一种提高SCCC系统迭代检测收敛性的方法

A method to improve iterative decoding convergence in SCCC

运用Turbo原理进行迭代检测是提高串行级联卷积码(SCCC)系统性能的一种全新思路,迭代检测的收敛性是决定SCCC系统性能的关键因素之一.针对短帧系统在迭代检测过程中存在的正反馈现象,提出了一种加权外信息交换的迭代方法,该方法通过对内译码器与外译码器之间传递的外信息概率进行加权控制来减少正反馈现象.理论分析与仿真结果表明,所提出的改进方法不仅能够有效地抑制SCCC系统中存在的正反馈现象,提高了系统的误比特性能,而且减少了平均迭代次数,提高了系统的实时性;同时避免了外信息在传递过程中从概率到似然比及其逆转换的过程,降低了译码算法的复杂度.

Iterative decoding based on the turbo principle is a novel approach for improving the performance of serial concatenated convolutional code （SCCC） systems. The speed of convergence of iterative detection is one of the key factors determining system performance. To reduce the high levels of feedback required during iterative decoding of short data blocks, a solution that improves the exchange of extrinsic information was proposed. In this method, inter decoder feedback is reduced by weighting the probability of extrinsic information between the inner and outer decoders. Theoretical analysis and simulation results show that the proposed method decreases the need for feedback and improves bit error rate performance. As a result, the average number of iterations was reduced and real time performance improved. In addition, the proposed method does not require conversion from probability to likelihood ratios, nor its inversion in the process of transforming extrinsic information. This significantly decreases the complexity of the decoding algorithm.