Use of adaptive learning radial basis function network in real-time motion tracking of a robot manipulator

In this paper, real time motion tracking of a robot manipulator based on the adaptive learning radial basis function network is proposed. This method for adaptive learning needs little knowledge of the plant in the design processes. So the centers and widths of the employed radial basis function network (RBFN) as well as the weights are determined adaptively. With the help of the RBFN, motion tracking of the robot manipulator is implemented without knowing the information of the system in advance. Furthermore, identification error and the tuned parameters of the RBFN are guaranteed to be uniformly ultimately bounded in the sense of Lyapunov's stability criterion.