基于聚类的多进制扩频伪码序列盲估计方法  

Clustering Based Blind Estimation of PN Sequences in Mary Spread Spectrum System

作　　者：李军伟[1] 张天骐[1] 朱洪波[1] 潘毅[1] 

机构地区：[1]重庆邮电大学,信号与信息处理重庆市重点实验室,重庆400065

出　　处：《科学技术与工程》2014年第2期32-36,共5页Science Technology and Engineering

基　　金：国家自然科学基金项目(61071196,611021 31);教育部新世纪优秀人才支持计划项目(NCET-10-09 27);信号与信息处理重庆市市级重点实验室建设项目(CSTC2009CA 2003);重庆市杰出青年基金项目(CSTC201 1jjjq40002);重庆市自然科学基金项目(CS TC2010BB2398,CSTC2010BB2409,CST C2010BB2411,CSTC2012JJA40008);重庆市教育委员会科研项目(KJ120525)资助

摘　　要：针对多进制扩频系统伪码序列估计的问题,借鉴无监督聚类分析的思想,提出了一种基于K均值聚类算法的多进制扩频系统伪码序列估计方法。该方法首先将多进制扩频信号分段成不重叠的信号向量构造数据集合,利用数据集合的自相关矩阵的检测统计量完成信号的盲同步;其次通过搜索基于K均值聚类的代价函数来估计伪码集合规模,代价函数曲线的拐点位置为伪码集合规模的估计值;最后利用数据集合的聚类特征完成伪码序列的估计。理论分析和计算机仿真表明,该方法可以准确估计多进制扩频系统伪码时延和伪码序列值。To solve the problem of blind estimation of the pseudo-noise（PN） sequence in Mary spread spectrum system（ MSSS）, relevving to the idea of unsupervised clustering analysis, a method of estimating PN sequences in Mary spread spectrum system was proposed based on K-means clustering algorithm. In this method, the MSSS sig-nal was divided into non-overlapped individuals to construct a data set, and the blind synchronization of the PN se-quence was completed by the detection statistic of the data set＇ s autocorrelation matrix ; then the PN sequences set＇ s scale was estimated by searching the cost function based on K-means clustering algorithm, and the cost function＇ s graph exhibited a significant inflection point, as well as its position indicated the scale of PN sequences set; final-ly the PN sequences was estimated by exploiting the clustering property of the data set. Theoretical analysis and computer simulations showed that this method could accurately estimate the time delay and PN sequences in MSSS.

关 键 词：多进制扩频 检测统计量 K均值聚类 代价函数 伪码估计 

分 类 号：TN911.7[电子电信—通信与信息系统]