基于压缩感知的分布式协同估计算法 - 电讯技术.pdfVIP

第57卷 第4期 电讯技术 Vol.57,No.4 2017年4月 Telecommunication Engineering April,2017 doi:10.3969/ j.issn.1001-893x.2017.04.001 引用格式:张亚东,姚彦鑫.基于压缩感知的分布式协同估计算法[J].电讯技术,2017,57(4):377-381. [ZHANGYadong,YAOYanxin.Adistrib鄄 uted collaborative estimation algorithm based on compressed sensing[J].Telecommunication Engineering,2017,57(4):377-381. ] 基于压缩感知的分布式协同估计算法* 张亚东,姚彦鑫** (北京信息科技大学 信息与通信工程学院,北京 100101) 摘摇 要:为了降低分布式协同估计算法的计算量并改善其收敛性能,提出了基于压缩感知(CS)和递 归最小二乘(RLS)的分布式协同估计算法。 该算法在传统RLS分布式协同估计算法的基础上引入 压缩感知技术,首先在压缩域中进行递归最小二乘运算,然后利用压缩感知重构算法得到未知参数 向量的估计值。 提出的算法能够在增量式策略和两种模式的扩散式策略下实现对未知向量的有效 估计。 理论分析和仿真结果表明,该算法一方面降低了RLS分布式协同估计算法的计算量,另一方 面保持较快的收敛速度与良好的均方误差性能。 关键词:分布式估计;压缩感知;递归最小二乘;增量式策略;扩散式策略 中图分类号:TN911摇 摇 文献标志码:A摇 摇 文章编号:1001-893X(2017)04-0377-05 A Distributed Collaborative Estimation Algorithm Based on Compressed Sensing ZHANG Yadong,YAO Yanxin (School of Information and Communication Engineering,Beijing Information Science Technology University,Beijing 100101,China) Abstract:Inordertoreducetheamount of calculation andimprovetheconvergenceperformance,adistrib鄄 uted collaborative estimation algorithm based on compressed sensing and recursive least square(RLS) is proposed. Based on the traditional RLS distributed collaborative estimation algorithm,this algorithm intro鄄 duces compressed sensing technology. Firstly,the recursive least square method is applied in the com鄄 pressed domain. Andthentheestimationoftheunknownparametervectorisachievedby compressedsens鄄 ingreconstruction algorithm. Theproposedalgorithmcan effectively estimatetheunknownvectorunderthe inc