MIMO系统中基于贝叶斯的稀疏信道估计算法  

Bayesian based Sparse Channel Estimation Algorithm in MIMO Systems

作　　者：侯小静[1] 吕治国[1,2] 孙泽宇[1] 邢萧飞[3] 

机构地区：[1]洛阳理工学院计算机与信息工程学院,河南洛阳471023 [2]西安电子科技大学综合业务网理论及关键技术国家重点实验室,西安710071 [3]广州大学计算机科学与教育软件学院,广州510006

出　　处：《光通信研究》2017年第4期62-65,共4页Study on Optical Communications

基　　金：国家自然科学基金资助项目(61503174);河南省高等学校青年骨干教师资助计划(2016GGJS-158);河南省教育厅自然科学基金资助项目(17A520044);广东省自然科学基金面上项目(2016A030313540);广州市科技计划项目(201707010284);广州市市属高校科研项目(1201430560)

摘　　要：针对未知支持集贝叶斯匹配追踪算法中支持集搜索过于复杂的问题,提出一种改进的贝叶斯稀疏信道估计算法。该算法利用信道的稀疏特性来提高信道估计精度,通过计算信道期望来估计高维稀疏信道。为了提高支持集的搜索准确度,改进算法在支配支持集的搜索过程中引入了维特比路径译码的思想,保留所有发生概率差异低于门限的路径,并对支持集上的信道估计值进行计算,最后计算出信道期望并作为最后的信道估计值。理论分析和仿真结果都表明,改进算法在降低算法复杂度的同时还保持了较高的估计精度。Due to the high complexity of the Support Agnostic Bayesian Matching Pursuit （SABMP） algorithm in the supportset search, a modified Bayesian sparse channel estimation algorithm is proposed in this paper. The modified Bayesian algorithmcan improve the channel estimation accuracy by exploiting the sparse characteristics of the channel. The high-dimensionalsparse channel is estimated by calculating the channel expectation. In order to improve the accuracy of the support set search,the idea of Viterbi path decoding is introduced in the search process of dominating support set. All the paths with probabilitydifference below the threshold are preserved. Then, the channel estimation on the chosen support set is carried out and thechannel expectation is calculated as the final channel estimate. Theoretical analysis and simulation results show that the lowercomputational complexity is maintained and the channel estimation accuracy is kept at the same time.

关 键 词：贝叶斯估计 压缩感知 维特比译码 贪婪算法 稀疏重建 

分 类 号：TN911.71[电子电信—通信与信息系统]