摘要
针对基于校验方程符合度方法前面交织关系识别会严重影响其后所有交织关系识别的不足,本文对该算法进行了改进.改进算法根据原算法性能受限的原因进行方法设计,利用识别可靠度找到交织识别错误位置,并从降低复杂度角度出发,根据编码器的记忆特性,仅针对部分相关数据采用Gibbs样本法得到条件概率进行纠错.仿真结果表明,改进算法较原算法性能改善明显,在相同信噪比条件下,达到相同识别性能时所需数据量减少一半;在相同数据量条件下,达到相同正确率时的信噪比增益约为1dB.
The method based on conformity of parity-check equation was improved for its defect that the front interleaver identification has a strong impact on all the interleaver behind. The improved algorithm designs an approach on account of the limited performance of the original algorithm. It uses the recognition reliability to find the error position of interleaver identification. And from the purpose of reducing the complexity, it uses Gibbs sampling method to get the conditional probability of some relevant information for error correction according to the memory feature of the encoder. Simulation results show that, the improved algorithm performs better than the original algorithm, and it only needs half of the data to achieve the same recognition performance under the condition of the same SNR and has 1 dB SNR gain to achieve the same accuracy under the condition of the same amount of data.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2018年第1期15-23,共9页
Acta Electronica Sinica
基金
国家自然科学基金(No.61401511)
关键词
TURBO码交织器
校验方程符合度
识别可靠度
Gibbs样本法
Turbo-code interleaver
the conformity of parity-check equation
the recognition reliability
Gibbs sampiing method
