A modified direction-of-arrival estimation algorithm for acoustic vector sensors based on Unitary Root-MUSIC

Abstract

A novel method applying for direction-of-arrival(DOA) using acoustic vector sensors(AVS) based on Unitary Root-MUSIC algorithm(URM) is proposed in this paper. AVS array has a characteristic named coherence principle of sound pressure and velocity, which can significantly improve the detection performance of DOA by reducing the influence of white Gaussian noise. We apply this characteristic and the extra velocity information of AVS to construct a modified covariance matrix. In particular, the modified covariance matrix need not extend the dimension in calculation of AVS covariance matrix which means saving the computing time. In addition, we combine the characteristics of modified matrix with URM algorithm to design a new algorithm, which can minimize the impact of environment noise and further reduce computational complexity to a lower order of magnitude. So the proposed method can not only improve the accuracy of DOA detection but also reduce the computational complexity, compared to the classic DOA algorithm. Theory analysis and simulation experiment show that the proposed algorithm for AVS based on URM can significantly improve the DOA resolution in low SNR ratios and few snapshots.