Adaptive equalization with RLS-MLSE for fast fading mobile radio channels

Abstract

The authors derive a novel adaptive equalization scheme from the maximum likelihood estimation theory. This scheme estimates the transmitted signal sequence from the received signal sequence which may suffer from multiplicative time-varying fading and additive white Gaussian noise. Time-varying fading is characterized as a simple Markov Gaussian process. The new scheme is a combination of the recursive least squares adaptation algorithm and maximum likelihood sequence estimation (RLS-MLSE) technique. The authors also report on the performance of the proposed RLS-MLSE as evaluated through laboratory experiments. An RLS-MLSE prototype is developed with a single-chip digital signal processor (DSP) to assess the feasibility of RLS-MLSE's application to 42 kb/s 3-channel TDMA (time division multiple access) mobile radio QPSK transmission in the 900 MHz band under frequency-selective fast fading environments. Experimental results show that RLS-MLSE is superior to a recursive least squares decision feedback equalizer.