Direction of Arrival Estimation Method for Circular Acoustic Vector Sensor Array Without Angular Ambiguity in Ambient Noise

This study presents a DOA estimation method without angular ambiguity using a transformed covariance matrix differencing approach, with the aim of mitigating the impact of the correlation between the noise of each channel of the circular acoustic vector sensor array (CAVSA) with the adjacent element spacing less than the half wavelength of the signal on the direction of arrival (DOA) estimation of the subspace-based methods in ambient noise. This method first transforms the noise covariance matrix into a real-valued matrix, and the CAVSA covariance matrix, which contains the transformed noise covariance matrix, is differential with its conjugate matrix, to remove the noise component. To avoid the angular ambiguity problem, using the eigenvectors that correspond to the bigger positive eigenvalues of the differenced covariance matrix to estimate the DOAs of the source, the matrix orthogonal to the signal steering vector is constructed. The simulation results reveal that at low signal-to- noise ratios, the proposed method achieves higher resolution and obtains better DOA estimation performance. Furthermore, through a lake experiment, the effectiveness of the proposed method is verified.