A Low-Complexity Robust Bearing Estimator Using Quadric Rotational Invariance of Covariance Matrix for the Distributed Source

摘要

In recent years, many methods were proposed to estimate the bearing of spatially distributed source. They usually suffer from heavy computational load and limit to small angle spread due to model-approximated error. The main contribution of this paper is twofold. First, the unimodal symmetric space frequency distribution is introduced to describe the source model. The exact expression of covariance matrix can be calculated without approximating processing even in the case of large angle spread. Second, using Toeplitz and quadric rotational invariance of covariance matrix, a novel estimator is proposed by solving a nonlinear least-squared problem in central space frequency. It only requires a FFT or one-dimension search to obtain the bearing estimate. Numerical results illustrate its asymptotic efficiency and robustness in the large spread case.