Novel decoding algorithm with weighting of extrinsic information for punctured turbo-codes

Turbo-code is known as the powerful error correcting code. The principle of turbo-code, which is the concatenated convolutional code, makes it possible to control the code rate by the use of the puncturing. When the puncturing is applied, the reliability of the decoded bits is not uniform over a frame, because some bits are decoded using only received information bits, and the others are decoded using both information and redundant bits. In the turbo-code decoding, the capability of error correcting is improved by the iterative decoding with updated extrinsic values. The extrinsic values are outputted by each elementary decoder and represent the reliability of decoded bits. In this paper we propose a novel turbo-code decoding algorithm, which compensates the di erence of the reliability by weighting the extrinsic values. By the computer simulation, we show that the proposed method achieves lower BER than the conventional decoding algorithms. key words turbo-codes, weighting, punctured codes, extrinsic information