Modeling and Vibration Control with Neural Network for Flexible Multi-Link Structure.

Abstract

Vibration control for flexible structures, such as flexible robot arms and flexible space structures, has been studied by many researchers. We proposed decentralized control for flexible structures by decoupling mode quantities of other linkS. But when modeling errors exist in the control system, the performance of the decoupling control goes down because the control method is based on the model. In order to consider of modeling error, the adaptive control based neural network is applied to the vibration control of flexible multi link system. By considering the axial forces in modeling, mode dynamics is given as nonlinear system. The effectiveness of the proposed control method is shown through simulations of 5 link robotic arm.