ISSN: 2278 – 1323

International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)

Volume 2 Issue 8, August 2013

A COMPARATIVE BEAMFORMING ANALYSIS OF

LMS & NLMS ALGORITHMS FOR SMART ANTENNA

1
Priyanka Yadav, 2Prof. Smita Patil

 beams with one beam

Abstract— Beamforming is a technique of directional signal
transmission and recitation. Now a days, because of an turned on towards the desired signal or a single beam
increasing demand on wireless communication (like RADAR, (formed by phase adjustment only) that is steered toward
satellite communication, mobile communication, broadcasting
the desired signal. The other type is the adaptive antenna
etc.) the beamforming technique plays a vital role for good
transmission and reception purpose. By changing the phase
array is an array of multiple antenna elements, with the
and amplitude of the exciting currents in each of the antenna received signals weighted and combined to maximize the
elements, it is possible to electronically scan the image lobe desired signal to interference plus noise power ratio [1].
and/or place nulls in any direction. In beamforming, each users A smart antenna system at the base station of a cellular
signal is multiplied with complex weights that adjust the mobile system is depicted in Fig.1[3]. It consists of a
magnitude and phase of the signal to and from each antenna. uniform linear antenna array for which the current
This causes the output from the array of antennas to form a amplitudes are adjusted by a set of complex weights using
transmit/receive beam in the desired direction and minimizes
an adaptive beam forming algorithm. The adaptive beam
the output in other direction. In this paper we use different
forming algorithm optimizes the array output beam
version of LMS adaptive algorithm (basically LMS and NLMS)
to steer the antenna beam in particular desired direction. The
pattern such that maximum radiated power is produced in
radiation pattern of for each case plotted in polar plot. Here we the directions of desired mobile users and deep nulls are
steer the beam of of the system in two way (1) Considering the generated in the directions of undesired signals
desired direction (position) are in uniformly distributed (2) representing co-channel interference from mobile users in
The desired direction are non uniformly distributed. We adjacent cells.
campier their simulation results of each case with each other. Prior to adaptive beamforming, the directions of users and
Adaptive signal processing sensor arrays, known also as smart interferes must be obtained using a direction-of- arrival
antennas .The smart antenna adaptive algorithms achieve the
(DOA) estimation algorithm.
best weight vector for beam forming by iterative means.

Index Terms— LMS algorithm, NLMS algorithm, Smart

antenna.

I. INTRODUCTION
Throughout the world, there is significant research and
development on smart antennas for wireless systems.
Smart antenna systems attract a lot attentions now and
believably more in the future, as it can increase the Fig.1 : Block diagram of smart antenna
capacity of mobile communication systems dramatically
[4].This is because smart antennas have tremendous The goal of direction-of-arrival (DOA) estimation is to use
potential to enhance the performance of future generation the data received on the downlink at the base-station
wireless systems as evidenced by the antennas’ recent sensor array to estimate the directions of the signals from
deployment in many systems. There are two basic types of the desired mobile users as well as the directions of
smart antennas. The first type is the phased array or multi interference signals [5]. The results of DOA estimation are
beam antenna, which consists of either a number of fixed then used by to adjust the weights of the adaptive beam
former so that the radiated power is maximized towards
Manuscript received Jul, 2013.
the desired users, and radiation nulls are placed in the
Priyanka Yadav M.E. Scholar, Electronics and Communication directions of interference signals. Hence, a successful
EngineeringI.E.S, IPS Academy,Indore, M.P., India. 08251833854.
design of an adaptive array depends highly on the choice of
Smita Patil Associate Professor, Department of Electronics and
Communication Engineering I.E.S, IPS Academy,Indore, M.P., India. the DOA estimation algorithm which should be highly

All Rights Reserved © 2013 IJARCET

2478
 ISSN: 2278 – 1323
International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 2 Issue 8, August 2013

accurate and robust. e(n) = d(n)-y(n) (2)

Adaptive smart antennas are the array antennas whose The estimated convolved signal
radiation pattern is shaped according to some adaptive r(n) y(n) yˆ (n) (3)
algorithms [1]. Smart essentially means computer control of
the antenna performance. The smart antenna radiation we arrive at the recursion for the LMS adaptive
pattern directs the beam towards the users of interest only & algorithm for updating the step as
nulls toward interference to improve the capacity of cellular
system. The adaptive beam forming algorithms takes the h(n) = h(n - 1) + 2*μe(n)x(n)
fixed beam forming process one step further & allows for the (4)
calculation of continuously updated array weights.
where μ is constant step and the filter taps can be
According to signal space information smart antenna can adaptively updated by using above recursive relation.
form directional beam in space with the adaptive beam
forming algorithm, achieving that the main beam aims at the
direction of the expected signal while the side lobe and nulls
aims at the interference. Now many adaptive algorithms have
been proposed on smart antenna. The NLMS algorithm and
LMS algorithm are most commonly used as a adaptive beam
forming algorithm .

II. LEAST MEAN SQUARE (LMS) ALGORITHM

Fig.2 : LMS adaptive Beamforming Network

The Least Mean Square (LMS) algorithm, is an adaptive

algorithm. LMS incorporates an iterative procedure that III. NORMALIZED LEAST MEAN SQUARE (NLMS)
makes successive corrections to the weight vector in the ALGORITHM
direction of the negative of the gradient vector which
eventually leads to the minimum mean square error. Beam The NLMS algorithm has been implemented in Matlab. As
forming is directly determined by the two factors. This the step size parameter is chosen based on the current input
algorithm can be applied to beam forming with the software values, the NLMS algorithm shows far greater stability
Matlab. The result obtain can achieve faster convergence and with unknown signals [6]. This combined with good
lower steady state error. The algorithms can be simulated in convergence speed and relative computational simplicity
MATLAB 7.10 version. make the NLMS algorithm ideal for the real time adaptive
Consider a L length LMS based adaptive filter in which ‘W’ echo cancellation system. As the NLMS is an extension of
is the weight vector updated in accordance with the statistical the standard LMS algorithm, the NLMS algorithms
nature of the input signal x(n) arriving from the antenna practical implementation is very similar to that of the LMS
array. An adaptive processor will minimize the error e(n) algorithm. Each iteration of the NLMS algorithm requires
between a desired signal d(n) and the array output y(n).The these steps in the following order [7].
knowledge of the received signal eliminates the need for 1. The output of the adaptive filter is calculated
beam forming, but the reference can also be a vector which is
somewhat correlated with the received signal. As shown in
Fig.2.
(5)
The output response of the uniform linear array is given by
y(n) = hˆ H (n)x(n) (1) 2.An error signal is calculated as the difference between
the desired signal and the filter output
We consider the adaptive filter where the input signal x(n) is E(n) = d(n) – y(n) (6)
convolved by an unknown h(n) filter (to produce y(n)) which
has an additive interference signal v(n) before being observed 3.The step size value for the input vector is calculated
as d(n).

The value of error signal estimation is

2479
www.ijarcet.org
 ISSN: 2278 – 1323
International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 2 Issue 8, August 2013

Beam pattern using NLMS (Amplitude response antenna pattern)

0

(7) -10

4.The filter tap weights are updated in preparation for the -20

next iteration.

Beam (dB)
-30
W(n+1) = W(n) + μ(n)e(n)x(n) (8)
-40

Each iteration of the NLMS algorithm requires 3N+1 -50

multiplications, this is only N more than the standard LMS -60

algorithm. This is an acceptable increase considering the
-70
gains in stability and echo attenuation achieve. NLMS -100 -80 -60 -40 -20 0 20 40 60 80 100
Angle (Degree)
converge faster than LMS because the step size is optimize
Fig.4 : NLMS Antenna Coverage Pattern
at each iteration. The computational complexity of the
more in NLMS than LMS.
Fig.4 shows the beam pattern for desired users with
signals. The plots correspond to the 200 iterations. The
IV. SIMULATION RESULTS
figure shows that the response is successfully steered
1. LMS ALGORITHM
towards the desired signal source and away from the
Beam pattern using LMS (Amplitude response antenna pattern)
0 strongest interfering sources. Number of snapshots are
-5 100, number of array elements N=16 wit h spacing
-10
between elements, d =0.05λ,SNR=0, JNR= 30. Working
frequency 3 GHz, Sample frequency 18 GHz.
-15
It has been noticed from results that sharper
Beam (dB)

-20
beams are directed towards desired signals as more
-25 elements are used in antenna array. Also, spacing between
-30 array elements has an effect on beam-former performance
-35 such that very small or very large spacing between array
-40
elements can degrade beam-former performance.
From different numerical calculation it has been
-45
-100 -80 -60 -40 -20 0 20 40 60 80 100 observed that element spacing of 0.05λ is a good value, the
Angle (Degree)
value of sharper beams (positive interference or desired
Fig.3 : LMS Antenna Coverage Pattern
user) is high in case of normalized least mean square
algorithm as compare to traditional least mean square
Here, Least Mean Square (LMS) algorithm uses a gradient
method.
based method of steepest decent. LMS incorporates an
iterative procedure that makes successive corrections to the 0
Beam pattern using NLMS (Amplitude response antenna pattern)

weight vector in the direction of the negative of the gradient

-10
vector which eventually leads to the minimum mean square
error. LMS algorithm is relatively simple; it does not require -20

correlation function calculation nor does it require matrix

Beam (dB)

-30

inversions.
-40
Fig.3 shows the radiation pattern for desired users
with signals arriving at angles -40,0,60 degrees with the -50

different signal strength. Number of snapshots are 2000, -60

number of array elements N=16 with spacing between

-70
elements,d =0.5λ. -100 -80 -60 -40 -20 0 20 40 60 80 100
Angle (Degree)

2. NLMS ALGORITHM Fig.5 : NLMS Antenna Coverage Pattern

Fig.5 shows the beam pattern for desired users with signals
arriving at angles 15,25,-15,0 degrees with the different
signal strength .Number of snapshots are 100, number of
array elements N=20 with spacing between elements ,d

All Rights Reserved © 2013 IJARCET

2480
 ISSN: 2278 – 1323
International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
Volume 2 Issue 8, August 2013

=1.0λ,SNR=0, JNR= 40,working frequency 4Ghz,Sample [10] Koteswara rao.Thokala , Ch.Jaya Prakash,“Steering an
frequency 18Ghz. adaptive antenna arrayby LMS NLMS and BBNLMS algorithms”,
Global Journal of Advanced Engineering Technologies”,ISSN:
It has been noticed from results that the radiation pattern
2277-6370, Volume1-Issue3-2012.
for NLMS case show faster convergence rate and can form [11] D. B. Salunke, R. S. Kawitkar,“Analysis of LMS, NLMS and
deeper nulls in the direction of interference for desired MUSIC Algorithms for Adaptive Array Antenna
users. System”,International Journal of Engineering and Advanced
Technology (IJEAT) ISSN: 2249 – 8958, Volume-2,Issue-3
February 2013.
V. CONCLUSION [12] Markus Rupp “The behaviour of LMS & NLMS algorithms in
the presence of spherically invariant process”, IEEE transaction on
In these algorithms, the LMS algorithm is the most popular signal processing volume-41Issue-3 ,march 1993.
[13] Kaustav Das, Gokulananda sahu, subrata kumar sahu“A
adaptive algorithm, because of their low computational Compression study of Antenna Beamforming Using LMS and
complexity. However, the LMS algorithm suffers from NLMS Adaptive Algorithm”.
slow and data dependent convergence behavior. The [14] S. Weiss, “An effecient scheme for broadband adaptive
beamforming,”IEEE, 1999.
NLMS algorithm, an equally simple, but more robust
variant of the

LMS algorithm, exhibits a better balance between simplicity

and performance than the LMS algorithm. Due to its good
characteristics the NLMS has been largely used in real-time
applications.
Priyanka Yadav has completed her B.E. in electronics &
VI. REFERENCES communication engineering in 2009 from SRITS Datia (RGPV
University Bhopal, M.P.). She is pursuing her M.E. in digital
[1]Shiann-Jeng Yu and Ju-Hong Lee,“Adaptive Array communication from IES, IPS Academy, Indore,India. At present
she is in 4th and final semester of M.E..
Beamforming Based on an Efficient Technique”, IEEE trans.
Smita Patil received the B.E. degree in 1997 from
Antennas and Propagation. 1996ˈ44(8) 1094-1101.
Department of Electronics and communication, S.S.G.M.CE,
[2] L. S. Reed, J. D. Mallett, “Rapid Convergence Rate in Adaptive
Shegaon, M.E. degree (Digital Communication) from I.E.T.
Arrays”, IEEE trans. Acoustics Aerospace and Electronic Systems.
DAVV, Indore. Presently working as Associate Professor in
1974, 10(6):853-863.
I.E.S., I.P.S. Academy, Indore, India.
[3] Komal R. Borisagar and Dr. G.R.Kulkarni, “Simulation and
Comparative Analysis of LMS and RLS Algorithms Using Real
Time Speech Input Signal”,.Global Journal of Researches in
Engineering Page 44 Vol.10 Issue 5 (Ver1.0) October2010.
[4] H. Wang and M. Glesner “Hardware implementation of smart
antenna systems”,Advances in Radio Science 4, 185– 188, 2006.
[5] Raed M. Shubair, Mahmoud A. Al-Qutayri, and Jassim M.
Samhan, “A Setup for the Evaluation of MUSIC and LMS
Algorithms for a Smart Antenna System”, Journal of
Communication, Vol. 2, No. 4, June 2007.
[6] Lee, K.A.; Gan,W.S; “Improving convergence of the NLMS
algorithm using constrained subbandupdates,” Signal Processing
Letters IEEE, vol. 11, pp. 736-739, Sept. 2004.
[7] D. Morgan and S. Kratzer, “On a class of computationally
efficient rapidly converging, generalized NLMS algorithms,” IEEE
Signal Processing Lett., vol. 3, pp. 245–247, Aug. 1996.
[8] Alok Pandey, L.D. Malviya, Vineet Sharma “Comparative
Study of LMS and NLMS Algorithms in Adaptive
Equalizer”,(IJERA) ISSN: 2248-9622,Vol. 2, Issue 3, May-Jun
2012, pp.1584-1587.
[9] M. Yasin, Dr. Pervez Akhtar, Dr. Valiuddin,“Performance
Analysis of LMS and NLMS Algorithms for a Smart Antenna
System”, International Journal of Computer Applications (0975 –
8887)Volume 4– No.9, August 2010.

2481
www.ijarcet.org