Comparison between Two RLS Methods for Adaptive Sinusoids Retrieval: Notch Filtering and Subspace Tracking

摘要

This paper presents a brief comparison between two modern methods in their RLS form for retrieving sinusoidal signals from background noise: notch filtering and subspace tracking. We consider the stabilized PAST (Projection Approximation Subspace Tracking) and PASTd (PAST based on deflation technique) methods and we compare their performances with the RLS notch algorithms. We show extended simulations results in order to compare all the algorithms' performances. We study the statistical properties of both the notch-type and the subspace tracking-type algorithms in stationary environments and for frequencies situated relatively far in the normalized spectrum.