长春理工大学学报(自然科学版)
長春理工大學學報(自然科學版)
장춘리공대학학보(자연과학판)
JOURNAL OF CHANGCHUN UNIVERSITY OF SCIENCE AND TECHNOLOGY(NATURAL SCIENCE EDITION)
2015年
1期
143-147
,共5页
LMS算法%自适应滤波%小波变换%收敛速度
LMS算法%自適應濾波%小波變換%收斂速度
LMS산법%자괄응려파%소파변환%수렴속도
LMS algorithm%adaptive filtering%wavelet transform%convergence rate
介绍自适应滤波算法、多尺度小波算法的基本原理和两种算法结合的实现过程。针对最小均方(LMS)自适应滤波算法不能同时提高收敛速度和收敛精度,提出变步长LMS自适应算法,在滤波过程中算法先用大步长跟踪,提高收敛速度,接近稳态时用小步长跟踪,提高收敛精度。为了能有更好的滤波效果,应该在算法的步长因子上有所突破。在抽样函数的基础上改进算法,并结合多尺度小波分解,使得滤波的效果更加理想。通过Matlab仿真实验,验证了改进算法具有更好的稳定性和优越性。
介紹自適應濾波算法、多呎度小波算法的基本原理和兩種算法結閤的實現過程。針對最小均方(LMS)自適應濾波算法不能同時提高收斂速度和收斂精度,提齣變步長LMS自適應算法,在濾波過程中算法先用大步長跟蹤,提高收斂速度,接近穩態時用小步長跟蹤,提高收斂精度。為瞭能有更好的濾波效果,應該在算法的步長因子上有所突破。在抽樣函數的基礎上改進算法,併結閤多呎度小波分解,使得濾波的效果更加理想。通過Matlab倣真實驗,驗證瞭改進算法具有更好的穩定性和優越性。
개소자괄응려파산법、다척도소파산법적기본원리화량충산법결합적실현과정。침대최소균방(LMS)자괄응려파산법불능동시제고수렴속도화수렴정도,제출변보장LMS자괄응산법,재려파과정중산법선용대보장근종,제고수렴속도,접근은태시용소보장근종,제고수렴정도。위료능유경호적려파효과,응해재산법적보장인자상유소돌파。재추양함수적기출상개진산법,병결합다척도소파분해,사득려파적효과경가이상。통과Matlab방진실험,험증료개진산법구유경호적은정성화우월성。
This paper describes the implementation process of adaptive filtering algorithms, the basic principles of multi-scale wavelet algorithm and a combination of two algorithms. As for the least mean square (LMS) adaptive fil-tering algorithm can not improve convergence speed and accuracy, so the proposed variable step size LMS algorithm, in the filtering process, the algorithm first with a stride length of track, improve the convergence speed, close to steady with a small step tracking, improve the convergence precision. In order to have a better filtering effect, it should be a breakthrough in the step factor algorithm. This paper improves the function on the basis of sampling algo-rithms, combined with multi-scale wavelet decomposition, so that the filtering effect is more ideal. By Matlab simula-tion to verify the improved algorithm has better stability and superiority.