On the complete decoding of constrained codes

AbstructA maximum likelihood algorithm for devising complete decoders for constrained codes and an algorithm to calculate the resulting error multiplication probabilities are presented. These algorithms enable us to reduce the error multiplication when decoding the symbols of constrained codes or the symbols of various combined error correcting, constrained coding schemes after reception over channels that introduce bit errors into the symbols. one mapping, fa : X -9 U, given fi : U + C. We here present algorithms to devise fa : X --f U such that we either a) inhibit the maximum possible error multiplication (i.e., the phenomenon that a single or a few channel errors may cause multiple decoding errors) for each decoded n-tuple Q; when channel errors occur, or b) minimize the average bit error rate in the decoded sequence (62).