Performance of a Variable-Constraint-Length Viterbi Decoding Algorithm

摘要

Document demonstrates that the computational complexity of the Viterbi decoding algorithm can be reduced for Maximum Likelihood Sequence Estimation on time variant, fading channels when the channel constraint length is reduced. An algorithm is developed that reduces the computational complexity of the Viterbi decoding algorithm. For time-variant, fading channels, the effective constraint length may expand or contract, and it is shown that we can expand or contract the state trellis diagram of the Viterbi algorithm accordingly. The derived algorithm requires knowledge of the channel's intersymbol interference patterns. These results apply not only to the hf channel, but to any channel characterized by convolutional encoding. In addition, channels that demonstrate catastrophic behavior are found and discussed.