Speed Control of Two-Mass Elastic System by On-line Trained Neural Network based on Kalman Filter Estimation
[ 1 ] Instytut Automatyki i Inżynierii Informatycznej, Wydział Elektryczny, Politechnika Poznańska | [ P ] pracownik
2014
referat
angielski
EN The article presents the problem of adaptive motor control for the two mass drive system with elasticity using online trained neural speed controller, where controlled signals obtain by Kalman Filter. Structure of the controller with training algorithm and idea of KF are shortly described. For ANN training process the Resilient Back Propagation algorithm (RPROP) was chosen. Only the measurement of motor position assumed as being sufficient to the possibility of torsional vibration damping. Results, in change to drive system parameters, are presented.
171 - 172