Adaptive Bayesian decision feedback equaliser based on a radial basis function network

Abstract
The authors derive a novel Bayesian decision feedback equalizer (DFE) for digital communications channel equalization. It is shown how decision feedback is utilized to improve equalizer performance as well as to reduce computational complexity. The relationship between the Bayesian solution and the radial basis function (RBF) network is emphasized and two adaptive schemes are described for implementing the Bayesian DFE using the RBF network. The maximum likelihood sequence estimator (MLSE) and the conventional DFE are used as two benchmarks to assess the performance of the Bayesian DFE.