Research And Implementation Of Robust Adaptive Beamforming Algorithm

Adaptive beamforming is a main research direction of array signal processing and it has been widely used in practice. The desired signal steering vector and array signal covariance matrix unavoidably have mismatch under the complex electromagnetic environment. The performance of digital beamforming degrades sharply when the mismatch exists. So it is of great importance to study the robust adaptive beamforming in the presence of various mismatches in array.The paper addresses the robustness of the adaptive beamforming algorithm research. First of all, on the basis of the adaptive array signal processing, three traditional robust adaptive beamforming algorithms are analyzed:diagonal loading algorithm, eigenspace-based algorithm and the worst-case performance optimization algorithm. These algorithms are lack of some strict theoretical basis for specific mismatch. Then algorithm based on iterative methods to achieve desired signal steering vector correction and the algorithm based on steering vector estimation directly are studied. These two algorithms take the essential reason for performance degradation that is steering vector mismatch as a breakthrough point to improve the robustness of the algorithm, especially the direct estimation algorithm only need the information about desired signal’s incidence area. Aiming at the above two algorithms with low computational efficiency defects, a kind of eigenspace-based robust algorithm which has quick convergence rate is proposed. The proposed algorithm not only estimate the desired signal steering vector, but also reconstruct the desired signal and interference covariance matrix. What’s more, it compensates the covariance matrix to improve the algorithm’s performance under the low SNR condition. Finally the adaptive beamforming system based on high-speed ADC+FPGA+DSP hardware architecture is designed. This paper expounds the overall design requirements and implementation scheme of the system. Then this paper analyzes the signal processing process, adopts Jacobi method as the solution to compute the eigenvalue and eigenvector of eigenspace-based algorithm realized in DSP and verifies the feasibility of the algorithm which is realized in DSP.