Selective Interference Rejection based Antenna Selection for MIMO over LTE Advanced Networks

Procedia Computer Science 93 ( 2016 ) 624 – 631

1877-0509 © 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the Organizing Committee of ICACC 2016 doi: 10.1016/j.procs.2016.07.249

ScienceDirect

6th International Conference On Advances In Computing & Communications, ICACC 2016, 6-8

September 2016, Cochin, India

Selective Interference Rejection based Antenna Selection for MIMO

over LTE Advanced Networks

Ramachandran Vijayarani*, Lakshmanan Nithyanandan

Department of Electronics and Communication, Pondicherry Engineering College, Puducherry and 605 014,India

Abstract

An important goal for long term evolution advanced (LTE-A) network is to enhance the cell edge user throughput while achieving the link reliability and high data rate communication. However, the cell edge users’ throughput is limited by inter-cell interference. To suppress the interference signal, interference rejection combining (IRC) receiver is an effective solution for Release 10/11. In Multiple-Input Multiple-Output (MIMO) technique, the IRC receiver utilizes multiple antennas, larger the number of antennas, the more interfering signals can be suppressed. Due to practical constraints of increase in radio-frequency units impose a limit in the number of antennas. In this paper, a new selective interference rejection based antenna selection (SIR-AS) scheme is proposed to perform the better selection of transmitting antenna for cell edge MIMO user and enhance the system throughput. The selective interference rejection is performed by incorporating the IRC receiver, where the covariance matrix (CM) of the received interference signals is estimated and the minimum of CM is selected. The corresponding antenna set with minimum of CM is selected for transmission. Simulation results show that the proposed algorithm can effectively detect intercellular interference at cell boundaries and select the optimal antenna to reduce block error rate (BLER) and improve throughput performance compared to ideal IRC and maximum ratio combining algorithm, making it suitable for LTE-A downlink receiver in multi-cell MIMO systems.

© 2016 The Authors. Published by Elsevier B.V.

Peer-review under responsibility of the Organizing Committee of ICACC 2016.

Keywords:antenna selection; inter-cell interference; interference rejection combining; long term evolution advanced ; multiple-input multiple output ; selective interference rejection;

* Corresponding author. Tel.: +91-7502123124; fax: +0-000-000-0000 .

E-mail address: [email protected]

© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

For LTE Advanced network, Multiple Input Multiple Output (MIMO) has been one of the emerging technology to meet the demands for higher data rate, spectral efficiency, network capacity and cell coverage without increasing the average transmit power or bandwidth 1_{. Furthermore, it has been widely incorporated to improve the average cell}

throughput and cell-edge throughput, including single-user MIMO (SU-MIMO) and multi-user MIMO (MU- MIMO)

2_{. It was proved that MIMO structure successfully constructs multiple spatial layers where multiple data streams are}

delivered on a given frequency time, resource and linearly increases the channel capacity and cell edge throughput1_.In

LTE-Advanced, the maximum number of layers supported in SU-MIMO is extended to eight to increase the spectral efficiency in twofold 3_{. In MIMO technology, the use of multiple antennas result in a significant increase in the channel}

capacity and performance, however the multiple antennas at the transceivers lead to additional requirement of radio-frequency (RF) chains, which increases the hardware complexity, implementation cost and energy consumption 4_{. To}

deal with this issue, antenna selection techniques are developed, in which only a small number of antennas among available antennas are selected for transmission, were proposed.

Antenna selection (AS) is a simple but powerful scheme as it could attain the benefits of the MIMO technique with only a small number of RF chains. In AS approach, only subsets of antennas are selected for transmission based on given selection parameters. Therefore, this technique achieves a low implementation cost and less feedback load compared with other beamforming/precoding techniques 5,6_{. Moreover, antenna selection is robust to channel}

estimation errors because the phase information is generally not required. Owing to these advantages, antenna selection has been considered for 4G LTE-advanced 7_{.Transmit antenna selection allows a reduction of the complexity of MIMO}

system as it was shown that it preserves the spatial diversity provided by a MIMO system while improving the channel capacity and system performance 8,9_{. Non codebook based precoding is used for multilayer demodulation reference}

signals(DMRS) which provides improved precoding flexibility and performance of Release 10 MIMO operations 3_{. In} 10_{, transmit antenna selection (TAS) with receive generalized selection combining have been discussed for cognitive}

decode and-forward (DF) relaying with Nakagami- fading channels. In cognitive MIMO networks TAS with maximal ratio combining (TAS/MRC) and with selection combining (TAS/SC) has been proposed which scales with the minimum number of antennas at the secondary users over Rayleigh fading channel 11_{. The secrecy outage performance}

with the consideration of the effect of weighting errors is analysed for transmit antenna selection/maximal-ratio combining system 12_.

On the other hand, the user equipments (UEs) located near the cell boundary is suffered from severe interference from adjacent cells which degrades the system performance. The use of IRC receiver for the cell-edge user is effectively improved the throughput under this condition 13_{. In} 14_{, the IRC receiver is investigated which suppresses}

the interference signals with the aid of multiple antenna branches while detecting the desired signal. Since the IRC receiver strictly generates the received weight matrix based on the correlation information of interference signals, the interference signals can be suppressed according to the spatial degrees of freedom. In 15_{, the IRC receivers for}

open-loop transmit diversity that employing the SFBC is investigated. The extended covariance matrix estimation has been performed where some unknown elements in the covariance matrix are determined with appropriate values and specific insertion of zero values is performed. However, these methods are not optimal for antenna selection in the presence of substantial co-channel interference, particularly for cell edge user.

In this paper, the selective interference rejection based antenna selection (SIR-AS) technique is proposed for cell edge user for SU-MIMO. The selective interference rejection has been performed by incorporating the IRC receiver, where the covariance matrix (CM) of the received interference signals is estimated and the minimum of CM is selected. The transmit antenna selection is achieved with the impact of interference covariance matrix. The antenna set with minimum of CM is selected for low complexity receivers. In the presence of channel estimation errors, the CM can be more easily estimated, hence CM based AS techniques give better error-rate performance for cell edge user than their channel state information based counterparts. The system performance is evaluated in terms of block error rate (BLER) and throughput under different channel conditions and the optimal AS is performed in practical receiver configurations at cell boundaries.

The rest of the paper is organized as follows. Section 2 describes the system model and operation of the IRC receiver. Section 3 presents the proposed SIR-AS algorithm. Section 4 shows simulation results and discussion. Finally, conclusions are drawn in Section 5.

Fig.1 Single user MIMO network

2. System Model

Consider a SU-MIMO spatial multiplexing model with K subcarriers, the eNodeBs (base station) employs M transmit antennas and serves the user equipment (UE), which has N receive antennas. Assume L is the number of OFDM symbols in a subframe. The system model is shown in Fig.1. The number of equipped transmit and receive RF chains are

N

Q_RF

d

M N

,

. This implies that, only Qinstead of N(M) transceiver chains have to be built, and also the signal processing can be simplified. The eNB allocates the resources and schedules the data for the downlink transmission to the users. It uses a feedback channel from the users in order to acquire channel state information (CSI). The receiver at each user estimates SINR received from each transmit antenna of eNB, and feedback using a reliable feedback channel to eNB. Assuming that a perfect channel estimation and error free feedback channel is available, the system model at serving UE can be written as

1,

( , )

( , )

( , )

( , )

( , )

( , )

B s s i s i i s

Y k l

G k l X k l

G k l X k l

z k l

z



¦



(1)

where is the effective channel matrix between the serving eNodeB and the target UE with MIMO channel matrix

H

_s

{

H H

₁

,

₂

,...,

H

_Q

}

and the corresponding applied precoding matrix

P k l

_s

( , )

. Similarly

G k l

_i

( , )

H k l P k l

_i

( , ) ( , )

_i is the channel matrix for the inter-cell interference from the ith _{interference cell.}

Both

P k l

s

( , )

and

P k l

i

( , )

are generated at eNodeB based on the UE feedback channel information. Xs(k,l) denotes

the transmit data vector at kth subcarrier of the lthOFDM symbol of a sub frame, which is called a resource element in LTE-A system and

z k l

( , )



N

u

1

independent identical distribution (i.i.d.) cyclic symmetric complex Gaussian random vector with covariance matrix of

V

2I.

The IRC receiver in 16 _{utilizes the knowledge of the covariance matrix of the total interference plus noise and}

prewhitens the received signals. The IRC receiver for the SU-MIMO transmission can be represented by





1 1

( , )

( , )

( , )

( , )

irc ZZ q H YX YY H H q s i i s

w

R R

k l G k l G

k l

R

k l

H

H

 



(2)

where

R

_YXis the cross correlation matrix between Y k l( , )and

X k l

_s

( , )

;

R

_YYis the auto-correlation matrix of Y k l( , );

ZZ

q

R

denotes the spatial covariance matrix of intercell interference noise for qth antenna from transceiver antenna

Table 1. Proposed selective interference rejection antenna selection algorithm

1. Input: H, G,

R

_zzq, Q, N, M

2. Initialization: A= {1,2,…, Q} % antenna set at Transmit & Receive

3. for n=1 : N 4. for m=1 : M

5. Estimate

R

_zzq for each n, m using eqn.[4] 6. Order the estimated ˆq

zz R by decreasing magnitude. 7. end for 8. end for 9. Check * {1,2...,Q} ˆ arg min { q} ZZ zz A R R 

10. For selected

R

*_zzin step:9 estimate SINR using eqn.[4]

11. Output: index set of selected antennas Q={n,m}

set, denoted as A {1,2,},Q}.

G

_idenotes the channel matrix of the interferers. The cell specific reference signal (CRS) based on estimation scheme has been presented in17 _{to estimate the covariance matrix of interference and noise.}

The use of CRS subcarriers is the key idea to estimate the covariance matrix of inter-cell interference and noise.

R

_zzq

is obtained through the average estimate of each pilot’s position. The estimation of

ˆ

ZZ q

R

is given by ,

ˆ

[

( , )

( , )

( , )]

1

ˆ

ˆ

[

( , )

( , )

( , )]

p p p p q zz _H k l N p _{p p p}

Y k l

H

k l X

k l

R

N



_{Y k l}

_H

_{k l X}

_{k l}

§



·

¨

¸

¨

_u

_

¸

©

¹

¦

(3)

where subscript p denotes the pilot subcarrier position, N_pis the number of pilot resource elements in a sub frame and

H

ˆ ( , )

_p

k l

is the estimated channel response at pilot subcarrier p.The signal-to-interference plus noise ratio(SINR) of the IRC receiver is expressed as

1 2 1,i ( , ) H( , )

( , )

( , )

s s i i B H s i s k l k l

H

H

G k l G

k l

I

U

V

 z § · ¨ ¸ ©

¦



¹ (4)

where

G

_i

[

g g

_{i1 i2}

...

g

_iB

]

is the channel matrix of the interferers. 3. Proposed Selective Interference Rejection Antenna Selection Technique

In proposed SIR-AS technique, the covariance matrix of received interference signals

R

_zzqis estimated at UE using eqn.(3) and the values are sorted by decreasing magnitude that is,

^

R

ˆ

_zz1

,

R

ˆ

_zz2

,

,

R

ˆ

_zzQ

`

. The transceiver antenna set

with minimum of CM is selected compared with other sets and SINR of the corresponding antenna set is estimated using eqn.(4), which provides the highest SINR at the receiver for transmission.This process is repeated until all N receiver branches have been examined. After the selection process terminates, the selected branches are sent for further receive processing and is fed back to eNB through feedback channel. The same estimation for AS is repeated for other transmit antennas i.e., m

^

1, 2,...,M

`

with different N values. The optimal selection of CM is expressed as

* {1,2...,Q} ˆ arg min { q} ZZ zz A R R 

(5)

From the transceiver antenna set, i.e.,A {1,2,},Q}, the qth antenna sets are selected for combining, wherein the selection algorithm depends on the type of receiver that is implemented. Table 1 summarizes the proposed antenna selection algorithm.

4. Simulation and Discussion

Consider a multi-cell LTE-A downlink MIMO system with 19 cells and each cell is equipped with one base station serving 3 sectors. The LTE-A systems is developed according to the requirements stated in 18 _{and the simulation}

parameters are summarized in Table 2. Simulation is carried out with two channel models, namely, the ITU Extended Pedestrian A (EPA) and Extended Vehicular A (EVA), whose maximum Doppler frequency is 5 Hz and maximum delay spreads are 45 ns and 357 ns, respectively 19_{. The downlink system bandwidth is 10MHz, which corresponds to}

50 RBs (Resource Blocks) in LTE-A standard.

The simulation results of SU-MIMO with different receiver based antenna selection in the target cell are depicted in Fig. 2 and Fig. 3. Receiver performance is examined in terms of BLER and throughput. The antenna selection with

m = n = 1 (1

u

1) and m = n = 2 (2

u

2) over 4

u

4 MIMO is discussed.

Fig. 2(a) shows BLER comparison of different AS algorithms in EPA-5 channel. It can be observed that SIR algorithm outperforms MRC and ideal IRC method. The antenna selection with SIR receiver using estimated selective CM outperforms the MRC using estimated CSI and ideal IRC using estimated CM. MRC antenna selection with estimated CSI is degraded significantly due to high interference at cell boundaries when compared against ideal IRC and proposed SIR method. Furthermore, the antenna selection algorithm of proposed scheme with selective CM yields better BLER compared with ideal IRC receiver scheme due to selection of minimum CM of interference to noise ratio, which maximizes the SINR of the corresponding antenna pair. For q = 2 antenna set provides better error performance than q = 1due to additional diversity gain offered by the use of multiple antennas, as well as utilizes MIMO advantage with the reduced hardware complexity. It can be observed that the average BLER is reduced as the number of transceiver antennas Q increases.

Fig. 2(b) shows BLER comparison of different AS algorithms in EVA-5 scenario. Due to the increased residual inter cell interference the complete system performance degrades in Fig. 2(b) compared with results in Fig. 2(a). Moreover, for EPA-5 case, the channel is flat in the frequency domain due to its small delay spread. Therefore the noise covariance matrix accurately reflects the interference for each subcarrier. For EVA-5 case the channel estimation is more noisy due to relative stronger residual inter cell interference which affects the

Table 2. Simulation Parameters

Parameters Values

Network pattern 19 cells, each with 3 sectors Carrier frequency 2 GHz

System bandwidth 10 MHz

Antenna configuration (M x N) 4

u

4 (Uncorrelated) Antenna gain 14 dBi

Channel Model EPA-5, EVA-5 Total eNodeB Tx power 46 dBm Receiver Noise Figure 9 dB Traffic model Full buffer

UEs/sector 10

Min. distance from UE to BS ≥ 35 m

Thermal noise spectral density -174 dBm / Hz

Fig. 2.BLER

performance of

different antenna

selection algorithms versus SINR (a) EPA-5; (b) EVA-5 channel

precoding operations. The proposed AS based on selective CM estimation degrades significantly due to interference from other cells. However, its performance is better than other AS schemes which indicates better accuracy of selected CM based AS is achieved.

Fig. 3 depicts the comparison of throughput of the different schemes against SINR in EPA-5 and EVA-5 channel scenario. From Fig. 3(a), it can be observed that the proposed SIR-AS outperforms other schemes. The degradation of throughput performance can be seen in Fig. 3(b) compared with Fig. 3(a) is due to high residual interference in EVA-5. It can be seen from the simulation results of MRC receiver, the average cell throughput is different for both channel conditions due to increase of interferers. When the ideal IRC receiver is used, some performance differences are revealed because of the IRC receiver is more sensitive to the interference. Table. 3 summarizes the comparison of the average user throughput of different AS scheme. The gain of proposed SIR receiver is approximately 5% and 21% improved over 1

u

1 SISO and 2

u

2 MIMO, respectively compared with the ideal IRC receiver. Therefore, jointly considering the overall performance results and system impacts on the inter cell interference, the proposed selective interference rejection based AS provides a better performance in MIMO over LTE-A.

Table 3. Comparison of Average User Throughput

(a) EPA-5

AS Scheme Average User Throughput (Mbps) Gain MRC receiver 16.01 0.0% Ideal IRC receiver 18.54 15.8% Proposed SIR receiver (1

u

1) 20.02 25.0% Proposed SIR receiver (2

u

2) 22.01 37.4% (b) EVA-5

AS Scheme Average User Throughput (Mbps) Gain MRC receiver 14.21 0.0% Ideal IRC receiver 16.02 12.7% Proposed SIR receiver (1

u

1) 18.20 28.0% Proposed SIR receiver (2

u

2) 20.01 35.0%

5. Conclusion

In this paper, the selective interference rejection based AS algorithm has been proposed for MIMO transmission over LTE Advanced networks. It is shown that the proposed antenna selection algorithm based on the selective estimated covariance matrix in the interference region selects the optimal antenna, which reduces the overall computational complexity and maintains the superior BLER performance simultaneously than that of selection based on the estimated channel coefficients for the interferers. Moreover, AS based on the minimum CM of the received signal potentially alleviates the interference problems at cell boundaries and effectively suppress the co-channel interference with SIR receiver, thus keeping cost, size, and energy consumption at low levels.

Simulation results have shown that the antenna selection technique can improve the performance of SU-MIMO systems of cell edge user and, hence, achieves a high gain in throughput and error-rate performance compared to other ideal IRC and MRC techniques. In future, the feedback channel for MIMO with lower delay can be designed for effective antenna selection.

References

1. Lee J, Han JK Zhang JC. MIMO technologies in 3GPP LTE and LTE-advanced. EURASIP J Wirele.Commun. Netw, 2009;1:1–10. 2. Lim C, Yoo T, Clerckx B, Lee B and Shim B. Recent trend of multiuser MIMO in LTE-advanced. IEEE Commun Mag 2013;51:127–135. 3. Liu L, Chen R, Geirhofer S, Sayana K, Shi Z and Zhou Y. Downlink mimo in lte-advanced: Su-mimo vs. mu-mimo. IEEE Commun. Mag.

2012;50:140-147.

4. Haccoun D, Torabi M and Ajib W. Performance analysis of multiuser diversity with antenna selection in MIMO MRC systems. PhysicalCommun. 2010;3:276–286.

5. Molisch AF and Win MZ. MIMO systems with antenna selection. IEEE Microwave Mag. 2004; 5:46-56.

6. Vithanage CM, Coon JP and Parker SC. On capacity-optimal precoding for multiple antenna systems subject to EIRP restrictions. IEEE Trans. Wireless Commun. 2008;7:5182–5187.

7. Mehta NB, Kashyap S and Molisch AF. Antenna selection in LTE: from motivation to specification. IEEE Commun. Mag. 2012;50:144-150. 8. Love D. On the probability of error of antenna-subset selection with space–time block codes, IEEE Trans. Commun. 2005; 53:1799–1803. 9. Torabi M. Antenna selection for MIMO-OFDM systems, Signal Process. 2008; 88: 2431–2441.

10. Deng Y, Wang L, Elkashlan M, Kim KJ and Duong TQ. Generalized Selection Combining for Cognitive Relay Networks Over Nakagami-Fading. IEEE Trans. Sig. Process. 2015;63: 1993–2006.

11. Yeoh PL, Elkashlan M, Kim K, Duong T, Karagiannidis G. Transmit antenna selection in cognitive MIMO relaying with multiple primary transceivers.IEEE Trans. Veh.Tech.2016;65: 483–489.

12. Hu Y, Tao X.Secrecy outage on transmit antenna selection with weighting errors at maximal-ratio combiners.IEEE Commun. Lett.2015; 19: 597–600.

13. Ohwatari Y. Morimoto A, Miki N, OkumuraY. Investigation on interference rejection combining receiver in heterogeneous networks for LTE-Advanced downlink. In IEEE 14th Workshop on Signal Processing Advances in Wireless Communications (SPAWC). 2013;p.315-319.

14. Shirakabe M, Morimoto A. Performance evaluation in heterogeneous networks employing time-domain inter-cell interference coordination and cell range expansion for LTE-advanced downlink. IEICE Trans. Commun.2012;95:1218–1229.

15. Ohwatari Y, Miki N, Sagae Y and Okumura Y. Investigation on interference rejection combining receiver for space-frequency block code transmit diversity in lte-advanced downlink. IEEE Trans. Veh. Technol. 2014; 63: 191–203.

16. Bai Z, Badic B, Iwelski S, Scholand T, Balraj R, BruckGH and Jung P. On the receiver performance in MU-MIMO transmission in LTE. inProc of the Seventh Int.Conf. on Wireless and Mobile Communications, ICWMC 2011; p. 128–133.

17. Reference receiver structure for interference mitigation on enhanced performance requirement for LTE UE. NTT DOCOMO; Approval R4– 115213. Oct. 2011.

18. 3GPP, ’Further advancements for E-UTRA physical layer aspects’, 3GPP TR 36.814 v9.0.0, Mar. 2009.