Tracking Control of Parallel Robot Manipulators Using RBF Neural Network

Abstract

Since parallel robots are the multibody systems with closed-loop structures, their movement equations usually are in the complex form of redundant coordinates and their dynamics parameters are usually uncertain. The aim of this paper is to improve the control quality for the parallel robot by applying RBF neutron network. Firstly, the movement equations of Rostock Delta robot are established as differential–algebraic systems of equations with redundant generalized coordinates. Then, the stableness of the control method based on sliding mode control law using neural network is proved. Finally, the error in tracking control of a specific Rostock Delta robot is simulated by using this method.