Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.

3, June 2014

BEAMFORMING WITH PER-ANTENNA POWER

CONSTRAINT AND TRANSMIT ANTENNA
SELECTION USING CONVEX OPTIMIZATION
TECHNIQUE
Suban and Ann Caroline Jenifer

Department of Electronics and Communication Engineering, Velammal College of

Engineering and Technology, Madurai.

ABSTRACT
In this paper, transmit beamforming and antenna selection techniques are presented for the Cooperative
Distributed Antenna System. Beamforming technique with minimum total weighted transmit power
satisfying threshold SINR and Per-Antenna Power constraints is formulated as a convex optimization
problem for the efficient performance of Distributed Antenna System (DAS). Antenna Selection technique is
implemented in this paper to select the optimum Remote Antenna Units from all the available ones. This
achieves the best compromise between capacity and system complexity. Dual polarized and Triple
Polarized systems are considered. Simulation results prove that by integrating Beamforming with DAS
enhances its performance. Also by using convex optimization in Antenna Selection enhances the
performance of multi polarized systems.

KEYWORDS
Distributed Antenna System, Beamforming, Convex Optimization, Antenna Selection, Dual polarized,
Triple Polarized.

1. INTRODUCTION
Wireless communication systems have been developing and evolving in a furious pace in these
recent years. The number of mobile subscribers is growing tremendously in the past decades. The
early wireless systems consisted of a base station with a high-power transmitter which serves a
large geographic area. Each base station could serve only a small number of users and was costly
as well. Today, due to the advancement in technology, the cellular system consists of a cluster of
base stations with low-power radio transmitters. Now the total number of users served is
increased because of channel reuse and larger frequency bandwidth.

Next generation broadband wireless access systems are evolving towards distributed architectures
as a promising solution to meet the ever increasing demand for the wireless connectivity. With
limited transmit power and bandwidth; one way to achieve the high data rate is to reduce the
radio transmission distance between the transmitter and the receiver which is made possible by
using the distributed architectures [1]. It is well established that multiple-input multiple-output
DOI : 10.5121/sipij.2014.5306 59
 Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.3, June 2014
(MIMO) technology provides high data rates and link reliability without additional bandwidth or
power [2]. Efficient allocation of the transmit power over the coverage area is possible by
introducing the MIMO technologies into new radio architectures especially for future mobile
communication systems [3]. Distributed Antenna System (DAS) is an evolving architecture
which serves the need of the future wireless communication systems.

Beamforming or spatial filtering is a powerful signal processing technique used in antenna arrays
or sensor arrays for directional signal transmission or reception. Beamforming helps in main lobe
enhancement, side lobe reduction and removal of interference caused by unwanted transmitters.

There are also a set of limitations that result from practical system design such as a Per-Antenna
power constraint that keeps the amplifier at each transmit antenna in its linear range [4]. In
designing the communication system, these types of constraints should all be taken into account.
Instead of Sum power constraint, in this paper we consider the more realistic Per-Antenna power
constraint on each of the transmitters at the Remote Antenna Units (RAUs), since in practice each
antenna is limited individually by its equipped power amplifier.

In paper [5], the transmitter optimization problem for multiuser downlink channels with multiple
transmit antennas at the base-station is considered. Assuming perfect channel knowledge at the
transmitter, this paper investigates two different transmission schemes under the Per-Antenna
power constraint: a minimum-power beamforming design for downlink channels with a single
antenna at each remote user and a capacity-achieving transmitter design for downlink channels
with multiple antennas at each remote user. It is shown that in both cases, the Per-Antenna
downlink transmitter optimization problem may be transformed into a dual uplink problem with
an uncertain noise. This new interpretation of duality gives rise to efficient numerical
optimization techniques for solving the downlink Per-Antenna transmitter optimization problem.

The benefit of coordinating base-stations across multiple cells in a multi-antenna beamforming

system, where multiple base-stations may jointly optimize their respective beamformers to
improve the overall system performance is considered in [6]. This paper focuses on the design
criteria of minimizing either the total weighted transmitted power or the maximum per-antenna
power across the base-stations subject to Signal-to-Interference-and-Noise-ratio (SINR)
constraints at the remote users. The main contribution of this paper is an efficient algorithm,
which is capable of finding the globally optimal downlink beamforming vector across all base-
stations.

2. COOPERATIVE DISTRIBUTED ANTENNA SYSTEMS

The Cooperative Distributed Antenna System (DAS) is a promising system for future wireless
communications. Unlike the conventional collocated antenna systems, where antennas are
centrally collocated, the Remote Antenna Units (RAUs) of DAS are geographically distributed in
the cell. RAUs are simple antenna units which carry out transmission and reception for its Base
Station (BS) or Central Unit (CU).RAUs are connected via optical fibers to the central unit (CU)
or the BS [1]. Thus it minimizes the communication overheads of system coordination.
Complicated signal processing algorithms are performed at the BS. DAS cellular model is shown
in Fig.1.

60
 Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.3, June 2014
Since all the RAUs are connected to the BS, spatial diversity can be achieved, thus improving the
system performance. Also, having some antennas located closer to the mobile station decreases
the average distance of propagation to or from the nearest antenna, thus reducing the required
uplink and downlink transmitted power for a fixed channel quality and creating more uniform
coverage inside the cell [7]. DAS is often used in scenarios where alternate technologies are
infeasible due to terrain or zoning challenges. The idea works because less power is wasted in
overcoming penetration and shadowing losses, and because a line-of-sight channel is present
more frequently, leading to reduced fade depths and reduced delay spread. In the cooperative
DAS the transmitted signals of a few RAUs, instantaneously linked to a reference MT, are
considered useful signals rather than interference.

Fig.1. DAS cellular model

DAS was originally proposed to improve the indoor performance of wireless communication
systems. In order to exploit the advantages of both the collocated MIMO and the DAS, a
cooperative DAS based on a radio over fiber (RoF) technique was introduced in China’s beyond
3G FuTURE project [8, 9] and tested via field experiments. Research on distributed MIMO has
been carried out in [10–12] by exploiting both microscopic and macroscopic spatial diversity. In
[3] an analytical capacity study was presented for DASs.

A set of RAUs within reachable distances of a reference MT can cooperatively communicate with
this MT to form a MIMO system. Note that in cooperative DAS the distributed MIMO system
differs from the traditional centralized MIMO since each of the RAUs experiences independent
macroscopic fading from each other [11]. Therefore, the distributed MIMO formed by several
RAUs at different geometric locations can make use of the statistically independent properties of
the channels more efficiently, and larger channel capacity would be expected than in the
traditional collocated MIMO. The theoretic and Monte Carlo simulation results presented in [3]
shows the benefits of the distributed MIMO.
61
 Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.3, June 2014

3. BEAMFORMING AND PER-ANTENNA POWER CONSTRAINT

Beamforming is a signal processing technique used in antenna arrays for directional transmission
and reception. It steers all the energy in one direction by directing a transmitting element in an
antenna array to reach a desired receiver in a given direction. Using Beamforming technique,
weight vectors are assigned for each channel. The following assumptions are made 1) Perfect
Channel State Information at the transmitter (CSIT) 2) The channel is assumed to be a frequency
flat quasi static channel 3) The number of users are greater than the number of transmit antennas
at the RAU.

Depending upon the CSIT, Beamforming technique allocates weight vectors for each channel.
Now, the transmit antenna element in the RAU whose channel is assigned a high value of weight
vector is chosen from among the available set of antenna elements and transmission takes place
between the selected antenna terminal in the RAU and the mobile terminal through this channel.

Also in practical multi-antenna implementations, each transmit antenna is usually equipped with
its own power amplifier. Thus, an individual power constraint on each antenna individually is
more realistic than a sum power constraint across all the antennas. There are certain benefits in
considering the Per-Antenna power constraint. They are as follows

• It helps in joint transmission and reception.

• Since the power is equally divided among the transmit antenna elements, all the antennas
will be active at a particular time.
Hence in this paper we consider the Per-Antenna power constraint.

4. SYSTEM MODEL

Consider a DAS scenario where a Base Station (BS) is located at the center of the macro cell and
the multiple antennas of the BS, the Remote Antenna Units (RAUs) are distributed throughout
the cell. The RAUs are connected to the BS by an optical fiber or cable. Each RAU consists of
one or more antenna terminals. Mobile Terminals (MTs) can be located anywhere in the cell. The
system model is shown in Fig.2.

Fig.2. System Model (showing a RAU with 2 antenna elements and a MT with 1 antenna)

62
 Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.3, June 2014
The general system model for a RAU with N antenna elements and K receivers, each with one
antenna can be given as,

where, yk represents the received signal (1x1), is a (1xN) beamforming vector, hk is a (Nx1)
channel response vector and nk represents the noise (1x1).

The channel capacity equation of the distributed antenna system is given as [15].

C(G)=log2 det(IM + ρ M D ) (1)

Where G is the downlink propagation channel, ρ is the SINR, D is the Large Scale Fading matrix
and M is the number of RAUs.

5. BEAMFORMING IN DAS AS A CONVEX OPTIMIZATION PROBLEM

Let wk denote the weight vectors (k=1, 2...K). The total weighted power at the transmitter is
given as Increasing the transmit power affects the linearity of devices (e.g.
amplifier) present in the RAU. This is an undesirable effect. Hence the aim is to reduce the total
weighted transmit power but at the same time ensure that the Signal to Interference-Noise ratio
(SINR) target is met at the receiver. We formulate the Beamforming in cooperative DAS as a
convex optimization problem to get the solution. Convex optimization technique finds the weight
vector that minimizes the total transmit power and also meets the SINR target. The two
constraints to be satisfied are the Per-Antenna power constraint and multiuser transmit
beamforming under individual SINR constraint. The problem can be expressed as follows

(2)

(3)

(4)

Eqn (2) defines the objective function (minimize transmit power), (3) defines the SINR constraint
and (4) defines the Per-Antenna power constraint. N denotes the number of transmit antennas, ck
denotes the target SINR, denotes the noise power. The problem becomes infeasible when 1)
the channel vectors of two or more channels are co-linear or highly correlated 2) SINR target is
too high 3) the number of users, K is much greater than the number of antenna terminals, N in the
RAU.

63
 Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.3, June 2014

6. TRANSMIT ANTENNA SELECTION IN DAS AS A CONVEX

OPTIMIZATION PROBLEM
Antenna Selection is a powerful signal processing technique that can reduce the cost and
complexity of RF chains associated with each RAU, but at the same time preserves the diversity
and multiplexing gains obtained from the actual system. Also it reduces the signalling burden
overhead at the BS. The transmit antenna selection problem in DAS can be formulated as a linear
programming problem and can be solved using the Interior Point algorithm. Linear programming
is a mathematical optimization technique to solve a linear objective function, subjected to the
linear equality and linear inequality constraints. Define as the antenna selection
variable for each RAU such that,

Define a M X M diagonal matrix ∆ for RAU selection at the Base Station which has ∆i as its
diagonal entries. The diagonal matrix ∆ is represented as follows,

The modified channel capacity equation as a function of antenna selection variable can be written
as

C(∆)=log2 det(IM+ ρ M ∆ D) (5)

The set of RAUs that maximizes the capacity of the system must be selected out of all the
available RAUs. The antenna selection problem is formulated as a convex optimization problem.
The objective function is to maximize the channel capacity. The two constraints defined are: 1) 0
≤ ∆ ≤ 1, the antenna selection variable is relaxed to a weaker constraint so that it can be solved in
polynomial time. 2) the sum of the diagonal elements of ∆ must be equal to N, the total number
of RAUs selected. The problem can be expressed as follows,

(6)

The solution to this problem would be that N optimal RAUs will be selected from the available
set to serve a particular user.

64
 Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.3, June 2014

7. RESULTS AND DISCUSSION

For the performance evaluation, the simulation parameters are set as follows: Number of users
K=2, Number of transmit antennas in the RAU Mt=2, Number of receive antennas Mr=1, SINR
target ck=2 dB. Two cases are taken to evaluate the formulated problem. Case 1: Two orthogonal
channels are taken. Case 2: Two highly correlated channels are taken. In Case 1, weight vectors
are found for link 1 and link 2. By using convex optimization technique, the optimized value of
the weight vector is found to be 0.7068. Optimal value of transmit power is found to be 2.8280
dB. Table 1 tabulates the optimized values obtained for Case 1.

Table I
Optimal Values Obtained for Case 1

In Case 2, two channels which are highly correlated are taken. As discussed before, this case
becomes infeasible. Hence no weight vectors are assigned. Simulation results for SNR in dB vs.
average Bit Error Rate (BER) is shown in Fig.3. Graphs are obtained for four different cases i)
1x1 MIMO ii) 2x2 MIMO iii) 2x2 MIMO-BF (Beamforming) iv) 4x4 MIMO-BF.

Fig.3.Plot of SNR in dB vs. Average BER

65
 Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.3, June 2014
It is evident from the graph that to attain a minimum average BER of 10-1, SNR needed by the
1x1 MIMO and 2x2 MIMO are 10 dB and 2dB respectively. Thus by increasing the number of
transmit and receive antennas, a minimized average BER can be obtained at a reduced SNR. Also
including Beamforming technique with MIMO further reduces the SNR (and hence the transmit
signal power) needed to achieve a minimized average BER. From Fig.3 we can see that to get a
minimum average BER of 10-1, SNR needed by 2x2 MIMO-BF and 4x4 MIMO-BF are -1 dB
and -7.5 dB. Table 2 shows the SNR values obtained for different system configurations.

Table 2 Comparison of SNR Values for Different System Configurations

Simulation result for SINR in dB vs. BER for two methods- sum-power constraint and per-
antenna power constraint is shown in Fig.4.From the plot it is seen that the performance of per-
antenna power constraint is better than the sum-power constraint. Therefore by implementing the
more realistic Per-Antenna power constraint, the amplifiers at the RAU are kept in the linear
range. Thus it is clear that by integrating Beamforming techniques with DAS, a reduced total
transmit power can be achieved with target SINR constraint. Also, it is clear that Beamforming
technique in DAS with per-antenna power constraint achieves a better performance.

Fig.4. Plot of SINR in dB vs. BER

66
 Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.3, June 2014
Simulation result between Selected RAUs and channel capacity without Antenna Selection is
shown in Fig.5. From this figure it is observed that the performance of 2SS-DP and 3SS-SP are
similar.

Fig.5. Plot of Selected Antennas vs. Capacity (without Antenna Selection)

Simulation results between Selected Antennas vs. Channel capacity with Antenna Selection is
shown in Fig.6. It is evident from the two plots that the performance of Dual Polarized and Triple
Polarized systems is enhanced by using Antenna Selection Techniques.

Fig.6. Plot of Selected Antennas vs.Capacity (with Antenna Selection)

67
 Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.3, June 2014

8. CONCLUSION
The problem to find the optimal beamforming weights which minimizes the total weighted
transmit power, satisfying the SINR constraint and Per-Antenna power constraint, is formulated
as a convex optimization problem and is solved using convex optimization tools (e.g. CVX). Two
different cases- orthogonal channels and highly correlated channels are discussed. Comparison
between Sum power and Per-Antenna power constraint is done. The results and graphs obtained
prove that the performance of the system is improved by integrating beamforming technique with
Cooperative Distributed Antenna System (DAS). Also convex optimization is used to optimize
the performance of DAS for maximizing the ergodic capacity. The optimum RAUs that
maximizes the channel capacity is selected by using optimization algorithm. The results are also
compared with spatially separated single polarized systems. The performance of multi polarized
system is enhanced by using convex optimization. Also the performance of triple polarized
system is found to be good under certain channel conditions.

REFERENCES

[1] H.Zhu, “On Frequency Reuse in Cooperative Distributed Antenna System”, IEEE Communications
Magazine, vol. 50, No. 4, pp.85-89, April 2012.
[2] S.Mahboob, R.Ruby and V.C.M. Leung, “Transmit Antenna Selection for Downlink Transmission in
a Massively Distributed Antenna System using Convex Optimization”, in Proceedings of Seventh
International Conference on Broadband, Wireless Computing, Communication and Applications,
2012, pp. 228-233.
[3] X.H.You, D.M.Wang, B.Sheng, X.Q.Gao, X.S. Zhao, and M.Chen, “Cooperative Distributed
Antenna Systems for Mobile Communications”, IEEE Wireless Communications, vol. 17, pp. 33-43,
June 2010.
[4] M. Khoshnevisan and J.N. Laneman, “Power Allocation in Multi-Antenna Wireless Systems Subject
to Simultaneous Power Constraints”, IEEE Transactions on Communications, vol. 60, pp 3855-3864,
Dec. 2012.
[5] W.Yu and T.Lan, “Transmitter Optimization for the Multi-Antenna Downlink with Per-Antenna
Power Constraints”, IEEE Transactions on Signal Processing, vol.55, pp. 2646-2660, June 2006.
[6] H.Dahrouj, W.Yu, “Coordinated Beamforming for the Multicell Multi-Antenna Wireless System”,
IEEE Transactions on Wireless Communications, vol.9, pp. 1748-1759, May 2010.
[7] R.Heath, S.Peter, Y.Wang and J.Zhang, “A Current Perspective on Distributed Antenna Systems for
the Downlink of Cellular Systems”, IEEE Communication Magazine, vol. 51, pp. 161-167, April
2013.
[8] X.H.You and X.Q.Gao, “Development of Beyond 3G Techniques and Experiment System: An
Introduction to the FuTURE project”, in ICT Shaping the World: A Scientific View, Wiley, Nov.
2008.
[9] X.H.Yu, G.Chen, M.Chen and X.Gao, “Toward Beyond 3G: the FuTURE Project in China”, IEEE
Communications Magazine, vol.43, pp.70-75, Jan.2005.
[10] W.Choi and J.G.Andrews, “Downlink Performance and Capacity of Distributed Antenna Systems in
a Multicell Environment”, IEEE Transactions on Wireless Communication, vol. 6, pp. 69–73, Jan.
2007.
[11] H. Hu, Y. Zhang, and J. Luo, Distributed Antenna Systems: Open Architecture for Future Wireless
Communications, CRC Press, June 2007.
[12] D.M.Wang, X.You, J.Wang and Y.Wang, “Spectral Efficiency of Distributed MIMO Cellular
Systems in a Composite Fading Channel,” in Proceedings of IEEE International Conference on
Communications ICC’ 08, 2008, pp.1259–64.

68
 Signal & Image Processing : An International Journal (SIPIJ) Vol.5, No.3, June 2014
[13] F. Boccardi and H. Huang, “Optimum power allocation for the MIMO BC zero-forcing precoder with
per-antenna power constraints”, 2006 40th Annual Conference on Information Sciences and Systems.
pp 504, March 2006.
[14] S-R.Lee, S.H.Moon, “Optimal Beamforming Schemes and its Capacity Behavior for Downlink
Distributed Antenna Systems”, IEEE Transactions on Wireless Communications, vol. 12, no. 6, pp.
2578-2587, June 2013.
[15] F. Rusek, D. Persson, B. K. Lau, E. G. Larsson, O. Edfors, F. Tufvesson,and T. L. Marzetta, “Scaling
up MIMO: opportunities and challenges with very large arrays,” IEEE Signal processing Magazine,
2012.

AUTHORS
A. Suban, received B.E in the department of Electronics and Communication
Engineering from Anna university, Chennai and M.E in the discipline of Wireless
Technology from Thiagarajar College of Engineering, Madurai in 2011. He is
currently working as Assistant Professor in the Department of Electronics and
Communication Engineering, Velammal College of Engineering and Technology,
Madurai-625009, Tamil Nadu, India. His area of interest in Signal processing mainly
focused on MIMO techniques with beamforming, OFDM and power control
techniques.

R.Ann Caroline Jenifer is currently pursuing her M.E degree in Communication

Systems at Velammal College of Engineering, Madurai. She received her B.Tech
degree in Electronics and Communication Engineering from Karunya University,
Coimbatore in 2012. Her area of interest is Digital Signal Processing.

69