DOA Tracking Algorithm for the Time-varying Number of Signal Sources

Abstract

At present, sparse Bayesian learning (SBL) is introduced into direction of arrival (DOA) estimation for both coherent and incoherent signals. Instead of directly extending DOA estimation to DOA tracking, we construct an array data model including DOA change of adjacent time to decrease computational complexity and avoid grid effect. Thus, we regard DOA tracking as a parameter estimation problem in terms of Taylor expansion and Bayesian rule. Owing to the existence of hidden variables, we adopt the Expectation Maximization (EM) algorithm to calculate the DOA change. More importantly, we realize DOA tracking by utilizing the estimated signal power and noise power when the number of signal sources varies with time. The proposed method is numerically evaluated with an assumption of uniform linear array. The results show that the proposed algorithm has higher tracking accuracy over conventional methods.