Research On Dynamic Model Based Control Method Of High-speed Parallel Manipulators

Under the support of National Natural Science Foundation of China,"Integrated optimized design method the high-speed parallel mechanism based on the performance parameters of the motor",this paper takes the three degrees of freedom parallel manipulator Delta for example,and takes the aspects of rigid body dynamics modeling,trajectory planning and motion control strategy into concern.Systematically studied the control method based on dynamic model to improve the operating speed and accuracy.And furtherly proceeded experiments to verify the theory above.The main results is as follows.First,in terms of rigid body dynamic model,established a complete rigid body dynamic model based on the principle of virtual work and came up with a matrix form of expression.Then conducted a deep study on simplify method for the dynamic model,analyzed the robot trajectory parameters and structural parameters’ effect on accuracy of simplified model,took the approximate extent of the torque calculated by simplified model and complete model as the optimization index,presented a new dynamics model simplify method based on combined mass distribution coefficient,which laid the foundation for the realization of the real-time dynamics control methods.Secondly,in terms of trajectory planning,aiming for the reduction of harshness,a new trajectory planning method in joint space is proposed.This method firstly mapped the parameters in operating space to joint space by inverse kinematics model,and then planed the trajectory based on the quintic non-uniform rational B-spline motion law in joint space,through the comparison with the original polynomial trajectory planning method in operating space,it is indicated that the new method effectively improve the precision and smoothness of the robot under high-speed operation.Then,as for the motion control aspect,analyzed the motor driven single branch chain,built the PID control model based on MATLAB,presented a self-tunning method based on genetic algorithm and tuned up the PID parameters.A feedforward controller based on the simplified rigid body dynamics model was designed,built the model of control and mechanic system in Simulink/SimMechanics,made a dynamic simulation for single branch in MATLAB,whose result showed that the torque feedforward control method has better dynamic response characteristics compared to closed-loop kinematic control,while this method can effectively reduce the tracking error caused by dynamic characteristics effects of the robot in high speed operation,improve operational accuracy.Finally,in terms of experiments,TwinCAT real-time control system was used to build an control system experimental platform,relevant control algorithms was designed and software was designed to verify the above-mentioned methods and theories.On the one hand the validity of the above simplified dynamics model and trajectory planning was verified,on the other it was confirmed that single branch chain torque feedforward control method can significantly improve the actual high-speed movement tracking accuracy and dynamic characteristics of parallel manipulators.