Neural network architectures for parameterestimation of dynamical systems
Jitendra R. Raol, H. Madhuranath · IEE Proceedings - Control Theory and Applications · 1996
Various recurrent neural network architectures for solving the problems of parameter estimation in dynamical systems are presented. The architectures based on precomputation of weight/bias information (Hopfield neural network), direct gradient computation with and without normalisation and output error method are developed. A typical computer simulation result is given.