Abstract
Intersymbol interference (ISI) is a major cause of performance degradation for both wireless and wireline communication systems. It can be mitigated by several different methods including equalization and multicarrier modulation, but the complexity and efficiency of all methods would depend on the length of the ISI channel. In this paper, we propose a general framework for channel shortening where we considered the channel as a rational transfer function (having infinite impulse response (IIR)), and the source and the noise as auto-regressive moving average (ARMA) processes. The aim is to shorten the channel using an IIR filter to a desirable length, so that computationally efficient post processing techniques can be applied to the resulting signal. The use of an IIR filter provides more degrees of freedom for channel shortening as compared to an FIR filter. © 2004 IEEE.