New subspace updating algorithm for adaptive direction estimation and tracking and its statistical analysis

Subspace estimation is of importance to high-resolution direction estimation in array processing. In this paper, a new recursive least-squares (RLS) algorithm is proposed for null space estimation, which is used to estimate or track the directions of coherent and/or incoherent signals impinging on a uniform linear array (ULA). Especially by investigating the expectation computation of an inverse matrix, the statistical analysis of the RLS algorithm is studied in the mean and mean-squares senses in stationary environment, and further the mean-square-error (MSE) and mean-square derivation (MSD) learning curves are derived explicitly. The theoretical analyses and effectiveness of the proposed RLS algorithm are substantiated through numerical examples.