Simulated Annealing-MVDR Beamformer Based on Maximum Likelihood DOA Estimation

In the presence of direction of arrival (DOA) mismatch, lack of samples or high input signal to noise ratio (SNR), the minimum variance distortionless response (MVDR) beamformer forms a mainbeam away from the desired signal direction and high rising sidelobe levels. It results in the output signal to interference and noise ratio (SINR) decreases. In this paper, the DOA estimation is done by the simulated annealing (SA) algorithm on the maximum likelihood (MA) function. The sample covariance matrix is adjusted to tackle the problem of the MVDR beamformer when the small number of samples or the strong desired signal exists. The SA algorithm which is a powerful technique in locating the global optimum solution terminates the adjustment until the minimum mean output power of array is attained with the constraint of the unity response in the look direction. Simulation results indicate that the presented SA-MVDR beamformer enhances the beampatterns significantly in the case of the low number of samples and high input SNR.