Service Communautaire d'Information sur la Recherche et le Développement - CORDIS

Iterative learning algorithm using non-causal filters to control large complex structures subjected to repetitive events

A new innovative control algorithm has been developed wherein non-causal filters have been introduced, capable to control large complex structures subjected to repetitive excitation events.

The control algorithm is an iterative learning algorithm suitable to control repetitive events. It uses information from previous events to enhance its performance when controlling new upcoming events. The innovation, which has been added by our research, is the introduction of non-causal learning filters in the controller.

These are filters which work with future information and have attractive properties to stabilise control transfer functions. In a conventional feedback controller, information from the future is not available and these filters cannot be used. The iterative learning controller is able to reconstruct ¿future¿ information from the past events.

The controller needs a trigger signal, which announces the next event. The non-causal filters cause the controller to respond to the event in advance. The response and the event need to be synchronised.

Informations connexes

Reported by

KULeuven, Faculty of Engineering, Department of mechanical engineering, PMA
Celestijnenlaan 300B
3001 Leuven
See on map