A novel approach to ML DOA estimation based on eigenfiltering and stochastic search

Abstract

The paper proposes a novel approach that a new likelihood function is derived from observation data after filtered with eigenfilters, and hybrid gravitational search algorithm (H-GSA) optimization is collaboratively applied to maximum likelihood (ML) estimation of the direction of arrival (DOA) parameters of multiple signals impinging on a sensor array. This method prevents the ML estimation performances from deteriorating severely where the angular separation between signal sources is small and the SNR / sample size are low. Simultaneously due to the use of H-GSA, we make direct maximization of likelihood realistic in practice. In order to examine the performances of the proposed method, four kinds of situations are designed. Simulation results indicate that the proposed method offers significant performance enhancement at low signal to noise ratios, and hybrid GSA stochastic search technique is therefore efficient and reliable.