Experimental Study of Active Vibration Control for Flexible Truss by Using a Stewart Platform Manipulator

Abstract

Experiments of active vibration control for flexible truss are studied, with a Stewart platform manipulator (SPM) employed as an active device at the bottom of the truss. A back-propagation neural network is designed to identity the controlled object. And on the basis of the identification model, this work constructs a neural network adaptive inverse controller (NNAIC). Then based on the dSPACE real-time simulation system, the control test system for the truss is built, on which the neural network adaptive inverse control experiments are carried out. The vibration amplitude decreases more than 93.8% and 89.6% with the structure continuously being excited at the first-order and second-order resonant frequency respectively. The results indicate the effectiveness of the control conducted by using SPM, and NNAIC has a good performance.