Beam Pattern Optimization Using MVDR and Simulated Annealing

Yu Yanli,Wang Yingmin

DOI: https://doi.org/10.1109/SOSE.2010.23

2010-01-01

Abstract:A novel numerical algorithm about pattern synthesis for sensor array with arbitrary geometric configuration is presented. Unlike conventional approaches, the method proposed is based on minimum variance distortionless response (MVDR) and Simulated Annealing. The method is aimed to minimize the side lobe peak of the beam pattern and compress the beam width. The method designs initial beam pattern using MVDR, take the weighting vector obtained as the initial weighting vector and then optimize the pattern further using Simulated Annealing. The method is flexible for optimal array pattern synthesis, and globally optimal solution can be determined efficiently due to Simulated Annealing optimization formulations. Array calibration based on the measurement array manifold is used to overcome the array model error. Compared to the conventional approach for array pattern synthesis, beam pattern designed by proposed method has better performance. Example about 24-element cross array is presented to illustrate the method. Obtained experimental results confirm the effectiveness of the approach in comparison with other techniques in the related literature.