DOA estimation for acoustic vector sensor array based on fractional order cumulants sparse representation

Abstract

Aiming at the problem that the existing direction of arrival (DOA) estimation algorithms are difficult to achieve high-precision estimation in environments with mixed Alpha-stable distribution noise and Gaussian-colored noise, a look ahead orthogonal matching pursuit algorithm based on Fractional Order Cumulants (FOC) is proposed for acoustic vector sensor (AVS) arrays. Firstly, the algorithm computes the FOC matrix of the observed data and exploits the semi-invariance of the FOC to separate Alpha-stable distribution noise and Gaussian-colored noise from the observed data. Furthermore, the property that FOC is insensitive to the Alpha-stable distribution processes and Gaussian processes is then exploited to suppress the Alpha-stable distribution noise and Gaussian-colored noise. Subsequently, the FOC matrix is reconstructed through the vectorization operator, and an FOC-based sparse DOA estimation model is derived. Finally, the look ahead orthogonal matching pursuit algorithm predicts the impact of each candidate atom on minimizing the residual. It selects the optimal atom to enter the support set, obtaining the DOA estimation of the target. The effectiveness of the proposed algorithm is verified through computer simulations. The simulation results show that the proposed algorithm has high estimation accuracy and success probability.