The research paper published by IJSER journal is about Interference Reduction and DOA Estimation in GSM Systems 1
ISSN 2229-5518
Nsionu I I, Onoh G N, Ekeh J, Eneh I I
Index Terms: Interference Reduction, Adaptive, Uniform Linear Array, Multiple Signal Classification, Direction of Arrival, Multipath,
—————————— ——————————
There is an unprecedented growth in the demand for wireless communication services. The obvious challenge for network providers is the provision of quality voice and high speed data services. Following initial system deployment, subscribers enrollment tend to increase on daily basis and attempts to accommodate every new subscriber gives rise to unavoidable interference. As a result, network capacity drops.
Interference is of three categories: Intersymbol-signal delay which arises due to multipath. This results in poor BER performance. Next is Co-channel interference which arises due to multiple access. This results in limitation in network capacity. Then Adjacent Channel interference which arises due to signals adjacent in frequency to the desired signal. This also results in network capacity limitation.
Interfering signals arrive at the receiver from different directions in space. Therefore, spatial processing will be required to check the menace. Each interfering signal has distinct spatial signature which will be exploited by adaptive uniform linear arrays (or Smart Antennas). One of the spatial signatures is the Direction of Arrival, DoA which will be explored using intelligent signal processing algorithms.
The two main functions of the adaptive antenna are Direction of Arrival Estimation and Array Pattern Synthesis. This paper concentrates on the Direction of Arrival Estimation. This spatial signature enables the system to determine which signal to accept or reject.
Let us consider 4 signals impinging on 32 element arrays with element spacing 0.5λ, where λ corresponds to the carrier frequency. Let the directions of arrival of the signals be 14o , 28o , 35o and 55o
The goal of DoA estimation is to use data at array to estimate 1,2 ,3and4. Of course, the inherent challenges in situations like this in real time are (1) unknown number of signals, (2) unknown directions (3) unknown amplitudes (4) signal corrupted by noise.
This paper shall consider the adoption of the subspace based method called music which stands for Multiple Signal Classification. The space is divided into signal
and noise subspace. Music exploits the noise eigenvector subspace. The noise space is spanned by eigenvectors corresponding to the smallest eigenvalues of the array
correlation matrix. Given that D signals impinge on M
IJSER © 2012
The research paper published by IJSER journal is about Interference Reduction and DOA Estimation in GSM Systems 2
ISSN 2229-5518
elements; the number of signal eigenvalues and
eigenvectors is D, and the number of noise eigenvalues and eigenvectors is M – D.
The array correlation matrix, Rxx is given by [1],
Obtain sample data from the array
Compute array (covariance)
correlation matrix.
R A R AH
2 I
Where Rss source correlation matrix
Is coherent signal present?
yes
R 2 I noise correlation matrix
Apply spatial smoothing technique
No
A [a(1)
vector.
a(2 )
a(3 )....
a(D )] is an M D steering
Perform the eigen decomposition of the estimated data correlation matrix (eigen values and eigen vectors)
Compute the DOA of the signals
Rss [s1 (k )
s2 (k )
s3 (k )...
s (k )]T is a D D matrix.
The array correlation matrix has D eigenvectors associated with signals and M D eigenvectors associated with noise
1 Direction of arrival estimation flowchart
Fig.
The subspace such that
M (M D) is spanned by the noise vectors
The music technique is simulated using Matlab. The simulation was run for 4 signals coming from different
EN [e1 e2
e3 eM D ]
directions as listed in the problem setup and which impinge
on a 32 element array. To run the program, it has to be
Noise space eigenvectors are orthogonal to the steering
ensured that all files regarding the Music algorithm are in
vectors at the angles of arrival
1,2 ,3......,D .
MUSIC
the same directory. The main programme is run in
formula from the point of view of orthogonality condition,
can be shown to be [1 ], [2]
Matlab. The number of elements is entered, spacing between elements, number of desired signals with their
Pmusic
[a( )H E
1
E H a( )]
which yields sharp peaks at
corresponding angles of arrival. The program gives the estimated number of signals as output along with the array
the angles of arrival
pattern plots. All that is required is to obtain the values of the highest peaks to get the direction of arrival [3]
The flowchart for the estimation of the direction of arrival is shown in fig. 1.
All 4 signals and their directions are resolved accurately as shown in the table below containing the desired signal direction and the actual output of the system [4]
Table 1: Table of values of the actual array output and the intended system output
IJSER © 2012
The research paper published by IJSER journal is about Interference Reduction and DOA Estimation in GSM Systems 3
ISSN 2229-5518
analysis shown, it is certain that finding the direction of
arrival of signals (including the desired and interfering signals) will help a great deal in the problem of interference suppression in gsm systems. Estimates with relatively high accuracy can be obtained with signals closely spaced in angles- co channel interference. This means that interferences in close vicinity of the desired signal can be mitigated
[1] Frank Gross, “Smart Antennas for Wireless Communications” with
Matlab by McGraw-Hill Companies Inc. 2005
[2] Nadeem Ather, “Null Steering and Pattern Control in Smart
Antenna Arrays”.
[3] Nikhil Shetty, “Direction of Arrival Estimation Algorithms” April 7
2004
[4] Nsionu I I ,” Improving GSM Network Capacity Using Smart
Antenna System” 2012
Fig. 2: Plot for estimated directions of arrival
[5] Ahmed El Zooghby , “Smart Antenna Engineering”, 2005 Artech
House Inc. 685 Canton street Norwood, MA USA
The estimation of the signal direction was carried using the
MUSIC algorithm was carried out in this paper. From the
14o , 28o ,35o ,55o along the peak of music spectra for 32-element array
Nsionu I I – Federal Polytechnic, Oko, Nigeria
Onoh G N- Enugu State University of Science & Technology Enugu, Nigeria Ekeh J- Enugu State University of Science & Technology Enugu, Nigeria Eneh I I- Enugu State University of Science & Technology Enugu, Nigeria
IJSER © 2012