Pilot-Assisted Communication Over ISI Channels Based on the Belief Propagation Algorithm

Application of the belief propagation (BP) algorithm is proposed for pilot-assisted reception of Gaussian symbols over unknown inter-symbol interference channels whose parameters are also Gaussian. The proposed Bayesian network graph for the demodulator includes the channel parameters explicitly, allowing the combined estimation of the channel and information bearing symbols. The proposed algorithm is parallel in nature, is suitable for sparse channels, and is easily expandable to coded systems for joint channel pilot-assisted demodulation and decoding. The numerical results show performance better than the performance of iterative semi-blind equalization with batch minimum mean-squared error estimation iterations.