Energy efficient architectures for the log-MAP decoder through intelligent memory usage

Abstract

Turbo decoding generally employs maximum a posteriori (MAP) and the soft output Viterbi (SOVA) algorithm in its soft-input soft-output (SISO) component decoders. This paper reformulates the implementation of a low power Log-MAP decoder with reduced storage requirement and based on the optimized MAP algorithm that calculates the reverse state metrics in the forward recursive manner. As a result, the authors present new low power derivatives of this decoder through a variation in the percentage of memory savings. Three low power architectures of the Log-MAP decoder not employing the sliding window technique have been developed and post layout power savings of approximately 44%, 40% and 36% with respect to the conventional implementation have been observed.