Controlling A Motion Platform Beyond Its Anti-Resonance

Abstract

Abstract In controlling motion platforms, the servo motor and the feedback sensor are in general, not co-located. Hence, there usually happens to be a mechanical anti-resonance in controlling the system in velocity mode. This anti-resonance hinders the controller design as the open-loop gain of the system is severely reduced. In this study, a novel approach based on self-recurrent wavelet neural networks (SRWNNs) has been proposed in order to re-shape the anti-resonance behavior of the open-loop plant in a robust manner. The approach has been verified both using simulations and experiments.