偏差补偿比例更新互相关熵算法  

Bias-Compensated Proportional Correntropy Algorithm

作　　者：郑栋桥 张志禹[1] 马文涛[1] 邱进哲 ZHENG Dong-qiao;ZHANG Zhi-yu;MA Wen-tao;QIU Jin-zhe(Scholl of Automation and Information Engineering, Xi ’an University of Technology, Xi ’an, Shaanxi 710048, China)

机构地区：[1]西安理工大学自动化与信息工程学院,陕西西安710048

出　　处：《信号处理》2019年第1期75-85,共11页Journal of Signal Processing

基　　金：国家自然科学基金重大项目课题四(41390454);陕西省自然科学基础研究计划资助项目(2017JM6033);陕西省教育厅专项科研计划项目(15JK1379)

摘　　要：为了解决输入信号含有噪声和非高斯输出噪声的稀疏系统辨识问题,本文提出一种偏差补偿比例更新互相关熵算法。基于互相关熵的自适应滤波算法可以消除非高斯观测噪声的影响,进一步应用无偏准则来解决含噪输入信号带来的估计偏差问题。另外,将比例更新机制引入算法,通过自适应调节步长参数以增强算法的跟踪性能。仿真结果表明所提算法对于输入信号受噪声干扰和非高斯输出噪声环境下的稀疏系统辨识问题具有强的鲁棒性和稳态性能。To address the sparse system identification problem under noisy input and non-Gaussian output measurement noise, a sparse bias-compensated normalized maximum correntropy criterion (MCC) algorithm is developed in this work. The MCC based adaptive filtering algorithm can eliminate the impact of non-Gaussian observation noise, and the unbiasedness criterion is introduced to solve the bias estimate issue caused by the noisy input. In addition,the proportionate update scheme is utilized in the proposed algorithm,which can improve the tracking ability by adjusting step-size parameter adaptively. Simulation results confirm that the proposed algorithm shows robustness and steady-state performance for the sparse system identification problem under noisy input and output noise with non-Gaussian environments.

关 键 词：偏差补偿 稀疏系统辨识 互相关熵 含噪输入 

分 类 号：TN911.7[电子电信—通信与信息系统]