Abstract:

A robust diagonal loading (DL) beamforming method is proposed, which can effectively over- come the influence of DOA(direction of arrival) mismatch and sensor position uncertainties on the per- formance of adaptive beamformer. The proposed method exhibits the enhanced robustness of the beam- formers with the traditional method of DL based on analysis of linear constrained minimum variance beam- former. Then the support vector machine (SVM) algorithm is applied to the robust beamforming, which is based on the principle of structural risk minimization. The simulation results show that the SVM-based beamforming method enhances the robustness in terms of desired signal array manifold vector errors in an ideal scenario of no-mismatch and an actual scenario of mismatch, respectively.

Key words: information processing, beamforming, diagonal loading, support vector machine, robustness