Practical Implementation of a Factorized All Pass Filtering Technique for Non-minimum Phase Models

Abstract

One of the problems of the inverse model-based control techniques is the stability of the identified inverse model after the estimation process by the recursive least square (RLS) method. One solution is to use the all pass filtering (APF) technique to transform non-minimum phase models into minimum phase models [9]. However, there are several cases not cured by the all pass filtering method when the all pass filter algorithm is implemented on the hardware. In this paper, an improved version of the all pass filtering technique is presented to deal with the non-minimum phase models. The factorized APF (fAPF) technique for non-minimum phase models is presented to suggest a simple solution. A simple method for avoiding the calculation of complex numbers is also presented for the easy implementation. Several examples are given to support the proposal.