The effect of the nonlinearity of signal-processing algorithms on the statistical characteristics of adaptive filters
Abstract
The statistical characteristics of adaptive filters are determined for linear and nonlinear signal-processing algorithms. These characteristics include the convergence time, the characteristics of the error signal amplitude, and the variance of the weight-coefficient fluctuations. It is noted that the fluctuations of the parameters of input signals do not produce a significant change in the tuning speed of the adaptive filter. The use of a nonlinear algorithm makes it possible to obtain high tuning speed and to increase the accuracy of tuning; however, this is accompanied by an increase in the variance of weight-coefficient fluctuations.
- Publication:
-
Radiofizika
- Pub Date:
- 1982
- Bibcode:
- 1982RaF....25..211M
- Keywords:
-
- Adaptive Filters;
- Algorithms;
- Nonlinear Filters;
- Signal Processing;
- Statistical Analysis;
- Convergence;
- Error Signals;
- Tuning;
- Weighting Functions;
- Electronics and Electrical Engineering