Robust Control Of Robot With Uncertainties Based On PD+Feedforward

The control problems of robotic manipulators have received great attention in theoretical research and engineering for many years. Robust control is just one of the most heated problems. It is well known that the robotic manipulator is a very complicated MIMO nonlinear system with time-varying strong-coupling and nonlinear dynamic characteristics, so the control for such a system is quite difficult, we have to face a lot of uncertainties. So the research on uncertain robot manipulators has theoretical and practical importance. In this dissertation, the system of robotic manipulators with entire dynamic model, namely, the robotic system with uncertainties is regarded as controlled plant and the various compensation schemes based on PD+feedforward are developed on the base of the references available. The dissertation gives a brief description about the developing situation and control method firstly, and then the models of robot and mathematics knowledge are introduced in detail. Subsequently three classes of control strategies with compensation control structure, which are based on PD+feedforward, are proposed. Namely, Variable Structure Compensation Control, Neural Network Compensation Control and Robust Adaptive Control. The overall idea is that the system of robotic manipulators is decomposed as two parts: one is parameter uncertainty and the other is unparameter uncertainty. We compensated them distinguishingly. Parameter linearization and intelligent methods are used for compensating parameter uncertainties. Robust control strategies are utilized to solve the unparameter uncertainties. Besides, the study on robot with bounded inputs is proposed in this dissertation. In the condition of torque saturation of robot, an effective controller can be designed. It provides a good way for saturation of execution mechanism when robot is applied in the practical occasion.