3-dof Light Parallel Manipulator Control Technology Research

With the rapid development of economy,the production efficiency and product quality have got great improvement.With the growing demand of people,More requirements are put forward for robot technology.Because low-dof parallel manipulator has many advantages like high rigidity,high precision,they have broad prospects in the food,pharmaceutical and electronics industries and have great research value.This paper take 3-DOF Delta robot for example,focus on control algorithm based on dynamic model and how to realize with hardware equipment.Firstly,this paper introduces structure of delta manipulator and analyzes inverse kinematics and simplified dynamic model.Matlab and Solidworks/Motion are used to verify kinematic accuracy.Simulink and Simmechanics are used to verify dynamic accuracy.Focusing on control strategy based on dynamic model,the controller is designed in Simulink and physical model is designed in Simmechanics.The co-simulation is realized in Simulin/Simmechanics.The control low and steady-state error of computed torque method are analyzed firstly.Aming to uncertain PD parameters and suboptimal tracking effect,the fuzzy decoupling controller is designed to optimize PD parameters and tracking effect.Then the RBF neural network is designed to compensate uncertainty of model and random disturbance of external environment.The law of network weights is developed based on Lyapunov theory and the system stability is guaranteed.Then co-simulation is realized under two conditions.The results show three algorithms can track desired trajectory well without the external interference.However,when interference exist,steady state error of computed torque method increases obviously.The tracking effect of RBF neural network is best but computation amount is large.The tracking effect of fuzzy control is suboptimal and computation amount is relatively small.Last,the hardware and software of control system are designed and the manipulator test platform is built.The control program is completed with TwinCAT3,C++ and Simulink.The trajectory tracking experiment is carried out to verify computed torque method and fuzzy control.The results show fuzzy control.is better.It can reduce errors effectively and realize the precise track of target trajectory.Meanwhile,It shows control system is reliable and effective and can handle heavy computational tasks.It also provides a solution for some complex algorithms of parallel manipulator.