Further results on the equivalence between SOVA and max-log-MAP decodings

Turbo codes are preferred for high rate data services in the third generation mobile communication system IMT-2000. Thus low decoding delay and low complexity interactive decoding is of paramount importance. The modified soft output Viterbi algorithm (SOVA) and the max-log-maximum a posterioi (MAP) algorithm are such decoding techniques. The equivalence of modified SOVA and max-log-MAP decodings has been proven on the assumption of no a priori information and infinite decoding depth. Since the a priori information has to be updated continuously in iterative decoding, we prove their equivalence when there exists a priori information. We further demonstrate the equivalence between the sliding window (SW) max-log-MAP algorithm and the modified SOVA for finite decoding depth