Title :
A system identification technique based on neural networks
Author :
Wabgaonkar, H. ; Stubberud, A.
Author_Institution :
California Univ., Irvine, CA, USA
Abstract :
A system identification technique based on neural networks is presented. A feedforward-type neural network is trained using an extended Kalman filter, so as to capture the input-output characteristics of a dynamical system. The proposed technique can be enhanced to simultaneously obtain the estimates of the plant states and those of the neural network parameters
Keywords :
Kalman filters; feedforward neural nets; identification; dynamical system; extended Kalman filter; feedforward-type neural network; input-output characteristics; parameter estimation; state estimation; system identification; Artificial neural networks; Equations; Feedforward neural networks; Iterative algorithms; Network synthesis; Neural networks; Nonlinear dynamical systems; State estimation; State-space methods; System identification;
Conference_Titel :
Systems Engineering, 1992., IEEE International Conference on
Conference_Location :
Kobe
Print_ISBN :
0-7803-0734-8
DOI :
10.1109/ICSYSE.1992.236884
