New results on the performance of a wellknown class of adaptive filters
Abstract
The adaptation algorithms studied are introduced with the aid of an idealized identification problem involving an unknown system with a continuous vector input of known dimension and a scalar output. A misalignment vector is defined. Proof is provided for a uniform exponential convergence to zero of the norm of the misalignment vector. Upper and lower bounds to the convergence rate are derived. Attention is given to the properties of the linear system in response to a special input.
 Publication:

IEEE Proceedings
 Pub Date:
 November 1976
 Bibcode:
 1976IEEEP..64.1583S
 Keywords:

 Adaptive Filters;
 Error Signals;
 Feedback Control;
 Linear Systems;
 Signal Analysis;
 Algorithms;
 Analog Computers;
 Communication Theory;
 Control Theory;
 Convergence;
 Linear Equations;
 Nonlinear Equations;
 Electronics and Electrical Engineering