Comparative Analysis of QR-RLS and IQR-RLS Algorithms
Long-term numerical instability is a critical problem in fast recursive least squares adaptive algorithms, which has received extensive attention in recent years. Both the QR-RLS and inverse QR-RLS algorithms have the property of numerical robustness. So, the QR-RLS and inverse QR-RLS algorithms are analyzed by means of two applications: One application is stationary system identification. Another is to track a time-varying system. Numerical simulation results in this paper demonstrate that both the QR-RLS and inverse QR-RLS algorithms preserve the desirable convergence properties of the standard RLS algorithm. As far as the rate of convergence is considered, the inverse QR-RLS algorithm is a better choice from the QR-RLS and inverse QR-RLS algorithms. However, the QR-RLS is better than the inverse QR-RLS in term of the squared error. Both the QR-RLS and inverse QR-RLS algorithms have good performance on tracking a nonstationary system if the forgotten factor is set as reasonable value.
QR-RLS inverse QR-RLS Adaptive filter
Wang Quai
Ministry of Education Key Lab of Network control & Intelligent Instrument, Chongqing University of Posts and Telecommunications, Chongqing, China No.2 Chongwen Road, Chongqing,China. 400065 Key Laboratory of Wireless Sensor Network & Communication, Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, 865 Changning Road, Shanghai 200050 China
国际会议
2010 Second Asia-Pacific Conference on Information Processing(2010年第二届亚太地区信息处理国际会议 APCIP 2010)
南昌
英文
252-255
2010-09-17（万方平台首次上网日期，不代表论文的发表时间）