Application of the alternating projection strategy to the Capon beamforming and the MUSIC algorithm for azimuth/elevation AOA estimation

Abstract

In this paper, the alternating projection (AP) scheme is applied to get the initial and the final estimates of the Capon beamforming algorithm and the MUSIC algorithm. In a previous paper, the AP scheme has been proposed as a computationally efficient method for the initialization and optimization of the cost function for the ML-based angle of arrival (AOA) estimation problem. We call the AP scheme in the previous paper the brute-force AP strategy. To improve the brute-force AP, we suggest the Newton AP which consists of the successive one-dimensional Newton search. The brute-force AP can be used for the initialization and the optimization of the cost function of AOA estimation, and the Newton AP can be used for the optimization of the cost function for the AOA problem. We compare the performance of the AP-based approach with the conventional implementation of the AOA estimation algorithm using two-dimensional brute-force search.