Adaptive beamforming using worst-case performance optimization with white-noise gain constraint

The performance of a robust adaptive beamformer using worst-case performance optimization (WCPO) depends heavily on the selection of the covariance matrix and steering vector uncertainty set parameters, and deteriorates significantly under low signal-to-noise ratio or large error conditions, which is equivalent to the performance degradation caused by the array receiving a certain amount of uncorrelated white noise. To address this issue, the original WCPO of using uniform values for covariance matrix uncertainty set parameters across all scanning angles is modified. By introducing a white noise gain constraint, the covariance matrix uncertainty set parameters are optimized independently for different angles, leading to a new extension of the WCPO robust adaptive beamformer. Furthermore, a reasonable strategy for selecting the parameters of the white noise gain constraint is provided. The results of simulated and experimental data both show that the new approach has a good performance in detecting the weak signal against strong interferences, provides accurate target power estimation, and shows strong robustness to steering vector errors. In addition, the algorithm performance is less sensitive to the parameters of the white noise gain constraint determined by the proposed strategy.