Title :
Adaptive filtering of nonlinear systems with memory by quantized mean field annealing [digital subscriber loop example]
Author :
Nobakht, Ramin A. ; Ardalan, Sasan H. ; Van den Bout, David E.
Author_Institution :
IBM Corp., Research Triangle Park, NC, USA
fDate :
2/1/1993 12:00:00 AM
Abstract :
A technique for adaptive filtering of nonlinear systems with memory that combines quantized mean field annealing (QMFA) and conventional recursive-least-squares/fast-transversal-filter (RLS/FTF) adaptive filtering is developed. This technique can efficiently handle large-order nonlinearities with or without memory. The nonlinear channel is divided into a memory nonlinearity followed by a dispersive linear system. QMFA is applied to obtain the coefficients and the order of the memory of the nonlinearity, and RLS/FTF is applied to determine the weights of the dispersive linear system. Statistical thermodynamic analysis that provides theoretical measures for making annealing algorithms computationally efficient. The method is applied to a full duplex digital subscriber loop. Simulations show a performance improvement of over 40 dB compared to ordinary RLS/FTF and steepest descent algorithms, and the solution is robust
Keywords :
adaptive filters; digital communication systems; filtering and prediction theory; least squares approximations; simulated annealing; subscriber loops; QMFA; RLS/FTF; adaptive filtering; dispersive linear system; full duplex digital subscriber loop; nonlinear systems with memory; quantized mean field annealing; recursive-least-squares/fast-transversal-filter; Adaptive filters; Algorithm design and analysis; Annealing; Computational modeling; DSL; Dispersion; Linear systems; Nonlinear systems; Resonance light scattering; Thermodynamics;
Journal_Title :
Signal Processing, IEEE Transactions on