Updated extrinsic information for iterative decoding of turbo codes

This paper proposes a novel iterative method for turbo codes. This method updates extrinsic information value based on the interim soft outputs to minimize the effects of decoding errors of elementary decoders. The bit error rate performances of turbo codes are simulated, where MAP, log-MAP and MAX-log-MAP based component decoders are employed. Moreover, performances of turbo decoder in the cases that extrinsic information value revised and unrevised are compared in an AWGN channel. It is concluded that the proposed method minimizes the extrinsic information errors, thus the turbo decoder based on the MAX-log-MAP algorithm performs close to that based on MAP and log-MAP algorithms respectively.