【摘要】 在差分协作系统中,为了解决硬判决多符号差分检测(multi-symbol differential detection,MSDD)译码过程中,部分信息丢失导致性能损失以及该算法中最大似然检测计算复杂度高的问题,提出了一种软输入软输出的多符号差分球形译码(SISO-MSDSD)算法.该算法利用球形译码思想,在一定搜索范围内计算编码比特的对数似然比(LLR),去除先验信息可获得编码比特的外信息,经内部译码器判决得到源信息.仿真分析表明,比较硬判决MAP算法,SISO-MSDSD算法性能明显提高,且在信道频偏较大时,随着分组长度增加能有效提高检测性能.此外,在低SNR区域,SISO-MSDSD算法对复杂度的改善也十分明显.因此,SISO-MS-DSD可作为一种较好的软检测算法应用于差分协作系统中.更多还原

【Abstract】 In order to solve the deficiency of performance loss caused by some information missed during the decoding process and reduce high computational complexity of maximum likelihood detection in hard decision multiple-symbol differential detection,we proposed a soft-input soft-output multi-symbol differential sphere decoding(SISO-MSDSD) algorithm in the differential cooperation system.The algorithm calculated the log-likelihood ratio of encoding bit in a certain search range by using the idea of sphere decoding.The extrinsic information of encoding bit could be achieved by removing a priori information.After the judgment,the internal decoder restored the encoding bit information.Simulation analysis shows that SISO-MSDSD has better performance than a hard decision MAP algorithm.The performance has an effective improvement with the increase of group length when channel frequency offset is large.Furthermore,SISO-MSDSD can obviously reduce computational complexity in the low SNR region.Therefore,SISO-MSDSD can be used in the differential cooperation system as an excellent soft decision algorithm.更多还原