Adaptive phase equalization using all-pass filters

Abstract

The feasibility of the application of all-pass filters (APFs) to channel equalization is studied. It is shown that the amplitude distortion of communication channels, which is known to be a function of the receiver timing phase, can be minimized by application of the available near-optimum timing recovery schemes. As a result, the contribution of the amplitude distortion to the resulting intersymbol interference (ISI) may be negligible as compared with that of the phase distortion. Under these conditions, the adaptive APFs may be effectively used for channel equalization. The general structure of two-dimensional adaptive phase equalizers is given and their corresponding LMS algorithm is also developed. Computer simulations confirm the unimodality of the performance surfaces of the phase equalizers in the region of convergence.