An Adaptive Method of Signal Separation Based on Spatial Filter

Abstract

This paper addresses the problem of parameter estimation and adaptive separation by antenna arrays. The technique of Sparse Bayesian Learning (SBL), with remarkable performance in low SNR and limited snapshots, is introduced to estimate Direction-of-Arrival (DOA), as no information about the statistical property or deterministic property is known in advance. The spatial filter is designed based on the DOA estimates to separate signals from different directions. It is shown that the spatial filter can separate the signals with the noise power decreased. To enhance the performance of separation, an iteration processing is utilized until satisfying the convergence criterion. Experimental results are used to evaluate the performance of the spatial filter.