Performance Comparison of Nature-inspired Optimization Algorithms Applied to MVDR Technique for Canceling Multiple Access Interference Signals
Background/Objectives: Minimum Variance Distortionless Response (MVDR) beam forming technique is among the most widely used in antenna array field. The conventional MVDR has poor performance, and low Signal to Interference plus Noise Ratio (SINR) gain in the condition of limited snapshots or Multipl...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Informatics Publishing Limited
2018
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/20609/ http://umpir.ump.edu.my/id/eprint/20609/ http://umpir.ump.edu.my/id/eprint/20609/ http://umpir.ump.edu.my/id/eprint/20609/1/Performance%20Comparison%20of%20Nature-inspired%20Optimization%20Algorithms%20Applied%20to%20MVDR%20Technique%20for%20Canceling%20Multiple%20Access%20Interference%20Signals.pdf |
| Summary: | Background/Objectives: Minimum Variance Distortionless Response (MVDR) beam forming technique is among the most widely used in antenna array field. The conventional MVDR has poor performance, and low Signal to Interference plus Noise Ratio (SINR) gain in the condition of limited snapshots or Multiple Access Interference (MAI) signals existing. Heuristic optimization algorithms are broadly used to solve many engineering problems. Methods/Analysis: In this work, two nature-inspired optimization methods, namely Particle Swarm Optimization (PSO) and Gravitational Search Algorithm (GSA) are applied to enhancing the conventional MVDR performance. In particular, the complex weight coefficients of the conventional MVDR solution are improved using both approaches. First, SINR calculated from MVDR basing linear antenna array configuration then the PSO and GSA implemented to minimizes the power of noise and interference in the constraint condition. The performance of the proposed methods is assessed based on various QoS criteria such as beampattern accuracy for azimuth and elevation scanning angles and SINR output. Findings: In comparison to conventional MVDR, the proposed algorithms have indicated that MVDRGSA providesfavorable agreement of synthesizing a maximum gain toward the desired real user angle while introducing deep null-forming in the undesired user directions. As a result, average SINR is evaluated over 20 runs in all simulation scenarios, the performance of MVDRGSA is better than the performance of MVDRPSO. Moreover, a good control over the null-forming level can be achieved by MVDRGSA for iteration number < 100 whereas MVDRPSO is simple and easy to implement but required more convergence time to get high SINR. Application/Improvements: In general, it was observed that MVDRGSA out performs the MVDRPSO with respect to solution quality, stability and convergence speed. |
|---|