Investigation On Dynamic Control Methods For A High-speed Parallel Manipulator

Aimed at a high-speed parallel manipulator, this dissertation deals with several dynamic control strategies to enhance its precision. Meanwhile, investigation on machine vision system has also been done to satisfy the demand of light industry. The following contributions have been made.With the aid of vector method, constraint equation, kinetic energy consideration, the inverse position model, Jacobian matrix and mass matrix have been developed. The reduced dynamic model which can be realized in real-time implementation has also been formulated by means of virtual work principle. This provides necessary theoretical foundation for the dynamic control.A model based control scheme employing computed-torque control has been presented, and its applications in three different path planning methods have also been discussed. Via combining the feedback of PID and nonlinearity, the strategy can compensate the following error which resulted from dynamic factors effectively. Furthermore, considering the effect of control, the most effective path planning method is selected to improve the performance of control.A kind of variable structure controller for high-speed parallel manipulator has been proposed. By employing the established dynamic model, introducing the reaching law, considering the model errors, parameter perturbations and other disturbances, the algorithm has invariance property in sliding mode. Based on above, to eliminate chattering phenomenon and ensure reaching speed, a single-input and single-output fuzzy controller has also been introduced.Vision software with real-time and high precision quality has been developed. Considering the request of vision theory and high-speed parallel manipulator, the vision software can match with control system and guide both the manipulator and belt.By using the simulation software ADAMS, the performance of manipulator which employing the designed control algorithms and the selected path planning method has been discussed. Simulation results reveal the control algorithm's validity and effectiveness.