Distributed Evolutionary Digital Filters for IIR Adaptive Digital Filters

Masahide ABE  Masayuki KAWAMATA  

IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences   Vol.E84-A   No.8   pp.1848-1855
Publication Date: 2001/08/01
Online ISSN: 
Print ISSN: 0916-8508
Type of Manuscript: Special Section PAPER (Special Section on Digital Signal Processing)
Category: Adaptive Signal Processing
adaptive digital filter,  IIR filter,  genetic algorithm,  evolutionary computation,  

Full Text: PDF>>
Buy this Article

This paper proposes distributed evolutionary digital filters (EDFs) as an improved version of the original EDF. The EDF is an adaptive digital filter which is controlled by adaptive algorithm based on evolutionary computation. In the proposed method, a large population of the original EDF is divided into smaller subpopulations. Each sub-EDF has one subpopulation and executes the small-sized main loop of the original EDF. In addition, the distributed algorithm periodically selects promising individuals from each subpopulation. Then, they migrate to different subpopulations. Numerical examples show that the distributed EDF has a higher convergence rate and smaller steady-state value of the square error than the LMS adaptive digital filter, the adaptive digital filter based on the simple genetic algorithm and the original EDF.