Low-complexity method of weighted subspace fitting for direction estimation

Abstract

In this paper, we consider a low-complexity method of weighted subspace fitting (WSF) for direction-of-arrival (DOA) estimation. With the properties of the multi-stage Wiener filter (MSWF), we derive a novel criterion function for the WSF method without the estimate of an array covariance matrix and its eigendecomposition. A new approach for noise variance estimation is also proposed. Numerical results indicate that by selecting a specific weighting matrix, the low-complexity WSF estimator can provide the comparable estimation performance with the conventional WSF method.