快重复调整NLMS演算法的分析.pdfVIP

類 93 年 24(1) 95-112 數例 量 量理論 度數 量 93 年08 20 93 年09 23 93 年09 30 95 Analysis of the Fast Data-Reusing NLMS Algorithm Heng-Chou Chen Department of Computer and Communication Engineering, Chienkuo Technology University Abstract This paper analyzes the performance of the data-reusing scheme used in NLMS algorithm, so that it becomes data-reusing normalized least-mean-square algorithm (DRNLMS). The computation equation deduced from such algorithm is found to have the repeated-updating scaling factors of its coefficients precalculated with 3 multiplications and 2 additions regardless of the input signals, and the computation complexity is the same as that of the NLMS. Both the analysis of convergence performance and computer simulation results demonstrate that DRNLMS algorithm with higher reusing times has a more rapid convergence speed. Key words: LMS, data-reusing, convergence performance, computational complexity 96 行識 數濾 來類濾 濾數 兩 量M M 濾 度度 度量M 2 率例 濾 度數 見行理 力行 量陸來 不量度降 濾數 兩裡 濾 ~ T ( ) ( ) y (n) w n x n e(n) d (n) −y (n) 濾數 ~ ~ w(n +1) w(n) +µe(n)x(n) x(n) [x (n) x (n −1) L x (n −M +1)]T d (n) w(n) y (n) 濾來d (n)