Title :
Neural modelling of dynamic systems with non-measurable state variables
Author :
Alippi, Cesare ; Piuri, Vincenzo
Author_Institution :
CSISEI, CNR, Milano, Italy
Abstract :
The paper deals with neural modelling of dynamic processes. Attention is focused on processes characterised by non-measurable states and their modelling with nonlinear recurrent neural networks. A relationship is developed which, for such models, correlates the actual prediction error with the past ones
Keywords :
identification; modelling; recurrent neural nets; state-space methods; actual prediction error; blackbox model; drum-type boiler model; dynamic systems; equivalent discrete time system; mean square error; neural modelling; neural output error; nonlinear recurrent neural networks; nonmeasurable state variables; regression-type static neural net; state-space representation; system identification; Approximation error; Computational modeling; Computer networks; Error correction; Hybrid power systems; Nonlinear dynamical systems; Nonlinear equations; Predictive models; Recurrent neural networks; System identification;
Conference_Titel :
Instrumentation and Measurement Technology Conference, 1997. IMTC/97. Proceedings. Sensing, Processing, Networking., IEEE
Conference_Location :
Ottawa, Ont.
Print_ISBN :
0-7803-3747-6
DOI :
10.1109/IMTC.1997.612371
