Research On Kinematics And Control Of 3-PSS Parallel Mechanism Based On Neural Network And Particle Swarm Optimization

Parallel mechanism is widely used in docking mechanism,vector propulsion,machine tool,astronomy,simulator and so on.But the synthesis,kinematics,dynamics and control of parallel mechanism are more complex.In this paper,the common 3-PSS parallel mechanism is selected as the research object.In the aspect of kinematics,neural network and particle swarm optimization are mainly studied.In the aspect of control,neural network predictive control is studied.Firstly,the kinematics of 3-PSS mechanism is analyzed in detail in this paper,which is the basis of modeling and programming.Using Solid Works to generate data point space,the result of point space based on formula discrete method is verified.Secondly,the paper studies the problem of learning motion data points by neural network,improves the structure of neural network application based on LabVIEW,transplants neural network by LabVIEW,and verifies the ability of solving space kinematics by SolidWorks.Thirdly,this paper uses LabVIEW to program particle swarm optimization,and verifies the feasibility of the program.Based on the program,the numerical solution of 3-PSS kinematics is realized.Compared with the original trajectory in Solid Works,this method can achieve a certain precision.In addition,the similarities and differences between neural network method and particle swarm optimization method,application occasions and combination methods are discussed.Finally,this paper analyzes the control performance based on the Simulink neural network predictive control toolbox.The neural network prediction part is improved based on LabVIEW.The results show that the improved neural network can achieve high precision rolling learning and rolling prediction.