DocumentCode :
1752300
Title :
Statistical analysis of neural network modeling and identification of nonlinear systems with memory
Author :
Ibnkahla, Mohamed
Author_Institution :
Electr. & Comput. Eng. Dept., Queen´s Univ., Kingston, Ont.
Volume :
1
fYear :
2001
fDate :
2001
Firstpage :
315
Abstract :
The paper presents a statistical analysis of neural network modeling and identification of nonlinear systems with memory. The nonlinear system model is comprised of a discrete-time linear filter H followed by a zero-memory nonlinear function g(.). The system is corrupted by input and output independent Gaussian noise. The neural network is used to identify and model the unknown linear filter H and the unknown nonlinearity g(.). The network architecture is composed of a linear adaptive filter and a two-layer nonlinear neural network (with an arbitrary number of neurons). The network is trained using the backpropagation algorithm. The paper studies the MSE surface and the stationary points of the adaptive system. Recursions are derived for the mean transient behavior of the adaptive filter coefficients and the neural network weights for slow learning. It is shown that the adaptive filter converges to a scaled version of the unknown filter H, and that the nonlinear neural network converges to an approximation of the unknown nonlinearity. Computer simulations show good agreement between theory and experimental results
Keywords :
Gaussian noise; adaptive filters; backpropagation; convergence of numerical methods; discrete time filters; identification; mean square error methods; neural net architecture; nonlinear systems; statistical analysis; transient analysis; Gaussian noise; MSE surface; backpropagation; convergence; discrete-time linear filter; learning; linear adaptive filter; mean transient behavior; neural network architecture; neural network modeling; neural network training; nonlinear systems identification; stationary points; statistical analysis; two-layer nonlinear neural network; zero-memory nonlinear function; Adaptive filters; Adaptive systems; Backpropagation algorithms; Computer simulation; Gaussian noise; Neural networks; Neurons; Nonlinear filters; Nonlinear systems; Statistical analysis;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Signal Processing and its Applications, Sixth International, Symposium on. 2001
Conference_Location :
Kuala Lumpur
Print_ISBN :
0-7803-6703-0
Type :
conf
DOI :
10.1109/ISSPA.2001.949841
Filename :
949841
Link To Document :
بازگشت