Off-Grid DOA Estimation of Acoustic Vector Sensor Array Based on Look Ahead Strategy in Impulsive Noise

Abstract

To address the issue of low direction-of-arrival (DOA) estimation accuracy for acoustic vector sensor arrays under impulsive noise and grid mismatch conditions, a look ahead orthogonal matching pursuit algorithm based on phased fractional lower-order moments (PFLOM) is proposed. First, the algorithm uses PFLOM to suppress impulsive noise and reconstructs the PFLOM matrix using a vectorization operator. Next, it introduces off-grid deviation by performing a first-order Taylor expansion on the steering vector matrix, constructing a PFLOM-based off-grid sparse DOA model. Then, the look ahead strategy is introduced into the algorithm to select the optimal atoms by predicting their impact on the residuals. Finally, the joint sparsity of the coarse DOA estimation and the off-grid deviation vector is exploited to calculate the corresponding off-grid deviation using the alternating direction iteration method, resulting in the DOA estimation for the off-grid targets. Computer simulations validate the effectiveness of the proposed algorithm, with experimental results showing higher estimation accuracy and success rate compared with existing methods.