A NEW ALGORITHM FOR ADAPTIVE IIR FILTERS

Abstract

This paper introduces a Simplified Hyperstable SEquential Regression (SHSER) adaptive algorithm designed for use with Infinite Impulse Response (IIR) digital filters, which offers a reduced computational load since it avoids direct matrix inversions, fulfills the hyperstability condition under certain circumstances, and has a high convergence rate. The proposed SHSER algorithm is a combination of the Simplified Hyperstable Adaptive Recursive Filters (SHARF) algorithm and the recursive version of the SEquential Regression (SER) algorithm for IIR adaptive filters. Some important comments associated with the proposed algorithm including the SPR condition, and the step size are briefly given. Simulation results are included comparing the proposed SHSER algorithm to the SHARF algorithm and SER algorithm with respect to the convergence rate behaviour.