A new decorrelation algorithm for a soft decision feedback blind equalizer

Abstract

This paper proposes a new decorrelation algorithm for a soft decision feedback blind equalizer (soft DFE) based on a DFE switching structure model. The algorithm is developed using joint entropy maximization (JEM) framework suitable for introducing a desirable nonlinearity directly into the stochastic gradient algorithm. Simulation results have shown that an efficient decorrelation based on JEM algorithm provides a better convergence rate and switching robustness than the existing hard DFE solution.