Paper Info
Title
A robust and effective fuzzy adaptive equalizer for powerline communication channels
Abstract
Fuzzy adaptive equalizers (FAEs) are adaptive equalizers that apply the concepts of fuzzy logic. The main merit of applying FAEs in powerline channel equalization is that linguistic information (fuzzy IF-THEN rules) and numerical information (input-output pairs) can be combined into the equalizers. The adaptive algorithms adjust the parameters of the membership functions which characterize the fuzzy concepts in the IF-THEN rules, by minimizing some criterion function. In this paper, we propose a new FAE, using the extended Kalman filter (EKF) algorithm for powerline channel equalization. The simulation results show that the EKF-based FAE has lower steady state bit error rate (BER) and faster convergent speed compared to decision feedback recursive least-squares adaptive equalizer, recursive least squares (RLS) based FAE and least mean squares (LMS) based FAE. We also propose a robust improvement scheme for the new FAE. Simulation results show that the performance of the proposed robust FAE is improved in powerline channel equalization and outperforms all the equalizers considered above. The BER of the proposed scheme is very close to the optimum performance.
Year
DOI
Venue
2007
10.1016/j.neucom.2006.12.018
Neurocomputing
Keywords
DocType
Volume
extended kalman filters,ekf-based fae,simulation result,powerline communication,effective fuzzy adaptive equalizer,fuzzy adaptive filters,channel equalization,adaptive algorithm,new fae,powerline communication channel,adaptive equalizer,fuzzy if-then rule,fuzzy adaptive equalizer,impulsive noise channels,fuzzy concept,powerline channel equalization,proposed robust fae,communication channels,adaptive filter,input output,extended kalman filter,fuzzy logic,membership function,least mean square,steady state,bit error rate,impulse noise
Journal
71
Issue
ISSN
Citations 
1-3
Neurocomputing
2
PageRank 
References 
Authors
0.45
16
2
Name
Order
Citations
PageRank
Wai Kit Wong139420.10
Heng-Siong Lim2459.65