Dynamic Decoupling Control Strategy For Six-Degree-of-Freedom Parallel Mechanisms

Parallel mechanism with six degrees of freedom has been widely used in the fields of motion simulation and precise positioning,because of its advantage of high positioning accuracy,stable structure,strong bearing capacity,and good dynamic response performance.The parallel mechanism always works with load and the load will cause dynamic coupling among different degree-of-freedoms of the mechanism,which seriously affects the accuracy of the system.At present,two main difficulties in solving the dynamic coupling problem are as follows.Firstly,the load is usually uncertain and its dynamic parameters are difficult to know a priori,which makes some model-based control strategies cannot be applied.Secondly,the parallel mechanism with load is a multi-input multi-output system in the physical space,and each channel of system cannot be controlled independently.For the above-mentioned problems,this thesis takes 6-DOF parallel mechanism with uncertain load as a research subject,and the main research contents are as follows:1.By using the vector method and Newton's Euler method,the kinematic analysis and dynamic analysis of the mechanism are presented.And the three basic equations of the valve-controlled hydraulic cylinder was used to design the hydraulic drive system of a single leg.Secondly,the hydraulically driven 6-DOF parallel mechanism with uncertain load is built in MATLAB/Simulink,and a corresponding simulation system was designed to verify the correctness of the modeling.2.The dynamic coupling of the system was analyzed,and its dynamic coupling model was built.In order to solve the problem that the load dynamics parameters are difficult to know a priori,a load dynamics parameter estimation method based on extended Kalman filter algorithm is proposed.This algorithm is used to identify the load centroid height and the main moment of inertia.The simulation was carried out in MATLAB /Simulink to verify the identification effect of the proposed algorithm.3.The control effect of the PD control strategy in the joint space is simulated and analyzed,and the dynamic coupling characteristics among the degrees of the system and the non-independence among the control channels are observed.Based on the known load dynamics parameters,a modal space computed torque control method based on coupling model is proposed.The control performance is simulated and verified in MATLAB /Simulink,and the simulation results were compared with the joint space PD control strategy.4.For situations where the load model is complex and parameter identification is difficult,a modal space neural network compensation control strategy is proposed.The controller is designed in the modal space and uses the neural network to compensate the load force in real time.The control strategy can eliminate the need to have priori knowledge about the uncertain load and can guarantee the independence of each control channel.The control performance of the modal space neural network control strategy is simulated and verified,and the simulation results are compared with the simulation results of the joint space PD control strategy.The trajectory tracking ability and dynamic coupling suppression ability are summarized.