Design of optimal relay location in two-hop cellular systems
Jane-Hwa Huang•Li-Chun Wang•
Chung-Ju Chang•Wen-Shan Su
Published online: 20 May 2010
Springer Science+Business Media, LLC 2010
Abstract Relay stations are usually used to enhance the signal strength for the users near cell boundary, thereby extending the cell coverage. However, transmission through a relay station needs two transmission phases. The first phase is from base station to relay station, and the second one is from relay station to mobile station. Thus, using relay station may decrease system capacity due to two-phase transmission time. As a result, whether or not data are transmitted by one-hop or two-hop phases should be determined according to both signal strength and throughput. In this paper, we investigate the optimal relay location aiming to maximize system capacity. We consider two relay selection rules for determining whether two-hop transmission will be used: signal strength-oriented and throughput-oriented selection rules. We find that the signal strength-oriented two-hop transmission may yield even lower system capacity than the one-hop transmission. In the throughput-oriented scheme, the two-hop transmission can achieve higher system capacity than the one-hop transmission. By simulations, we determine the optimal
relay location and show the coverage enhancement by the relaying network. Extensive simulations are performed to investigate the impacts of relay transmission power and the number of relay stations on system capacity and optimal relay location. The simulation results reveal important insights into designing a relaying network with high system capacity.
Keywords Multi-hop cellular systems Relaying networks Relay location design
1 Introduction
Recently, the relay transmission technique is widely used in the next-generation wireless systems [1–7], because relay stations (RSs) can help the base station (BS) forward data and improve signal quality for mobile stations (MSs) near cell boundary to extend coverage. Compared to BS, deploying RS can reduce infrastructure cost. In addition, since RSs do not need wireline connections to the backhaul network, the relaying networks can be quickly deployed on a large scale.
However, transmission through a relay station needs two transmission phases, which may degrade system capacity. Therefore, one interesting issue in the relaying networks is to determine whether a two-hop transmission is necessary. Furthermore, it is an important task to investigate the impact of RS location on link reliability and system capacity. Specifically, if the relay stations are deployed far away from BS, the user at cell boundary can receive stronger signal from RS. However, the longer hop distance between BS and RS will decrease the relay link capacity. Therefore, determining appropriate relay location to achieve the tradeoff between communication reliability and
J.-H. Huang (&)
Graduate Institute of Communication Engineering, National Chi Nan University, Nantou County 545, Taiwan, ROC
e-mail: jhhuang@ncnu.edu.tw L.-C. Wang C.-J. Chang
Department of Communication Engineering, National Chiao Tung University, Hsinchu 300, Taiwan, ROC
e-mail: lichun@cc.nctu.edu.tw C.-J. Chang
e-mail: cjchang@cc.nctu.edu.tw W.-S. Su
Accton Technology Co., Ltd., Hsinchu 300, Taiwan, ROC e-mail: vincent_su@accton.com.tw
system capacity is an essential issue in the relaying networks.
This paper aims to determine the optimal relay location in a two-hop network to maximize system capacity subject to the outage probability requirement. We consider two relay selection principles for deciding whether a two-hop relaying transmission will be used. The first one is the signal strength-oriented selection rule, which is to compare the received signal strength directly from BS and that from RS. The link with stronger signal strength will be adopted for data transmission to improve link reliability. The sec-ond scheme is the throughput-oriented selection rule. According to this selection scheme, each user will estimate the effective transmission rate for the link directly from BS and that via the two-hop communication. Then, the user selects the link with higher effective transmission rate. We also consider two slot allocation schemes: equal time-duration and equal user-throughput schemes. The former scheme allocates each user with the same fraction of time for data transmission. The later one allocates the radio resource (e.g. time-slot) to make each user with the same throughput. We investigate the impact of time-slot alloca-tion scheme on the optimal relay locaalloca-tion. Moreover, we show the coverage enhancement while the RSs are deployed at the optimal locations.
Traditionally, the relaying communications are mostly used in the ad hoc networks [5, 6, 8, 9]. Recently, as deploying ubiquitous broadband wireless networks has become a critical topic, the relaying transmission is also widely exploited by the infrastructure-based wireless net-works [3,7,10]. In the literature, the performance studies for relaying networks mainly focus on capacity enhance-ment [10–12] and coverage extension [13,14]. In [15], the authors discussed three typical resource allocation schemes for relaying networks, including relaying in time-domain, relaying in frequency-domain, and hybrid time/frequency-domain relaying schemes. The work in [16] studied the impacts of the number of RSs and relay transmission power on throughput, given the locations of RSs. In [14], it was shown that the relaying node selection method can signif-icantly affect the system coverage.
Different from the previous works, the main contribu-tion of this paper is to investigate the impact of relay location on system capacity with considering the two-phase transmission overhead. In addition, by simulations we compare the capacity performance for two relay selection schemes, and we determine the optimal relay location. We also investigate the impacts of radio channel, relay trans-mission power, and the number of deployed RSs in a cell on the optimal relay location and the system capacity. The results obtained from extensive simulations can provide useful guidelines for network design to enhance system capacity in the relaying networks.
The rest of this paper is organized as follows. Section2 describes the relaying network architecture and the radio channel effects. In Sect.3, we clarify the tradeoff between link reliability and system capacity, and perform the opti-mization design for relay location. In Sect. 4, we discuss the relay selection rules and time-slot allocation schemes. Performance evaluations are shown in Sect.5. Conclusions are given in Sect.6.
2 Relaying network architecture and assumptions
2.1 Network architecture and multihop relay operation
We consider a two-hop relaying network as shown in Fig.1. Each cell consists of one BS and K RSs, each with an omni-directional antenna. The coverage radii of BS and RSs are rbs and rrs, respectively. The RSs are regularly
deployed around the BS with the separation distance dbr
between BS and RS. There are N MSs uniformly distrib-uted in the cell. If one-hop transmission is used, the transmission rate between BS and MS is Rbm. If two-hop
transmission is selected, the transmission rate between BS and RS is Rbr, and that between RS and MS is Rrm. Clearly,
the relaying communication can improve link reliability due to shorter hop distance, but may lower system capacity due to two-phase transmission time. Therefore, one important task in relaying networks is how to decide whether a user should use the two-hop communication. Besides, since the hop distance affects the transmission rates Rbrand Rrm, another challenge is designing the relay
location to achieve the tradeoff between link reliability and system capacity.
In general, there are two typical relaying operation modes in the multihop cellular networks, namely, trans-parent and non-transtrans-parent relay modes [1, 2]. The transparent mode aims at enhancing throughput, while the non-transparent mode is mainly used to extend coverage. In the transparent relay mode, an MS can directly receive the
(a) (b)
Fig. 1 The architecture of a two-hop relaying network with one BS and K RSs. a A two-hop relay netweork. b Transmission rate between transmitter and receiver
control information from BS, which identifies the duration and the subchannel for each RS to forward users’ data. By this resource allocation information, an MS can estimate the channel quality from each RS according to the asso-ciated subchannel pilot symbols. In the non-transparent relay mode, each RS needs to transmit the control infor-mation including its relay preamble. Thus, an MS can estimate the channel quality of each RS directly by the relay preamble, and then decide which RS will be used.
2.2 Radio channel effects
This paper considers the radio channel effects of path loss and shadowing [17]. Subject to path loss, the received power strength decays with the propagation distance r between the transmitter and the receiver. Shadowing is caused by obstacles between the transmitter and the receiver. Generally, shadowing can be modeled by a log-normal random variable 10n/10. Therefore, supposing that the transmission power is Pt, the received power can be
written as
Pr¼ Pt ðLrÞ1 10n=10: ð1Þ
where Lr represents the path loss. n is a Gaussian
distributed random variable with zero mean and standard deviation (r); and the probability of density function (pdf) is defined as fnðnÞ ¼ 1 ffiffiffiffiffiffi 2p p rexpð n2 2r2Þ: ð2Þ
According to the system model for IEEE 802.16j with relay stations [18], the path loss can be modeled as
LrðdBÞ ¼ 20 log10ð4pr k Þ for r r 0 0 Aþ 10c log10ðrr0Þ þ MLfcþ MLht for r [ r 0 0 ð3Þ where k is the wavelength in meter; r0= 100 (m) is a
reference distance; c, r00, and A are the parameters
depending on the propagation environment. Assume that the antenna height of the transmitter (BS or RS) is hb. For
flat terrain, c is defined as c¼ 3:6 0:005hbþ
20 hb
: ð4Þ
Let fcbe the carrier frequency, and htbe the antenna height
of the receiver (RS or MS). According to [18], r0 0 and A are given as r00 ¼ r010ð MLfc þMLht 10c Þ ð5Þ A¼ 20 log10ð 4pr00 k Þ ð6Þ where MLf c ¼ 6 log10ð fcðMHzÞ 2; 000 Þ ð7Þ MLh t ¼ 10 log10ð ht 3Þ; for ht 3 m 20 log10ð ht 3Þ; for ht[ 3 m: ð8Þ 2.3 Link reliability
The transmission is outaged if the signal-to-noise ratio (SNR) is lower than a specified threshold zth. Let N0be the
noise power. Subject to the radio channel effects, the out-age probability for the transmission can be calculated as
Poutage¼ Pr½SNR\zth ¼ Pr Pt ðLrÞ 1 10n=10 N0 \zth " # : ð9Þ
In this paper, we consider the area outage probability [19]. Suppose that the receivers are uniformly distributed in a circular area centered at the transmitter with radius rc.
The area outage probability is defined as
Po¼ 1 pr2 c Zrc 0 Pr n\10 log10zthLr N0 Pt 2prdr ¼ 2 r2 c Zrc 0 Z 10 log10zthLrN0Pt 1 fnðnÞdn " # rdr ¼ 2 r2 c Zrc 0 ð1 Qð10 log10zthLr N0 Pt r ÞÞrdr ¼ fzthðrcÞ: ð10Þ where QðzÞ ¼ Z1 z 1 ffiffiffiffiffiffi 2p p ey2=2dy: ð11Þ
Suppose that the outage probability requirement is Po,req.
The corresponding maximum reception range is rMAX¼ fz1
th ðPo;reqÞ for a modulation and coding scheme (MCS)
with the SNR threshold zth. That is, if the separation
dis-tance between receiver and transmitter is less than rMAX,
the average outage probability can be less than Po,req for
the given MCS. In (10), the outage probability Po is a
nonlinear function of coverage radius. By the numerical method, we can obtain the maximum reception range
rMAX¼ fz1th ðPo;reqÞ.
This paper considers the effects of path loss and shad-owing on the system-level performance (such as the cell capacity and coverage) for simplicity, as the system-level analysis in [20,21]. If the multipath fading is considered, a proper fading margin is required. Therefore, the cell cov-erage and per-user throughput will decrease. However, the
proposed optimal relay-location design approach in Sect.3 still can be straightforwardly applied to the case including the effect of small-scale fading. We also note that the link-level simulations to find the SNR threshold zth actually
have incorporated the impact of multipath fading [22].
3 Optimal relay location design
Link reliability and throughput are both essential factors in designing a relaying network. From a link reliability per-spective, deploying RSs far away from BS can improve the signal strength for the users near cell boundary. However, from the link capacity standpoint, it is better to deploy the RSs near the BS since a higher relay link capacity between BS and RS can forward more traffic for the users.
To achieve the tradeoff between link reliability and throughput, we formulate an optimization problem to find out the optimal relay location, aiming to maximize system capacity subject to the outage probability requirement. The system capacity C is defined as the aggregated throughput of BS. In a relaying network, the system capacity indeed depends on the relay location, the relay link selection rule, and the resource allocation scheme as detailed in Sect.4. The decision variables in the optimization problem include the relay location dbr, the direct link set Sbm, and the relay link set Srm. According to the relay link selection rule, the users in the set Sbmwill receive their data via the direct link from BS, and those in Sbrm will use the two-hop trans-mission. Based on these considerations, the capacity maximization issue can be formulated as a nonlinear pro-gramming problem as expressed in the following:
MAX
dbr;Sbm;Sbrm
C(Overall throughput of a cell) ð12Þ
subject to Po;bm Po;req; if using direct link,
Po;br; Po;rm Po;req; otherwise.
ð13Þ In (13), Po,br, Po,rm, and Po,bmrepresent the outage
proba-bilities for the link between BS and RS, that between RS and MS, and that between BS and MS, respectively. The constraint states the link reliability requirement that the outage probability in each link should be less than an outage probability threshold Po,req.
4 Relay selection rules and time-slot allocation schemes
In a two-hop relaying network as shown in Fig.1, the data packets can be delivered directly from the BS or relayed by the RS. The system capacity and per-user throughput are dependent on the rate adaptation, relay link selection rule,
and time-slot allocation scheme, as the procedures shown in Fig. 2. We explain these three steps in the following.
4.1 Distance-based rate adaptation
In a wireless multihop network, the separation distance between the transmitter (BS/RS) and the receiver (RS/MS) will affect the throughput and the transmission PHY mode (i.e., the modulation and coding scheme). In this paper, we consider seven transmission PHY modes in the IEEE 802.16 System as listed in Table1. In the table, the SNR thresholds and the net data rates for the transmission PHY modes are shown [22]. According to the transmission power (Pt), the SNR threshold (zth), and the outage
prob-ability requirement (Po,req), we can determine the
maxi-mum reception range rMAX,jfor each transmission mode as
discussed in Sect. 2.3, where rMAX,1[ rMAX,2[ _ [
rMAX,7. In principle, the communication pairs with a
shorter separation distance can adopt a transmission mode with higher net data rate. Suppose that the separation dis-tance between transmitter and receiver is dTX,RX. By the
distance-based rate adaptation, the transmission PHY mode is determined as
Transmission PHY mode¼ j; if rMAX;jþ1\dTX;RX rMAX;j: ð14Þ ∑
≤
Fig. 2 Procedures of the resource allocation for a user, including three steps: (1) rate adaptation, (2) relay link selection, and (3) time-slot allocation
By this distance-based rate adaptation, the outage proba-bility requirement for each link can be ensured. Further-more, once the transmission PHY mode is given, from Table1we can obtain the net data rate for each link, i.e., the net data rate for the direct link between BS and MS (Rbm), that between BS and RS (Rbr), and that between RS
and MS (Rrm).
4.2 Relay link selection rules
• Throughput-Oriented (TO) Selection Rule:
In this scheme, data will be delivered by two-hop transmission if forwarding the packet by relay station can achieve higher effective transmission rate. Clearly, the objective of this scheme is to increase per-user throughput. Let P be the packet size. With two trans-mission phases, the effective transtrans-mission rate Rbrmfor
the two-hop communication can be given as Rbrm¼ P tBSRSMS ð15Þ ¼ P P Rbrþ P Rrm ¼ 1 Rbr þ 1 Rrm 1 ð16Þ
where tBSRSMS¼RPbrþRPrm is the total two-phase transmission time. In this scheme, data will be delivered by two-hop communication, as long as
Rbrm [ Rbm: ð17Þ
Therefore, by the throughput-oriented selection rule, the transmission rate Re(n) for user n is equal to
ReðnÞ ¼ maxðRbm; RbrmÞ: ð18Þ
• Signal Strength-Oriented (SSO) Selection Rule: In this scheme, the user will adopt the two-hop transmission to deliver data if the received signal strength from RS (SignalRS) is stronger than that from
BS (SignalBS). Clearly, this scheme aims at improving
the link reliability. In addition, the effective transmission rate Re(n) is equal to
ReðnÞ ¼
Rbrm; if SignalRS[ SignalBS Rbm; Otherwise:
ð19Þ
4.3 Time-slot allocation schemes
In this paper, two time-slot allocation schemes are con-sidered: equal time-duration and equal user-throughput allocation schemes. The radio resource is allocated in a time division multiplexing (TDM) fashion. At each time slot, only one transmitter will send data. Therefore, the intra-cell interference is not considered in this paper. In other multihop cellular systems, BS and RSs may transmit data in parallel to improve spectrum efficiency and system capacity. However, such a concurrent transmission system should face a difficult challenge in managing the compli-cated interference.
• Equal Time-Duration (ETD) Allocation:
As shown in Fig.3(a), in this scheme each user is allocated with the same fraction of time for data transmission, no matter whether the data are transmit-ted directly from BS or by two-hop communication. Consider a cell with N users, each with the effective transmission rate Re(n). Since all the users evenly share
the radio resource, the system capacity is given as
C¼ PN
n¼1ReðnÞ
N : ð20Þ
• Equal User-Throughput (EUT) Allocation:
The main concept of this scheme is to allocate the time-slot so that all the users have the same throughput. This
Table 1 The SNR threshold and the net data rate for seven modulation and coding schemes in the IEEE 802.16 System
MCS Modulation Code rate SNR (dB) Net date rate
(Mbit/s) 1 BPSK 1/2 0.0 1.29 2 QPSK 1/2 2.5 2.59 3 QPSK 3/4 6.0 3.88 4 16-QAM 1/2 9.0 5.18 5 16-QAM 3/4 12.0 7.77 6 64-QAM 2/3 16.0 10.37 7 64-QAM 3/4 21.0 11.66 (a) (b)
Fig. 3 Frame structure for the equal time-duration and equal user-throughput allocation schemes with N users. (a) Equal time-duration allocation scheme: each user can send data in one slot of each frame. (b) Equal user-throughput allocation scheme: each user can send a packet with size P in each frame
scheme can achieve the fairness for each user in terms of throughput. Suppose that during a time interval, each user can send a packet with the same size P as shown in Fig.3(b). The total transmission time for N packets is equal toPNn¼1RP
eðnÞ. Therefore, the system capacity can
be expressed as
C¼Total transmitted data bits Total transmission time
¼PNN P n¼1RePðnÞ ¼ N X N n¼1 1 ReðnÞ1 !1 : ð21Þ 5 Simulation results
In this section, we compare the system capacity according to the throughput-oriented and signal strength-oriented relay selection rules, with considering the effect of relay location. Besides, we investigate the impacts of radio channel, the number of RSs, and the transmission power of RS on the system capacity and the optimal relay location. The system parameters are listed in Table2. We consider seven transmission PHY modes as listed in Table1. The SNR threshold for outage events is zth= 0 dB. The outage
probability requirement is Po,req= 0.1. With this
require-ment, the coverage radius of BS is rbs= 1750 (m) when
the transmission power of BS is 40 dBm. Unless otherwise specified, there are 8 RSs in a cell, each with the trans-mission power of 37 dBm. The RSs are deployed around the BS shown in Fig.1, with the separation distance dbr
between BS and RS. In this paper, we consider N = 20 active MSs uniformly distributed in the cell as an example. The MSs follow the resource allocation procedures in Fig.2to decide whether the relaying transmission is used. We obtain the system performance by averaging the results from the 10,000-round simulation for a given relay
location. In addition, by the exhaustive search method, we can find the optimal relay location.
5.1 Impact of relay link selection rule on system capacity and optimal relay location
Figure4 shows the achieved system capacity versus the separation distance between BS and RS, where the equal time-duration allocation scheme is used. Clearly, the throughput-oriented two-hop transmission can improve system capacity. Furthermore, there exists an optimal relay location to maximize system capacity. In this example, the optimal relay location is 1,273 m. Besides, the throughput-oriented selection scheme can achieve higher optimal system capacity than the signal strength-oriented scheme by 5.5%, and than the case without relay by 25.4%. This figure also shows that the signal strength-oriented two-hop transmission may yield lower system capacity than the one-hop transmission. For example, if deploying RSs at the distance of 450 m, the system capacity is lower than that without RS by about 30%. These phenomenons are due to the fact that by the signal strength-oriented scheme, an MS will choose the relaying transmission as long as the MS is closer to the RS rather than the BS. Thus, when the MS is between BS and RS, the MS may still choose the two-hop communication although using the direct communication can achieve higher throughput. In result, the signal strength-oriented scheme achieves lower overall system capacity while it can improve signal quality. On the other hand, the throughput-oriented scheme will choose the relaying transmission only if the RS can help increase per-user throughput. Therefore, the throughput-oriented scheme can always achieve higher system capacity than the signal strength-oriented scheme.
Table 2 System parameters
Item Nominal value
Carrier frequency (fc) 3.5 GHz System bandwidth (BW) 3.5 MHz BS transmission power 40 dBm RS transmission power 37 dBm Coverage radius of BS (rbs) 1750 m BS antenna height 30 m RS antenna height 15 m MS antenna height 2 m
Outage probability requirement (Po,req) 0.1
Standard deviation for shadowing (r) 8 dB
Noise power (N0) -102 dBm 0 400 800 1273 1750 3.5 4 4.5 5 5.5 6 6.5
Distance between BS and RS (m)
System Capacity (Mb/s)
Throughput−oriented Scheme Signal Strength−oriented Scheme Without RS
Fig. 4 Achieved system capacity for different relay selection rules, where the equal time-duration allocation scheme is used
Figure5 illustrates the system capacity for the equal user-throughput allocation scheme. In the figure, the opti-mal relay location is 1,305 (m) for the throughput-oriented selection rule. It is shown in Figs.4 and 5 that the equal user-throughput allocation scheme achieves lower system capacity than the equal time-duration scheme. This is because the equal user-throughput scheme allocates more radio resource for the users with lower transmission rates to achieve the same throughput for each user, thereby low-ering the overall system capacity.
5.2 Impact of relay selection rule on effective data rate and SNR
Figure6 shows the probability mass function (PMF) of effective data rate Re(n) for various relay selection rules
using the equal time-duration scheme, where the RSs are deployed at the optimal location. In the multihop cellular system, the users near BS can be served directly by BS with higher data rates. Hence, in the figure the probability of higher data rate (e.g., 9*12 Mb/s) for all the cases are almost the same. For the case without relay, the users near cell boundary are served by BS with very low data rates. Accordingly, many users have lower effective data rates. In this example, about 55% of the users experience the effective data rate lower than 4 Mb/s. Relay station can improve the effective data rate especially for the users near cell boundary. This figure shows that by using RSs, most of users can have effective data rate of 4*6 Mb/s. Note-worthily, according to the throughput-oriented selection rule, 13% of the users can have effective data rate of 6*8 Mb/s. However, by the signal strength-oriented scheme, only 1% of the users can achieve 6*8 Mb/s effective data rate. In the signal strength-oriented rule, the
users near the RS will use the two-hop transmission with stronger signal strength. In result, the effective data rate decreases due to higher overall transmission time.
Figure7 shows the complementary cumulative distri-bution function of the received SNR for the users. One can see that relay station can improve the received SNR. In the throughput-oriented scheme, the users near the relay may still use direct transmission from BS to improve per-user throughput. Because of longer hop distance, the received SNR in the throughput-oriented scheme is slightly lower than that in the signal strength-oriented scheme. However, the throughput-oriented scheme can increase system capacity as shown in Figs.4and5.
0 400 800 1305 1750 2.5 3 3.5 4 4.5 5 5.5
Distance between BS and RS (m)
System Capacity (Mb/s)
Throughput−oriented Scheme Signal Strength−oriented Scheme Without RS
Fig. 5 Achieved system capacity for different relay selection rules, where the equal user-throughput allocation scheme is used
0 1 2 3 4 5 6 7 8 9 10 11 12 0 5 10 15 20 25 30 35 40 45 50
Effective Data Rate, Re(n) (Mb/s)
Probability (%)
Signal Strength−oriented Sheme Throughput−oriented Scheme Without RS
Fig. 6 The PMF of effective data rate Re(n) according to different
relay selection rules, where the equal time-duration allocation scheme is used and the RSs are deployed at the optimal locations
−15 −10 −5 0 5 10 15 20 25 30 0.5 0.6 0.7 0.8 0.9 1 SNR (dB)
Signal Strength−oriented Sheme Throughput−oriented Scheme Without RS
Prob(SNR
≥
Abscissa)
Fig. 7 The complementary CDF of users’ SNR according to various relay selection rules, where the equal time-duration allocation scheme is used and the RSs are deployed at the optimal locations
5.3 Impact of relay transmission power and the number of RSs on system capacity
Figure8 shows the system capacity against the relay location for various RS transmission power, where there are 8 RSs in a cell. Obviously, there exists optimal relay location to maximize the system capacity. In addition, the optimal system capacity can be improved as the RS transmission power increases. Besides, to improve the throughput and the signal strength of outer users, the RS should be deployed closer to cell boundary as the trans-mission power decreases. This figure also shows that increasing the transmission power (e.g., more than 3 W) may not significantly increase system capacity. This phe-nomenon is due to the fact that in a relaying network, only the users near cell boundary benefit from the RS. Since the link capacity between BS and RS is limited, the improve-ment of effective data rate for outer users gradually diminishes for a higher RS transmission power.
Figure9illustrates the system capacity against the relay location, where the number of RSs varies from 4 to 12. The total transmission power of RSs is fixed at 20 W. In the figure, since the hop distance between RS and the outer user can be reduced as the number of RSs increases, the system capacity increases. In addition, as more RSs are deployed each with a lower transmission power, the opti-mal distance between BS and RS should be increased to improve link quality for the outer users. Noteworthily, deploying more RSs (e.g., greater than 8 RSs) will not remarkably improve system capacity. This is because limited by the relay link capacity between BS and RS, per-user throughput and system capacity will not significantly increase as more RSs are deployed.
5.4 Impact of radio channel on system capacity and coverage
In Fig.10, the system capacity against the relay location for various shadowing standard deviation r is shown. The figure shows that the optimal system capacity and the coverage of BS increase for a smaller r. In addition, the optimal relay location is increasing as the standard devia-tion r decreases. This is because the case with a smaller r can have better link reliability, thereby enhancing system capacity and coverage of BS.
Figure11shows the coverage radius of a two-hop cell for various shadowing standard deviation r, where the relays are deployed at the optimal locations and the outage prob-ability requirement for each link is Po,req= 0.1. Different
0 400 800 1200 1750 4.8 5 5.2 5.4 5.6 5.8 6 6.2
Distance between BS and RS (m)
System Capacity (Mb/s) RS transmission power (5W) RS transmission power (4W) RS transmission power (3W) RS transmission power (2W) RS transmission power (1W)
Optimal Relay Location: 1273
1289
1308
1335
1381
Fig. 8 Achieved system capacity and optimal relay location for various relay transmission power, where there are 8 RSs in a cell
0 400 800 1200 1750 5 5.2 5.4 5.6 5.8 6 6.2
Distance between BS and RS (m)
System Capacity (Mb/s) 12RSs (5/3W) 10RSs (2W) 8 RSs (2.5W) 6 RSs (10/3W) 4 RSs (5W)
Fig. 9 Achieved system capacity and optimal relay location for different number of deployed RSs, where the total transmission power of RSs is fixed at 20 W 0 400 800 1200 1750 2070 2365 5 5.2 5.4 5.6 5.8 6 6.2 6.4 6.6
Distance between BS and RS (m)
System Capacity (Mb/s)
Shadowing Standard Deviation (2dB) Shadowing Standard Deviation (5dB) Shadowing Standard Deviation (8dB)
Fig. 10 System capacity versus relay location for various shadowing standard deviation r, using the throughput-oriented selection scheme and the equal time-duration allocation scheme
from Figs.4,5,6,7,8,9and10which focus on the capacity enhancement by deploying RSs in a fore-deployed cellular system, here we aim at understanding the coverage exten-sion by RSs. This is a useful information for network plan-ning to reduce the infrastructure cost since we can deploy fewer BSs in an area. This figure shows that by deploying the RSs at the optimal location as shown in Fig.10, the coverage radius of two-hop cell can be significantly increased. In this example, compared to the case without RS, RS can extend the coverage by about 36% for r = 8 dB.
6 Conclusions
In this paper, we investigate the impact of relay location on system capacity and discuss how to decide whether the two-hop relay communication should be used to deliver data. The contributions of this work are described in the following. First, we compare the signal strength-oriented and throughput-oriented link selection rules. Simulation results show that the signal strength-oriented scheme can improve SNR, whereas may achieve lower system capacity than the one-hop transmission. By contrast, the throughput-oriented scheme can yield higher system capacity, although the received SNR is slightly lower than that in the signal strength-orient scheme. Second, we perform optimization design for relay location to maximize system capacity. It is shown that properly designing relay location can signifi-cantly improve system capacity; however, deploying the RS at random location may even degrade system capacity. Third, the impacts of RS transmission power, the number of RSs in a cell, and the radio channel on system capacity and coverage are investigated. Some important observa-tions are obtained as follows:
• Due to capacity limitation in the link between BS and RS, continually increasing RS transmission power and the number of RSs may not significantly improve system capacity. In the simulation example, eight RSs in a cell can be a good choice.
• For different number of RSs, various propagation environments and transmission power, the relay loca-tion should be appropriately designed to maximize system capacity.
• With a proper location design, RS can not only improve capacity but extend coverage. In the example, RS can extend system coverage by about 33% for various shadowing standard deviations r.
Acknowledgements This work was supported in part by the MoE ATU Plan, the Program for Promoting Academic Excellence of Universities (Phase II), and the National Science Council under Grant 98W803C, Grant NSC 97-2752-E-009-003-PAE, Grant 97-2221-E-009-097-MY3, Grant 97-2221-E-009-099-MY3, and Grant 96-2628-E-009-004-MY3.
References
1. Peters, S. W., Heath, J., & Robert W. (2009). The future of WiMAX: Multihop relaying with IEEE 802.16j. IEEE Commu-nications Magazine, 47(1), 104–111.
2. Genc, V., Murphy, S., Yu, Y., & Murphy, J. (2008). IEEE 802.16j relay-based wireless access networks: an overview. IEEE Wire-less Communications, 15(5), 56–63.
3. Pabst, R., Walke, B., Schultz, D. C., et al. (2004). Relay-based deployment concepts for wireless and mobile broadband radio. IEEE Communications Magazine, 42(9), 80–89.
4. Yanikomeroglu, H. (2002). Fixed and mobile relaying technolo-gies for cellular networks. In the 2nd Workshop on applicaiions and services, in wireless networks (ASWN) (pp. 75–81). Paris, France, July 3–5.
5. Huang, J.-H., Wang, L.-C., & Chang, C.-J. (2006). Capacity and QoS for a scalable ring-based wireless mesh network. IEEE Journal on Selected Areas in Communications, 24(11), 2070– 2080.
6. Huang, J.-H., Wang, L.-C., & Chang, C.-J. (2008). Throughput-coverage tradeoff in a scalable wireless mesh network. Journal of Parallel and Distributed Computing, 68(3), 278–290.
7. Huang, J.-H., Wang, L.-C., & Chang, C.-J. (2008). QoS provi-sioning in a scalable wireless mesh network for intelligent transportation systems. IEEE Transactions on Vehicular Tech-nology, 57(5), 3121–3135.
8. Grossglauser, M., & Tse, D. N. C. (2002). Mobility increases the capacity of ad hoc wireless networks. IEEE/ACM Transactions on Networking, 10(4), 477–486.
9. Li, H., & Yu, D. (2002). Performance comparison of ad-hoc and cellular based routing algorithms in multi-hop cellular networks. In International symposium wireless personal multimedia com-munications (WPMC’02). Hawaii.
10. Irnich, T., Schultz, D., Pabst, R., & Wienert, P. (2003). Capacity of a relaying infrastructure for broadband radio coverage of urban areas. In Proceedings of 10th WWRF meeting. New York. 11. Gastpar, M., & Vetterli, M. (2002). On the capacity of wireless
networks: The relay case. In Proceedings of IEEE INFOCOM 2002, vol . 3 (pp. 1577–1586). New York.
1 2 3 4 5 6 7 8 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400
Shadowing Standard Deviation (σ)
Coverage of a Two−hop Cell (m)
Throughput−oriented Selection with Equal Time−duration Allocation Throughput−oriented Selection with Equal User−throughput Allocation Signal Strength−oriented Selection with Equal Time−duration Allocation Signal Strength−oriented Selection with Equal User−throughput Allocation Without RS
Fig. 11 Coverage of a two-hop cell for various shadowing standard deviation r, where 8 RSs are deployed at the optimal locations and the outage probability requirement is Po,req= 0.1
12. Wittneben, A., & Rankov, B. (2003). Impact of cooperative relays on the capacity of rank deficient mimo channels. In IST mobile and wireless communications summit (pp. 421–425). Portugal: Aveiro.
13. Sreng, V., Yanikomeroglu, H., & Falconet, D. (2002). Coverage enhancement through two-hop relaying in cellular radio systems. In IEEE wireless communication and networking conference (WCNC’02) (pp. 881–885). Orlando, USA.
14. Sreng, V., Yanikomeroglu, H., & Falconet, D. (2003). Relayer selection strategies in cellular networks with peer-to-peer relay-ing. In IEEE VTC fall (pp. 1949–1953), Orlando, FL.
15. Pabst, R., Esseling, N., & Walke, B. H. (2005). Fixed relays for next generation wireless systems-system concept and perfor-mance evaluation. Journal of Communications and Networks, 7(2), 104–114.
16. Viswanathan, H., & Mukherjee, S. (2005). Performance of cel-lular networks with relays and centralized scheduling. IEEE Transactions on Wireless Communications, 4(5), 2318–2328. 17. Wang, L. C., Chen, A., & Huang, S. Y. (2007). A cross-layer
investigation for the throughput performance of CSMA/CA-based wireless local area networks with directional antennas and cap-ture effect. IEEE Transactions on Vehicular Technology, 56(5), 2756–2766.
18. Senarath, G., Tong, W., Zhu, P., et al. (2007). Multi-hop relay system evaluation methodology (channel model and performance metric). IEEE 802.16j-06/013r3.
19. Stu¨ber, G. L. (2001). Principles of mobile communication, 2nd edn. London: Kluwer.
20. Chandrasekhar, V., & Andrews, J. G. (2009). Uplink capacity and interference avoidance for two-tier femtocell networks. IEEE Transactions on Wireless Communications, 8(7), 3489–3509. 21. Viterbi, A. M., & Viterbi, A. J. (1993). Erlang capacity of a
power controlled CDMA system. IEEE Journal on Selected Areas in Communications, 11(6), 892–990.
22. Ball, C. F., Humburg, E., Ivanov, K., & Mu¨llner, R. (2005). Rapid estimation method for data capacity and spectrum effi-ciency in cellular networks. In Proceedings of 14th IST summit. Dresden, Germany.
Author Biographies
Jane-Hwa Huangreceived the B.S., M.S., and Ph.D degrees in electrical engineering from the National Cheng-Kung Univer-sity, Tainan, Taiwan, R.O.C., in 1994, 1996, and 2003, respec-tively. He joined the Department of Communication Engineering, National Chiao-Tung Univer-sity, Taiwan, as a Postdoctoral Researcher from 2004 to January 2006, and a Research Assistant Professor since January 2006. From August 2009, he has been with the Graduate Institute of Communication Engineering, National Chi Nan University, Taiwan, as an Assistant Professor. His current research interests are in the areas of wireless networks, wireless multi-hop communications, and radio resource management.
Li-Chun Wang received the B.S. degree from the National Chiao-Tung University, Hsin-chu, Taiwan, R.O.C., the M.S. degree from the National Taiwan University, Taipei, Taiwan, and the M.Sci. and Ph.D. degrees from the Georgia Institute of Technology, Atlanta, in 1986, 1988, 1995, and 1996, respec-tively, all in electrical engineer-ing. From 1990 to 1992, he was with Chunghwa Telecom. In 1995, he was affiliated with Northern Telecom, Richardson, Texas. From 1996 to 2000, he was with AT&T Laboratories, where he was a senior technical staff member in the Wireless Communications Research Department. Since August 2000, he has been with the Department of Communication Engineering, National Chiao-Tung University, Taiwan, as an associate professor and has been promoted to a full professor since August 2005. He was a corecipient of the Jack Neubauer Best Paper Award from the IEEE Vehicular Technology Society in 1997. His current research interests include cellular archi-tectures, radio network resource management, cross layer optimization for cooperative, and cognitive wireless networks. He is the holder of three US patents with three more pending. He is a senior member of the IEEE.
Chung-Ju Changwas born in Taiwan, ROC, in August 1950. He received the B.E. and M.E. degrees in electronics engineer-ing from National Chiao-Tung University (NCTU), Hsinchu, Taiwan, in 1972 and 1976, respectively, and the Ph.D degree in electrical engineering from National Taiwan Univer-sity (NTU), Taiwan, in 1985. From 1976 to 1988, he was with Telecommunication Laborato-ries, Directorate General of Telecommunications, Ministry of Communications, Taiwan, as a Design Engineer, Supervisor, Project Manager, and then Division Director. There, he was involved in designing digital switching system, RAX trunk tester, ISDN user-network interface, and ISDN service and technology trials in Science-Based Industrial Park. In the meantime, he also acted as a Science and Technical Advisor for the Minister of the Ministry of Communica-tions from 1987 to 1989. In 1988, he joined the Faculty of the Department of Communication Engineering and Center for Tele-communications Research, College of Electrical Engineering and Computer Science, National Chiao-Tung University, as an Associate Professor. He has been a Professor since 1993. He was Director of the Institute of Communication Engineering from August 1993 to July 1995, Chairman of Department of Communication Engineering from August 1999 to July 2001, and the Dean of the Research and Development Office from August 2002 to July 2004. Also, he was an Advisor for the Ministry of Education to promote the education of communication science and technologies for colleges and universities in Taiwan during 1995–1999; he is acting as a Committee Member of
the Telecommunication Deliberate Body, Taiwan. He serves as Editor for IEEE Communications Magazine and Associate Editor for IEEE Trans. Vehicular Technology. His research interests include
performance evaluation, wireless communication networks, and broadband networks. Dr. Chang is a member of the Chinese Institute of Engineers (CIE) and IEEE member.
