Blind equalization of encoded sequences by hybrid Markov models

Abstract

An algorithm for blind equalization of encoded sequences is presented. The main feature is the fact that the statistics of the encoder are taken into account, and the equalizer is a simple extension of the well known Viterbi decoder. A hybrid Markov model with state having both discrete and continuous components, is at the basis of the algorithm. It can easily lend itself to a parallel implementation using two DSP chips.