Eigenstructure-based DOA estimation for circular acoustic vector sensor array with element attitude errors

To enhance the direction of arrival (DOA) estimation performance of a circular acoustic vector array (CAVA) with element attitude errors, this paper proposes eigenstructure-based DOA estimation methods utilizing the eigendecomposition of the amplitude-weighted covariance matrix. In the scenario of isotropic ambient noise, the amplitude weighting processing is applied to the CAVA to reduce the impact of the inequality between the noise powers of the pressure and velocity channels. An objective function is constructed based on the orthogonality between the actual steering vector and the noise subspace. Using this objective function, the optimization problem concerning attitude errors is established for a preset azimuth. The closed-form solution of the vector containing the attitude error parameter is then solved to minimize the objective function, thereby enabling the estimation of the source DOAs. Furthermore, the accuracy of DOA estimation can be further improved through joint iterative processing. The simulation results demonstrate that the proposed methods exhibit strong robustness in the presence of element attitude errors, and obtain better DOA estimation performance and higher azimuth resolution. The experimental results further verify the effectiveness of the proposed methods.