Statistical analysis of the characteristics of an adaptive predictive filter
Abstract
The paper examines the statistical characteristics of an adaptive transversal predictive filter with complex weight coefficients at the delayline taps; an analog complex algorithm of least meansquare error is used for the automatic adjustment of the weight coefficients. The input signals are represented as an additive mixture of several narrowband random processes and broadband noise. Analytical expressions are obtained for stationary mean values of the weight coefficients, the complex transfer function of the adaptive filter, and the errorsignal power. The effect of fluctuations of the adaptively adjusted parameters on these characteristics is examined.
 Publication:

Radiotekhnika i Elektronika
 Pub Date:
 August 1982
 Bibcode:
 1982RaEl...27.1563M
 Keywords:

 Adaptive Filters;
 Signal Processing;
 Statistical Analysis;
 Algorithms;
 Broadband;
 Least Squares Method;
 Prediction Analysis Techniques;
 RootMeanSquare Errors;
 Weighting Functions;
 Electronics and Electrical Engineering