Adaptive filtering using modified conjugate gradient

An adaptive filtering algorithm is described that uses the modified Conjugate Gradient (CG) algorithm. It has the ability to perform sample-by-sample updating of the filter coefficients more efficiently than previously described CG methods. Its performance can be comparable to the RLS and LMS-Newton algorithms, giving fast convergence for highly correlated input signals, while maintaining low misadjustment. Simulations demonstrating its performance and the influence of various parameter choices are shown. A convergence criterion is also derived.

[1]  David G. Luenberger,et al.  Linear and nonlinear programming , 1984 .

[2]  M. Al-Baali.  Descent Property and Global Convergence of the Fletcher—Reeves Method with Inexact Line Search , 1985 .

[3]  Robert J. Plemmons,et al.  FFT-based RLS in signal processing , 1993, 1993 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[4]  W. K. Jenkins,et al.  Preconditioned conjugate gradient methods for adaptive filtering , 1991, 1991., IEEE International Sympoisum on Circuits and Systems.

[5]  S. Haykin,et al.  Adaptive Filter Theory , 1986 .

[6]  M. Srinath,et al.  Conjugate gradient techniques for adaptive filtering , 1992 .

[7]  R. Fletcher.  Practical Methods of Optimization , 1988 .