A fast recursive least-squares with a weighting factor based an a 2-D NSHP AR model and its application to spectral estimation

Abstract

This paper proposes a fast RLS based on a two-dimensional AR model in order to estimate a power spectrum of signals about time and space. The proposed method introduces a non-symmetric half plane (NSHP) model and an exponential weighting factor about time. Moreover, the calculation cost of the conventional FRLS is reduced.