Bounded metric Viterbi decoding of trellis coded modulation in presence of intersymbol interference

A new approach to a reduced states Viterbi decoder (VD) is introduced which is based on a bounded metric. This approach is most attractive for combatting intersymbol interference (ISI) using a Viterbi algorithm, specifically if trellis coded modulation (TCM) is used. It has been shown that in the case of ISI the overall compound Viterbi decoding of ISI and TCM is much better than any other decoding algorithm, for example decision feedback. The most aggravating drawback of such an overall VD is its large number of states, resulting from both the TCM encoder and ISI memory. However, this large number of states can be substantially reduced by the authors´ method of a bounded metric, with a negligible performance degradation