Low Complexity Stopping Rule for Turbo Decoding: the Max-log Criterion

This paper presents a new stopping criterion for turbo decoding. It is based on the selection of the maximum log-alphas calculated by the log-MAP algorithm. The sum of these maximum alphas is compared with a threshold value. Then, a decision on the end of decoding is taken. Simulation results show that the max-log criterion offers the same performance as the sum-alpha and sum-log criteria, while maintaining the same complexity level. The max-log criterion uses only the max operator to select maximum alphas and a summation. Therefore, the proposed criterion is faster and offers lower complexity.