Title :
Fast convergence transversal adaptive filtering algorithm for impulsive environment based on T distribution assumption
Author :
Sanubari, J. ; Tokuda, Keiichi
Author_Institution :
Dept. of Electron. Eng., Satya Wacana Univ., Indonesia
Abstract :
In this paper we propose an adaptive filter based on assumption that the error is t-distributed with X degree of freedom, The optimal system is updated by using the LMS-like algorithm. When the input is impulsive signals, the convergence of the algorithm with small X is faster than when large X is used. Simulation results also show that the convergence of the proposed method is faster than other LMS-variant that has been earlier proposed
Keywords :
FIR filters; adaptive filters; convergence of numerical methods; errors; filtering theory; least mean squares methods; LMS-like algorithm; T-distribution assumption; fast convergence adaptive filtering algorithm; impulsive environment; optimal system updating; transversal adaptive filtering algorithm; Adaptive filters; Atmosphere; Atmospheric modeling; Computer errors; Computer science; Convergence; Filtering algorithms; Least squares approximation; Noise reduction; Signal processing;
Conference_Titel :
Circuits and Systems, 2001. ISCAS 2001. The 2001 IEEE International Symposium on
Conference_Location :
Sydney, NSW
Print_ISBN :
0-7803-6685-9
DOI :
10.1109/ISCAS.2001.921160
