Research On Uncertain Robot Based On Intelligent Variable Structure Compensation Control

The control problems of robot have received great attention in theoretical research and engineering at all times. It is well known that the model-based scheme popularly known as Computed Torque Control (CTC) is effective and its performance is excellent in various control strategies for robot. However, in practice, unfortunately, it is impossible to obtain a perfect or even reasonably accurate dynamic model of a robot. Furthermore, the parameters of dynamics model of robot may also be subject to change when the manipulator goes about its task. Meanwhile, the system can be influenced by uncertainties such as external disturbance and payload change. Thus, it is essential for CTC to be improved.In this dissertation, the system of robot with entire dynamic model, namely, the robotic system with uncertainties is regarded as controlled object, and the various intelligent variable structure compensation schemes based CTC are emphatically developed on base of the references available.Firstly, the dissertation gives a brief description about the developing situation and control theory of robot, and then the underlying idea and main characteristic of CTC and CTC adding variable structure compensation are introduced in detail. Subsequently three control strategies with compensation control scheme which are based on CTC adding intelligent variable structure are proposed. Namely, fuzzy variable structure compensation control, Radial basis function neural network(RBFNN) compensation control and fuzzy neural network direct adaptive variable structure compensation control. The basic idea is that the robot system with uncertainty is decomposed as two parts: one is nominal system with perfect knowledge of dynamic model and the other is system with uncertainties. CTC is used to control nominal system. For uncertainties system, we adopt intelligent control method such as fuzzy andneural network, and variable structure control to design different intelligent variable compensation controllers. The outputs of the two parts together are regarded as the control input of the robotic system. This proposed control algorithm can not only solve the dithering brought by adopting variable structure compensation control solely, but also make robot system achieve be global consistent, more to be bounded or asymptotic steady. The stability of controller designed is analyzed, and the simulation results are presented for the same 2-DOF serial robotic manipulator, which validate the effectiveness and feasibility of the proposed schemes.Besides, a new algorithm based on distance fuzzy learning algorithm for solving the inverse kinematics problem of robot is proposed in the dissertation.