat as can be seen in Fig. 3.
Fig. 3. Mobile node sending CTS to BS and relays.
Subsequently, the BS sends another pilot message DTS
(delay– time symbol) received by all non-retreated relays and
the mobile (Fig. 4). The relays then forward the DTS message
to M. The mobile computes the delay between the direct signal
from the BS and the signals forwarded by the relays. The
relays with delay difference bigger than ������� or equal to
�������� , will be discarded by the mobile.
Finally, M will reply with SFR (select-for-relay) message
containing the ID list of the selected relay or relays (Fig. 5).
This will conclude the relay selection phase and the
communication between the BS and mobile can start through
the selected relays.
Fig. 4. DTS transmitted by BS and forwarded by non-retreated relays to the
mobile.
Besides, if more than (d – 1) relays remain after the DTS
phase, the mobile can choose the best relays in terms of SNR.
The CRS protocol will be repeated every T seconds depending
on the channel state variations of the relay-mobile links. The
selection procedure can also be triggered if the mobile’s
position changes significantly.
Fig. 5. SFR transmitted by the mobile to notify the selected relay(s).
Other selection protocols presented above (BRS, RSOD
and RSER) did not take into consideration the delay difference
between the signals received from the selected relays, which
may reduce the diversity to one for a relative delay of
one �.
Another important characteristic of the proposed CRS protocol
is that the relays do not need to be synchronized with each
other and with the BS. This will reduce considerably the
signaling overhead between the different nodes in the
cooperative network.
In the following section, we are going to compare the
proposed CRS protocol to the RSER protocol in terms of
needed time to accomplish the selection process and the error
rate performance in the communication phase.
IV. NUMERICAL RESULTS
The signaling complexity during the CRS selection process
is shown in Table I. Let us differentiate the transmitted
messages and the received messages . For instance, the
first column reflects the fact that the BS sends the RTS signal
and the N relays with the mobile receive this signal. However,
the CTS, sent by the mobile, will reach only � relays,
and � non-retreated relays receive the DTS from the BS and
forward it to M. At last, relays are selected by the mobile.
It is easily noticed that: � � .
Table I. Signaling messages of CRS protocol
RTS CTS DTS SFR
� � �
� 1 1 � 1
One important characteristic of the CRS protocol is the time
efficiency through the way of exchanging pilot messages and
the definition of a �������between the signals coming from
the base stations and the relays. This will speed up the
selection procedure, since some relay nodes may experience
processing delays while executing multiple tasks. Therefore a
well-defined value of ������� can solve this problem.
Next, we illustrate the approximate time needed by each
protocol to complete its selection process and start a
communication session.
In CRS protocol, we confine ������� to six symbol periods.
It is assumed that the maximum delay for the RSER
protocol may change from 0 to 10 �, since there is no
limitation concerning this issue. We assume a GSM system
with a cell radius of 1000 meters and the symbol period is
equal to � .
The propagation delays are in the order of 5 and a random
processing delay is considered at the relays, BS and mobile.
Fig. 6 shows the average time needed by CRS and RSER
protocols to finish the relays selection process for different
maximum relative delay � . As can be seen from the
simulation results, our proposed CRS protocol chooses the
appropriate relays faster than the RSER does. Moreover, the
needed time for signaling becomes the same for CRS for
delays bigger than 6 � because CRS does not take into
account signals arriving after �������. However, the
signaling time of RSER increase with high relative delays
Hereafter, we compare the error rate performance of the
CRS protocol with the RSER protocol. A diversity order of 2
is required, therefore only the best relay is chosen in addition
to the BS. The Alamouti code [12] is used in the
communication session and 4-QAM symbols are sent between
the BS and the mobile. All the channels are considered to be
Rayleigh fading with unit variance (σ2 = 1).
In Fig. 7, the Symbol Error Rate (SER) and Bit Error Rate
(BER) are plotted for both CRS and RSER protocols.
It can be noticed from the figure that CRS selection protocol
outperforms the RSER protocol. This is due to the fact that
CRS eliminates the signals arriving at a delay difference of
�������� � that has a destructive effect on the diversity,
but the RSER does not eliminate the relays with critical
delays.
Fig. 6. Signaling time comparison between CRS and RSER protocols.
Fig. 7. Error rates of CRS and RSER protocols.
V. CONCLUSION
Relay selection is a critical operation to ensure better
performance in cooperative networks. In this paper, we have
designed an asynchronous relay selection protocol that takes
into consideration both SNR and delay while choosing the
most appropriate relays for a given cellular communication.
The different steps of the centralized relay selection protocol
were shown. The relays with low SNR may retreat in the early
stages of the selection process. Moreover, the receiver will
choose the relays with signals arriving within a certain delay
threshold and a relative delay not equal to some critical values.
The number of needed signaling messages was also computed
for the proposed protocol. We have computed the average time
spent in the selection process and compared it with an existing
protocol. Not only our protocol has lower signaling time, but it
was also shown that it has better error rate performance in the
communication phase where the selected relays are involved.
REFERENCES
[1] R. U. Nabar, H. Bolcskei, and F. W. Kneubuhler, "Fading relay channels:
performance limits and space-time signal design," IEEE Journal on
Selected Areas in Communications, vol. 22, pp. 1099-1109, Aug. 2004.
[2] J. N. Laneman, D. Tse, and G. W. Wornell, "Cooperative diversity in
wireless networks: efficient protocols and outage behaviour," IEEE
Transactions on Information Theory, vol. 50, no. 12, Dec. 2004.
[3] Y. Liu, R. Hoshyar, X. Yang, and R. Tafazolli, "Integrated radio resource
allocation for multihop cellular networks with fixed relay stations," IEEE
Journal on Selected Areas in Communications, vol. 24, pp. 2137-2146,
Nov. 2006.
[4] A. Adinoyi and H. Yanikomeroglu, "Cooperative relaying in multi-
antenna fixed relay networks," IEEE Transactions on Wireless
Communications, vol. 6, no. 2, pp. 533-544, Feb. 2007.
[5] M. Nahas, A. Saadani, and G. Rekaya, "Bounded Delay-Tolerant Space
Time Block Codes for Asynchronous Cooperative Networks," IEEE
Transactions on Wireless Communications, vol. 10, no. 10, pp. 3288-
3297, Oct. 2011.
[6] A. Bletsas, A. Khisti, D.P. Reed, and A. Lippman, "A simple cooperative
diversity method based on network path selection," IEEE Journal on
Selected Areas in Communications, vol. 24, no. 3, pp. 659-672, Mar.
2006.
[7] A. Papadogiannis, A. Saadani, and E. Hardouin, "Exploiting dynamic
relays with limited overhead in cellular systems," in Proceedings of
IEEE GLOBECOM Workshops, Honolulu, United States, Dec. 2009.
[8] Helmut Adam, Christian Bettstetter, and Mohammad Senouci Sidi,
"Adaptive relay selection in cooperative wireless networks," in
Proceedings of IEEE International Symposium on Personal Indoor and
Mobile Radio Communications (PIMRC), Cannes, France, Sep. 2008.
[9] Xijun Wang, Wei Chen, and Zhigang Cao, "A simple probabilistic relay
selection protocol for asynchronous multi-relay networks employing
rateless code," in Proceedings of IEEE International Conference on
Communications (ICC), Kyoto, Japan, Jun. 2011.
[10] Wai P. Tam, Tat M. Lok, and Tan F. Wong, "Flow-optimized
asynchronous relay selection protocol for parallel relay networks," in
Proceedings of IEEE International Conference on Communications
(ICC), Beijing, China, May 2008.
[11] M. Nahas, A. Haj-Ali, T. Chakerian, and M. Rihani, "Asynchronous
relay selection protocol for distributed cooperative networks," in
Proceedings of International Symposium on Wireless Communication
Systems (ISWCS'12), Paris, France, Aug. 2012.
[12] S. Alamouti, "A simple transmit diversity technique for wireless
communications," IEEE Journal on Select. Areas In Communications,
vol. 16, no. 8, pp. 1451–1458, Oct. 1998.
[13] F. Oggier, G. Rekaya, J-C. Belfiore, and E. Viterbo, "Perfect space time
codes," IEEE Transactions on Information Theory, vol. 52, no. 9, pp.
3885–3902, Sep. 2006.
0 1 2 3 4 5 6 7 8 9 100
5
10
15
20
25
∆ / Ts
Sig
na
lin
g tim
e (
ms)
Time spent to select the relays
CRS protocol
RSER protocol
0 5 10 15 20 2510
-5
10-4
10-3
10-2
10-1
100
SNR (dB)
Err
or
rate
s
Error rates comparison
SER for CRS protocol
SER for RSER protocol
BER for CRS protocol
BER for RSER protocol