High Precision Control Of High-speed Parallel Manipulators

This dissertation investigates into the high-precision control of high-speed parallel manipulators for pick-and-place operations in many sectors such as electronics,food,pharmaceutical and other light industries,with particular focuses upon normalized modeling and analyses of the mechanical system,modeling and parameter identification of the servo system,and high-precision control technique.The following contributions have been made.The normalized kinematic and rigid-body dynamic models of high-speed parallel manipulators are established,based on which the characteristic of diagonal dominance of the manipulator's inertia matrix is demonstrated throughout the workspace.A novel method of calculating the equivalent load inertia of the servo motor is presented,which gives full consideration to the coupling effects from other chains and provides proper system parameters for the design of high-precision control technique.The general electromechanical model of the servo system is established and a parameter identification method is proposed by introducing the error evaluation function for purpose of minimizing the fitting error.By eliminating the undesirable high frequencies,the certain band for identification is determined,thus overcomes the sensitivity of full-band data fitting to the test noise and guarantees that the identification results will reflect the system's steady performance accurately.An approach to determine the feasible region of the servo driver's tunable parameters is presented in virtue of the Routh criterion,thus ensuring the anti-interference capability of the velocity loop.A control strategy based upon fuzzy regulation and feedforward regulation is proposed.The anti-interference capability of the fuzzy module is enhanced by using the non-uniform membership function.The fuzzy reasoning process is simplified by establishing the fuzzy control look-up table.The joint torque is compensated on coarse interpolation level by real-time amending the value of the corresponding variable.An approach for parameter tuning of the motion controller is proposed by minimizing the mean square tracking error along typical trajectories.The Fibonacci algorithm is adopted to develop a fast optimizing procedure,which overcomes the complex calculation of multi-objective optimization.The results of simulation verify remarkable effects of the proposed control strategy and the position tracking error of the joint is reduced largely.An open CNC system framework based upon “IPC+Turbo PMAC” architecture is proposed.The hardware platform of the system is constructed and the experimental prototypes are developed.Online regulation of the motion controller's PID parameters is realized by writing the fuzzy algorithm into the PLC 0 program directly.The results of experiment show that the proposed control strategy can effectively improve the control precision of the joints.The above-mentioned outcomes lay a solid foundation for the continuous and wide application of high-speed parallel manipulators.