Adaptive Decision Feedback Equalizer for SUI channel models

This paper investigates the performance of the Non-Stationary Recursive Least Squares (NSRLS) algorithm in the adaptive equalization of realistic wireless channels. As in general the wireless channels are assumed to vary in time according to a Markov model, the NSRLS algorithm may represent a favorite candidate since it is designed to track Markovian time varying channels. The Stanford University Interim (SUI) channels are considered in this paper. To obey the constraints of the realistic transmission context, we propose in this paper a generalized version of the NSRLS algorithm. The performances of the Decision Feedback Equalizer (DFE) updated by the proposed NSRLS algorithm are compared with those of the conventional RLS-DFE through simulations. The reported results demonstrate the efficiency of the generalized NSRLS algorithm to capture the time variations of the SUI-1 and SUI-2 channels. Indeed, the Bit Error Rate (BER) is significantly reduced with the NSRLS-DFE. Moreover, it is shown that a high order Markov model is required to well represent the non-stationarity of the SUI channels.