Study On Adaptive Tracking Control Of Uncertain Robot Systems

Robot system is not only a kind of complex system which is multivariable, nonlinear, highly-coupled and time-varying, but also subjected to various kinds of uncertainties, such as measurement errors, modeling errors, varying payload, external disturbances, and so on. So it is very difficult to establish the precise and integrated dynamics of robot system. With the rapid development of modern industry, the requirements of control accuracy and stability for robot systems are more and more paid attention to. Therefore, the study on trajectory tracking control of uncertain robot systems is important to both theory and practice. In this paper, based on a series of theories, which include Lyapunov stability theory, adaptive control theory, fuzzy control theory, integral variable structure control theory, backstepping control theory, robust control theory, and so on, some adaptive trajectory tracking control schemes are proposed for a class of uncertain robot systems. The main contents are outlined as follows.Firstly, based on the principle of variable structure control, using the simplified dead-zone model and the approximation capability of fuzzy systems with linear adjustable parameters which are used to approximate unknown plant function, a novel integral variable structure adaptive fuzzy control strategy is proposed for a class of uncertain robot manipulator with unknown dead-zones and external disturbances. The approach does not require the optimal approximation error to be square-integrable. Meanwhile, by utilizing Young’s inequality, the number of adjustable parameters in the fuzzy systems and the realization of complexity are reduced. Then, using the Lyapunov method, the closed-loop control system is proved to be semi-globally uniformly ultimately bounded, with the tracking errors converging to a small neighborhood of zero by appropriately choosing design constants. Simulation results show that the proposed approach is feasible and practical.Secondly, the problem of adaptive fuzzy control for a class of uncertain robot manipulator with external disturbances is discussed. Based on the principle of sliding mode control, by using a regularized matrix inverse, a novel fuzzy adaptive tracking control strategy is proposed. The fuzzy systems with linear adjustable parameters are used to approximate process unknown functions. The approach does not require the M(q) to be known and the optimal approximation error to be square-integrable. The robust adaptive compensation term is adopted to minify the influence of modeling errors and parameter estimation errors. Then, using the Lyapunov method, the closed-loop control system is proved to be uniformly ultimately bounded, with the tracking errors converging to zero. Simulation results show that the proposed approach is feasible and practical.Lastly, based on the principle of T-S fuzzy model, using the approximation capability of fuzzy systems with linear adjustable parameters which are used to approximate unknown plant function, according to the design of model reference tracking control system, a novel robust tracking control strategy combination of states observer, adaptive fuzzy control and Backstepping control is proposed for a class of uncertain robot manipulator with external disturbances. The proposed approach only requires the feedback of link position states without the velocity information. The design is based on the principle of the adaptive states observer and the technique of Backstepping control. The immeasurable link velocity for the robot model is estimated by the observer. Then, the control law is determined using the states estimates and the Backstepping control technology. By using of the Lyapunov method, the closed-loop control system is proved to be uniformly ultimately bounded, with the observer errors and tracking errors converging to a small neighborhood of zero. Simulation results show that the proposed approach has stronger robustness and good tracking performance.All the strategies suggested in this thesis are proved strictly in theories. And their practicality and validity are verified in simulation. Through the research in this thesis, new ideas are provided for dealing with some practical problems faced by robot manipulators in engineering.