Performance evaluation of the error-forecasting decoding on Markov channels

摘要

Interleaving random-error-correcting block codes is known as a countermeasure against burst errors. In general, each of interleaved block codes is independently decoded at the receiver. However, it has been shown that the error-forecasting decoding can improve decoding performance by taking into account correlation of burst errors. In this report, we derive expressions of the probability of incorrect decoding for the error-forecasting decoding on Markov channels by constructing some Markov chains derived from the original Markov channel model. The derived expressions are examined for three-state and four-state Fritchaman channel models for interleaved BCH codes. The optimal degree of interleaving which minimizes the probability of incorrect decoding is also numerically evaluated.