2D DOA Estimation for Coherent Signals with Acoustic Vector-Sensor Array.

In this paper, a fast ESPRIT algorithm is presented for two-dimensional (2D) direction-of-arrival (DOA) estimation of coherent signals with a linear acoustic vector-sensor array. Conventional space smoothing ESPRIT (SS-ESPRIT) algorithm for acoustic vector-sensor array has two obvious weak points: (1) eigenvalue decomposition (EVD) of high-order covariance matrix; (2) frequently occurring parameters mismatch which leads to failed 2D DOA estimation. The proposed algorithm can remove the two shortcomings. Firstly, we use two selection matrices to separate received data and form a low-order cross-covariance matrix. Then, ESPRIT algorithm is used to estimate the elevation angles of incident signals. At last, using the estimated elevation angles, a modified array manifold matching method is used to estimate the azimuth angles of coherent signals. During the estimation process of azimuth angles, the proposed method is not required to perform any peak search or EVD. Moreover, azimuth angles are automatically matched with the estimated elevation angles. Compared with the conventional SS-ESPRIT algorithm, the proposed algorithm has lower complexity. However, the estimation accuracy of proposed method almost is the same as the conventional SS-ESPRIT algorithm and far better than PM algorithm. Numerical simulations are presented to demonstrate the performance of the proposed algorithm. [ABSTRACT FROM AUTHOR]