A New DOA Estimation Algorithm Based on PSO-Gauss-Newton

Abstract

Direction-of-Arrival (DOA) estimation is a basic and important problem in sensor array signal processing. In order to solve this problem, many algorithms have been proposed. Among them, the Stochastic Maximum Likelihood (SML) algorithm has become one of the most concerned algorithms because of its high DOA accuracy. However, the computational complexity of SML algorithm is very high, so Gauss-Newton algorithm is used as the analytical algorithm of SML in this paper. The traditional Gauss-Newton algorithm used in DOA estimation has some defects: (1) over reliance on the choice of initial values (2) fall into local optimum easily. In order to solve these defects and further reduce the computational complexity, this paper proposes a new DOA estimation algorithm based on PSO-Gauss-Newton. First of all, a limited solution space is proposed based on the precondition that the estimated signal must be non-negative definite. Then, according to the idea of PSO(Particle Swarm Optimization) algorithm, multiple scattering points are randomly distributed in the limited solution space. Each initial particle performs Gauss-Newton algorithm iteration separately. Finally, the global optimal solution is determined by comparison of all the convergence values. Simulation results the computational complexity of this algorithm is almost comparable to that of MUSIC algorithm.