Bidirectional Viterbi Decoding using the Levenshtein Distance Metric for Deletion Channels

Abstract

In this paper, we present a bidirectional Viterbi decoding algorithm using the Levenshtein distance metric for a regular convolutional encoding system. For a deletion channel, this decoding algorithm can correct an average of 30 deletion errors within a 6000 bit frame, when using an r = 0.67 regular convolutional code; and it can correct an average of 80 deletion errors within a 4000 bit frame, when using an r = 0.25 regular convolutional code