Cooperation Strategies for Multi-user Transmission in Manhattan

Environment

∗

Jaewon Chang and Wonjin Sung

Department of Electronic Engineering, Sogang University, 1 Sinsu-dong, Mapo-gu, 121-742 Seoul, Korea

Keywords:

Cooperative Transmission, MU-MIMO, CoMP, Manhattan Environment.

Abstract:

One of the major drawbacks for wirelesscommunication systems in Manhattan environment with tall buildings

lining both sides of the streets is the performance degradation caused by the penetration loss and the effects of

inter-sector interference. To overcome such a degradation, cooperation among the sectors is under an active

investigation as an efﬁcient means to provide an enhanced coverage as well as increased spectral efﬁciency. In

this paper, we describe various type of cooperation sector location for cooperative multi-user transmission, and

determine cooperation strategies for the cooperative operation among sectors by evaluating and comparing the

types of microcell in Manhattan environment. The results shows that the more suitable number of cooperation

sectors is determined by signal-to-interference plus noise ratio (SINR), outage probability, and throughput

comparison. Performance variations based on different numbers of sector density under cooperation are also

presented to suggest an efﬁcient inter-sector cooperative transmission strategy in Manhattan environment.

1 INTRODUCTION

In wireless communication systems, mobile user’s

demands are rapidly increasing for high data rate

and quality network services in urban areas. Ur-

ban micro-cellular system in Manhattan-like environ-

ment have difference features, compared with macro-

cellular systems. The micro-cellular network requires

higher system capacity for the dense trafﬁc channel,

where a signiﬁcant amount of access users to wireless

link occurs. The micro-cellular system is developed

in the metropolitan area, where tall buildings are ag-

gregated and lined both sides of the streets. Thus, the

propagation models of line-of-sight (LOS) and non

line-of-sight (NLOS) wireless channel between base-

stations (BSs) and mobile-stations (MSs) are adopted

for practical performance investigations.

The performance evaluation of micro-cellular net-

works in Manhattan-like environment and related

works have been widely studied. Standardization

groups of IEEE802.16m and 3GPP-LTE advanced

are also developing the next generation cellular sys-

tems considering Manhattan environment. Capacity

and interference statistics of street cross-shaped mi-

cro cells is analyzed in (Ahmed, 2003), and spatial

∗

This work wassupported by the grant from the National

Research Foundation of Korea (NRF) funded by the Korean

government (MEST) (no. 2011-0016146).

multiple-input multiple-output (MIMO) channel ca-

pacity statistics in Manhattan environment are repre-

sented (Chizhik, 2003). The efﬁcient resource con-

trol and channel allocation method are also proposed.

Considering user mobility characteristics, the perfor-

mance focused on average number of handoffs has

been analyzed by (Cho, 2000). The propagation

model and deployment strategies of transmitters was

suggested in (Chiu, 2009), and multi-hop relay net-

works in Manhattan-like environment is investigated

in (Fu, 2007). However, these studies are based on the

non-cooperative operation among BSs without con-

sidering the cooperativetransmission and interference

mitigation.

In this paper, we present the efﬁcient coopera-

tion strategy in Manhattan environment from the per-

formance investigation of the cooperative multi-user

transmission. First, the various cooperation strategies

are presented as increasing the number of coopera-

tion sector antennas, where the uniformly distributed

MSs are located along the cross shape street consider-

ing the urban micro-cellular model of IEEE802.16m

system. For the different BS density in this environ-

ment, cooperation scenarios according to the cooper-

ation sector antenna variation are also described. Sec-

ond, the presented cooperation strategies are applied

to cooperative multi-user transmission using zero-

forcing beamforming and their performance in terms

415

Chang J. and Sung W..

Cooperation Strategies for Multi-user Transmission in Manhattan Environment.

DOI: 10.5220/0004970304150419

In Proceedings of the 16th International Conference on Enterprise Information Systems (ICEIS-2014), pages 415-419

ISBN: 978-989-758-028-4

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

of signal-to-interference plus noise ratio (SINR), out-

age probability, and throughput is compared, with an

aim of determining a feasible cooperation strategy in

Manhattan environment. Third, the performance vari-

ations based on different numbers of BS density and

the cooperative multi-user transmission scheme are

evaluated, to conﬁrm the performance manner in dif-

ferent operational settings.

The organization of the paper is as follows. Sec-

tion II gives the system model for the cooperative

multi-user transmission and the propagation model in

Manhattan environment. In Section III, cooperation

scenarios are explained as the variation based on the

number of cooperation sector antennas and BS den-

sity. The performanceresults are compared and inves-

tigated in Section IV. Concluding remarks are given in

Section V.

2 SYSTEM MODEL

2.1 Channel Propagation Model

Manhattan network consists of 100 buildings and 41

BSs with 4-sector antennas, as shown in Fig. 1. The

building block size is 200m × 200m and the street

width are set to be 30m. BSs are located at the main

street-crossing, which are positioned at every two in-

tersection (BS density η = 1/2). A BS with 4-sector

antennas covers 4 building blocks including a cross-

shaped area and a cell coverage is about 430 × 430

square meters. All cooperative BS sets are assumed

to use the same multi-user transmission scheme. MSs

located uniformly distributed along streets and never

enter the buildings considering outdoor environment

only.

2.2 Signal Model

We consider downlink transmission from sector BSs

with single transmit antennas to target MSs equipped

with single receive antennas in Manhattan environ-

ment. For the cooperative multi-user transmission us-

ing N

sector BSs, the i-th sector antenna of the l-th

cooperation set is denoted by BS

(l)

, and MS

(l)

de-

notes the k-th MS among N

MSs communicating

with a given l-th group of cooperating sector BSs.

In general, the signal model for a cooperative

multi-user transmission with N

single antenna sec-

tors and N

single antenna MSs is given by

r = H

(c)

∑

l6=c

(l)

+ n (1)

−1 −0.5 0 0.5 1

−1

−0.8

−0.6

−0.4

−0.2

0.2

0.4

0.6

0.8

Distance [km]

Figure 1: Manhattan deployment scenario.

where l denotes the index of the cooperation set, and

(l)

, W

(l)

, and s

(l)

are the N

× N

channel matrix,

corresponding N

× N

precoding matrix for the co-

operative multi-user transmission, and N

× 1 trans-

mit signal vector from the cooperative BSs of the l-th

cooperation set, respectively. n denotes an additive

white Gaussian noise (AWGN) vector and c is the in-

dex of the desired cooperation set.

2.3 MIMO Transmission

Cooperative ZF. In this paper, we consider the co-

operative multi-user transmission using zero-forcing

beamforming. By utilizing the pseudo-inverse ma-

trix W = H

(HH

)

−1

for the channel matrix, zero-

forcing beamforming efﬁciently nulls transmission.

Cooperative zero-forcing beamforming is the

practical multi-user MIMO scheme for enhancing the

bandwidth efﬁciency with multiple data transmission

and mitigation of the multi-user interference, simul-

taneously. However, cooperative zero-forcing beam-

forming is not able to use full power transmission at

each sector which is subject to per antenna power con-

straints. It is shown that one of the sector antenna

among cooperativesectors always use full power. The

probability of signal power for the other sector an-

tenna is distributed at the transmit power range below

the normalized maximum power value ‘1’, and the

probability of the low power transmission increases

as the number of cooperation sectors. Thus, the

transmit signal from the other sectors, where a trans-

mit antenna utilizes lower transmit power than the

maximum value ‘1’, impairs the cooperative channel

gain and SINR performance of the cooperative zero-

forcing beamforming.

ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems

416

Cooperative ZF-DPC. Cooperative zero-forcing

dirty paper coding is a non-linear scheme for the

multi-user MIMO transmission, which can utilize the

full power transmission. For cooperative zero-forcing

dirty paper coding from N

sectors to N

multiple

MSs, the LQ decomposition of the channel matrix H

can be utilized to transmit the multiple data for mul-

tiple MSs, and dirty paper coded transmit signal nulls

out the inter-user interference signal. Using LQ de-

composition, can be decomposed as H = LQ, L is

×N

lower triangular matrix and triangular matrix

and Q is N

×N

unitary matrix. For cooperativezero-

forcing dirty paper coding, the diagonal elements of

the lower triangular matrix L are beamforming gain

and the unitary matrix Q can be used for W = Q

3 COOPERATION STRATEGIES

In Manhattan street environment, we consider the BS

collaboration which contains sector antennas to trans-

mit signal for MSs located at the outdoor street. At

the center of the intersection, 4-sector antennas of a

BS transmit signal to target MSs, and a target MS

can receive the desired sinal from a serving sector an-

tenna, when a MS is selected by scheduling among

uniformly distributed MSs in the coverage area of

each sector antenna. Thus, the number of coopera-

tive serving sectors is same as the number of target

MSs (N

= N

), and it is to be determined how many

sector antenna can be utilized for the efﬁcient cooper-

ation strategy.

To determine the efﬁcient cooperation strategy,

the SINR distributions are evaluated and compared

using the 2, 4, 8, and 16-sector cooperation models

using full frequency reuse. The signal is transmit-

ted from each sector with 40 dBm power over the

ﬂat Rayleigh fading channel. The thermal noise of

−104 dBm power is added at the receiver, which

corresponds to 10 MHz transmission bandwidth of

2 GHz center frequency. We assume the MSs of in-

terest is located along the street of the coverage area

for a center cooperation group in Manhattan grid net-

work, and both the transmitter and receiver have the

perfect knowledge of the channel. Employing the co-

operative zero-forcing transmission, each cooperation

sector is subject to per antenna power constraints.

The topology of cooperative sector antennas and

collaboration strategies can be presented by various

scenarios according to the BS density (η) variation.

As the variation of the BS density, these topologies

of sector BSs for the cooperative transmission, which

is adopted for the regular pattern of the cooperation

group in entire Manhattan grid network, are chosen

−0.2 −0.1 0 0.1 0.2

−0.1

0.1

0.2

0.3

0.4

(a)

−0.2 −0.1 0 0.1 0.2

−0.2

−0.1

0.1

0.2

0.3

(b)

−0.2 −0.1 0 0.1 0.2

−0.2

−0.1

0.1

0.2

0.3

(c)

Figure 2: Cooperation scenarios for the BS density η = 1:

(a) 2-sector cooperation, (b) 4-sector cooperation, (c) 8-

sector cooperation.

−10 0 10 20 30 40 50

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

SINR [dB]

Cumulative Distribution Function

NCT

CT: 2−Sector

CT: 4−Sector

CT: 8−Sector

Figure 3: SINR performance for the cooperative zero-

forcing transmission (η = 1).

for the performance comparison.

Case I (η = 1). Considering the maximum BS den-

sity, Each BS is located at every intersection, where

the distance between the nearest two BSs is 230m. In

this case, a sector coverage is about 230× 100 square

meters including 100 × 30 street area. The topology

of 3 cooperation scenarios; (a) 2-sector collaboration

facing each other at edge points of the straight line

street, (b) 4-sector collaboration of 3 BSs located at

a center point and 2 edge points of the straight line

street, and (c) 8-sector collaboration of 5 BSs located

at a center point and 4 edge points of the cross-shaped

street area can be presented.

The cumulative distribution functions (CDFs) of

the SINR for cooperative zero-forcing beamforming

CooperationStrategiesforMulti-userTransmissioninManhattanEnvironment

417

−0.2 −0.1 0 0.1 0.2

0.1

0.2

0.3

0.4

0.5

(a)

−0.2 −0.1 0 0.1 0.2

0.1

0.2

0.3

0.4

0.5

(b)

−0.4 −0.2 0 0.2

−0.2

0.2

0.4

(c)

−0.5 0 0.5

−0.5

0.5

(d)

Figure 4: Cooperation scenarios for the BS density η = 1/2:

(a) 2-sector cooperation, (b) 4-sector cooperation, (c) 8-

sector cooperation, (d) 16-sector cooperation.

−10 0 10 20 30 40 50

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

SINR [dB]

Cumulative Distribution Function

NCT

CT: 2−Sector

CT: 4−Sector

CT: 8−Sector

CT: 16−Sector

Figure 5: SINR performance for the cooperative zero-

forcing transmission (η = 1/2).

using different the number of cooperation sectors are

plotted in Fig. 3. The SINR performance gain of co-

operative transmission using 4-sector antennas over

the non-cooperative transmission amounts to 7 dB

at the median value of 0.5, demonstrating the inter-

and intra-cooperation sector interference signal re-

duction performing cooperative multi-user transmis-

sion. However, it is observed that no noticeable gain

is obtained by performing 2-sector and 8-sector coop-

eration over 4-sector cooperation, since the interfer-

ence impact cannot be signiﬁcantly reduced under the

maximum BS density condition.

Case II (η = 1/2). The Manhattan deployment sce-

nario in Fig. 1 and 4 employ the BSs location at every

two intersection (η = 1/2), where the horizontal or

vertical distance along the street between two BSs is

460m. A sector antenna covers 2 building blocks in-

cluding 200 × 30 street area and a sector coverage is

about 430× 200 square meters. The topology of 4 co-

operation scenarios; (a) 2-sector collaboration facing

each other at edge points of the straight line street,

(b) 4-sector collaboration of 4 BSs located at 4 edge

points of the cross-shaped street area, (c) 8-sector col-

laboration of 6 BSs located at connected cross-shaped

streets of a parallelogram area, and (d) 16-sector col-

laboration of 9 BSs located at a diamond shape area

can be presented.

In Fig. 5, it can be observed that 4-sector coop-

eration outperforms 2, 8, 16-sector cooperation. It

is also shown in the ﬁgure that SINR of cooperative

transmission is substantially higher than that of the

non-cooperativetransmission. At the median value of

0.5, the amount of SINR gain for 4-sector cooperation

exceeds 7.4 dB when compared to non-cooperative

transmission.

4 PERFORMANCE

COMPARISON

In this section, the per-user throughput, total through-

put, and outage probability are evaluated and com-

pared with 4 and 8-sector cooperation models using

full frequency reuse.

Figure 6 shows the CDFs of the per-user through-

put for different BS density of 4-sector and 8-sector

cooperation. The per-user throughput can be ob-

tained by

∑

k=1

log

(1+ γ

), where γ

is the SINR

value of MS

(c)

. It is conﬁrmed that 4-sector coopera-

tion exhibits the maximum per-user throughput for all

cases of the BS density. The per-user throughput per-

formance is enhanced as decreading the BS density,

since the signal power from the interfering coopera-

tion sector is reduced by the sparse number of BS. In

particular, the performance gap between 4-sector and

8-sector cooperation increases at the large BS den-

sisty.

The distributions for the total throughput consid-

ering the BS density variation are plotted in Fig. 7.

The total throughput can be quantiﬁed by the sum of

throughput for every MS in about 2.6 × 2.6 square

kilometers area. Increasing the BS density, the target

MSs received the desired signal are densily located

in the considering area. At the cooperation cover-

age area of eta = 1/4 case for 4 MSs, 16 MSs can

be located to receive the desired signal considering

eta = 1/2 case. Thus, it is observed that the total

throughput is improved by the large BS density con-

ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems

418

0 2 4 6 8 10 12 14

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

η = 1

η = 1/2

η = 1/4

Per−User Throughput [bps/Hz]

Cumulative Distribution Function

CT: 4−Sector

CT: 8−Sector

Figure 6: Per-user throughput for the cooperative zero-

forcing transmission.

0 0.5 1 1.5 2 2.5 3

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

η = 1/4 η = 1/2 η = 1

Total Throughput [Mbps/Hz]

Cumulative Distribution Function

CT: 4−Sector

CT: 8−Sector

Figure 7: Total throughput for the cooperative zero-forcing

transmission.

dition due to the high link-access capability of MSs.

The performance of the outage probability is eval-

uated and compared condiering the BS density varia-

tion. At the entire range of outage threshold γ

, BS

density eta = 1/4 using 4-sector cooperation exhibits

the minimum outage probability, where the amount of

the strong interference signal is reduced by the sparse-

ness of BSs. At the 0.1% outage probability, the SINR

gain for the 4-sector cooperation of eta = 1/4 case ex-

ceeds 6 dB when compared to eta = 1 case.

Employing the cooperative zero-forcing dirty pa-

per coding, the performanceof the effective SINR and

the outage probability are evaluated using the 2, 4, 8,

16-sector cooperation models for BS density η = 1/2,

when all of the cooperation sectors use full power

transmission.

5 CONCLUSION

In Manhattan environment, we describe the various

kinds of cooperation strategy as increasing the num-

ber of cooperation sectors and transmission node den-

sity of the network. Considering both the practical

linear scheme using cooperative zero-forcing beam-

forming and non-linear scheme using zero-forcing

dirty paper coding, it is determinedthat 4-sector coop-

eration can be feasible strategy for cooperative multi-

user transmission in Manhattan network.

REFERENCES

Ahmed, B. T. et al. (2003). On the capacity and interference

statistics of street W-CDMA cross-shaped micro cells

in Manhattan environment (uplink analysis). In Proc.

IEEE Vehicular Technology Conf.

Chizhik, D. et al (2003). Multiple-input-multiple-output

measurements and modeling in Manhattan. IEEE J.

Select Areas Commun., vol. 21, pp. 321–331.

Cho, H.-S. et al. (2000) Performance analysis of cross- and

cigar-shaped urban microcells considering user mobil-

ity characteristics. IEEE Trans. Veh. Technol., vol. 49,

no. 1, pp. 105–116.

Chiu, W.-Y. and Chen, B.-S. Mobile positioning problem in

Manhattan-like urban areas: uniqueness of solution,

optimal deployment of BSs, and fuzzy implementa-

tion. IEEE Trans. Signal Process., vol. 57, no. 12, pp.

4918–4929.

Fu, I.-K. et al. Deployment and radio resource reuse in

IEEE 802.16j multi-hop relay network in Manhattan-

like environment. In Proc. IEEE ICICS, 2007, pp. 1–5.

CooperationStrategiesforMulti-userTransmissioninManhattanEnvironment

419