A robust adaptive beamforming method based on variable constraint of uncertainty set

The performance of adaptive beamformer is seriously degraded under mismatched conditions with covariance matrix error and steering vector error.The Worst-Case Performance Optimization(WCPO) approach has greatly improved robustness of adaptive beamformer under mismatched conditions.However,WCPO has a problem of systematic overestimation of signal power.Furthermore,two key parameters of WCPO which determine the uncertainty sets of covariance matrix and steering vector respectively,have to be specified by the users,and are lack of solving methods and data adaptiveness.In this paper,a theoretical analysis of power overestimation is provided as well as a corresponding power correction method.Then an adaptive method based on matrix reconstruction to estimate the uncertainty set of covariance matrix is given.Finally,a WCPO-based robust adaptive beamformer with a combination of power correction and variable constraint of uncertainty set(WCPO-PCVC) is proposed.By contrast with WCPO,WCPO-PCVC is a user parameter-free approach that can eliminate the systematic deviation of signal power estimation.Results of simulated and experimental data show that our approach has excellent interference and sidelobe rejection capability under mismatched conditions,better angular resolution than conventional beamformer and more accurate power estimation than adaptive beamformer.