Research On Energy Control Strategy Of Uncertain Robot

Robotic system has time-varied, strong-coupled, highly nonlinear characteristics. When we establish the robot model, because of various uncertainties, it is very hard to formulate a precise and complete mathematical model of robot. Design methods based on the concept of dissipative control pay more attention to the energy attenuation process of the error system, and take advantage of the physical structure characteristics of the robot to construct the energy function. This can simple derivation of the controller, and has been widely applied in robot control.This paper studies trajectory tracking of uncertain robotic systems with application of the dissipative theory. The main research works can be summarized as follows:Aiming at uncertain robot system, a neural network robust control based on the theory of dissipative scheme is proposed. In this control program, based on Lyapunov stability theory, the controller is designed starting from stability of the whole closed-loop system. It uses neural network to adaptively learn the uncertainties of robot system, and takes approximate error as system's exterior disturbance, with the theory of dissipative interference suppression. This controller can guarantee the asymptotic convergence of the tracking error to zero, and ensure the stability of the whole system.When the external disturbance and unmodeled dynamics exist, first the controller is designed according to passivity. And then two cases are discussed. one kind is that the upper bounds of system uncertainties are known. Fuzzy control is introduced based on the traditional sliding mode control, in which the sliding surface is regarded as the input of fuzzy controller while the output is the weight of compensation controller. This controller can effectively eliminate chattering of sliding mode control. Another case is that the upper bounds are unknown. RBF network is used to adaptively learn the unknown upper bounds. This control scheme overcomes the limit which the conventional sliding mode control scheme need to known the upper bounds of system uncertainties, and eliminates chattering of the controller. The system obtains desired performance.