A predictive controller was designed for car velocity control. The cost function of the predictive controller consists of three terms: sum of the absolute value of the control error during the prediction horizon, sum of the absolute value of the control increments during the prediction horizon plus the absolute value of the error at the end of the prediction horizon. The process was modeled by a neuronal net. The artificial network was trained off-line by the method of Levenberg-Marquardt. The controller was tuned by an evolutional algorithm. Simulation results are presented.
Nonlinear system control using artificial neural network and differencial evolution
International Conference Process Control, 9 ; C030a/1-C030a/7
2010
7 Seiten, 8 Bilder, 7 Quellen
Aufsatz (Konferenz)
Datenträger
Englisch
Ship manoeuvrability analysis using the differencial approach
Tema Archiv | 1987
|Predicting the Evolution of Suspended Sediment Concentration Using Artificial Neural Network
HENRY – Bundesanstalt für Wasserbau (BAW) | 2010
|Nonlinear flight control using adaptive neural network
British Library Online Contents | 1998
|Artificial Neural Network Based Control Systems
SAE Technical Papers | 2003
|