Low complexity decoding for intersymbol interference channels

The author first discusses the performance of the M-algorithm for intersymbol interference (ISI) channels. As a specific example, he considers the magnetic recording channel for four different densities. If decoding error probability is the performance criterium, then he concludes that for this type of channel only a few estimates need to be considered by the decoder. He gives simulation results for a modified maximum likelihood sequence estimation (MLSE) decoding algorithm that uses a subset of the states as defined by the channel impulse response. Simulations results indicate that this type of decoding can be very efficient. He also shows that interleaving, combined with a minimum amount of redundancy, gives rise to surprisingly good decoding performance for a direct detection with soft decisions