Prototyping a Digital Communication System with a Novel Blind Adaptive Equalizer

In this paper, an undergraduate research project, sponsored by Arkansas NASA Space Grant, is presented. A novel blind adaptive channel equalizer, which is based on constant modulus algorithm (CMA) with variable step size, is first proposed. The step size is designed to follow an exponential function, with a large initial value for fast convergence rate and a small steadystate value for a low convergence level. A comprehensive digital communication system is further implemented, which includes a digital modulator for the source signal, a multipath channel, a noise generator and a blind equalizer. The prototype builds around software defined radio concepts on both a PC and a DSP platform. It provides flexibility for comparing different equalization algorithms in various realistic environments. Both off-line simulation in Matlab and real-time experiment on DSP board are conducted and presented. It is observed from the simulation results that the proposed blind equalizer performs significantly better than the conventional fixed-step CMA approach.