Turbo-codes decoding with weighting of extrinsic information for punctured turbo-codes

摘要

Turbo-codes are known as the powerful error correcting codes. The principle of turbo codes, which are concatenated convolutional codes, makes it possible that the coding rate is controlled by puncturing. In the case that punctured coding are adopted, the reliability of the decoded bits are not uniform over a frame, because the some bits are decoded only from information hits, and the other hits are decoded from both information bits and redundant bits. In the turbo decoding, the capability of error correcting is improved by the iterative decoding with updating the extrinsic information. The extrinsic value is outputted by each elementary decoder and represents the reliability of decoded bits. The difference of the reliability concerning the extrinsic information has not been considered yet. In this paper, we propose a novel turbo decoding algorithm, which compensates the difference of the reliability by weighting the extrinsic information. By the computer simulation, we show that the proposed method can achieve lower BER than the convolutional punctured turbo decoding.