DOA Estimation of multiple sources based on multiset canonical correlation analysis

Abstract

It is an urgent task to improve the direction of arrival (DOA) estimation precision. To solve the problem of time difference DOA estimation is sensitive to interference and noise, this paper proposes a novel multiple sources DOA estimation algorithm based on multiset canonical correlation analysis (MCCA). The key technique of this DOA estimation method is blind separated the mixed signals from multiple moving sources by MCCA, translated the multiple sources DOA estimation problem to each source DOA estimation in turn, and removed the effect of interference and noise. The simulation results suggest that the estimation of time difference after MCCA processing is more accuracy and stabilization than direct correlation and after CCA processing, can greatly improve the performance of DOA estimation.