SIMO系统吉布斯盲迭代均衡算法  

Iterative Blind Equalization Based on Gibbs Sampler for SIMO Communication Systems

作　　者：乔良[1] 郑辉[1] 

机构地区：[1]西南电子电信技术研究所盲信号处理重点实验室,四川成都610041

出　　处：《四川大学学报（工程科学版）》2015年第3期123-129,共7页Journal of Sichuan University (Engineering Science Edition)

基　　金：国家自然科学基金资助项目(61172140)

摘　　要：针对符号间干扰信道的多天线分集接收问题,提出一种单输入多输出(SIMO)系统盲迭代均衡算法。该算法利用吉布斯样本法处理思路,在SIMO条件下推导了信道冲击响应、发送符号等未知参数的条件后验分布,根据该条件概率逐个参数进行随机采样,通过不断迭代更新来逼近最大后验概率(MAP)估计的结果。该算法的一个显著特点是具有软输入软输出(SISO)结构,因此在编码系统中可以与信道译码结合,通过联合迭代进一步提升均衡的性能。计算机仿真结果表明,在严重符号间干扰信道条件下,SIMO系统盲迭代均衡算法的性能非常接近于已知信道时迭代均衡算法的性能,距离理想无符号间干扰信道分集合成的性能差距只有约1 d B。To solve the problem of symbol detection in the intersymbol interference( ISI) channel of spatial diversity systems,an iterative blind equalization algorithm was proposed for single-input multiple-ouput( SIMO) communication systems based on the Gibbs sampler method. The conditional posterior distributions of all unknown quantities such as channel impulse response,transmitted symbol sequence were derived in SIMO systems. The unknown quantities were updated one by one from such conditional distributions,so that the maximum a posteriori( MAP) estimates of these unknowns were accomplished in an iterative manner. A salient feature of the equalization algorithm was that it had a soft-input soft-output( SISO) structure. Hence,it was well suited for iterative processing in a coded communication system,which allowed the blind equalization to improve its performance. Simulation results showed that the iterative blind equalization algorithm performs closely to the algorithm with channel response perfectly known to the receiver in severe ISI channels. Performance gap between this approach and channels with no ISI is only 1 d B.

关 键 词：单输入多输出 盲均衡 吉布斯采样 软输入软输出 迭代均衡译码 

分 类 号：TN911.5[电子电信—通信与信息系统] TN92[电子电信—信息与通信工程]