Research On Stewart Platform Semi-Physical Simulation Control Based On Particle Swarm Optimization

The Stewart platform is a parallel six degree of freedom robot driven by six linear brakes.The parallel robot is widely used in aerospace,education and entertainment simulation industries.Current research on the Stewart platform are mostly pure simulation,and the real-time performance is poor.In addition,the Stewart platform has a high requirement for the stability of the system,so the selection of the parameters of the PID controller is particularly important.In this thesis,the simulation of the xPC platform can improve the real-time performance of the system,and using particle swarm optimization algorithm to optimize PID controller parameters also can improve the stability of the control system.The main contents of this thesis are as follows:(1)The xPC real-time simulation platform of two machine mode was built,and the correctness of the real-time simulation platform was verified by testing the xpcosc model.(2)The kinematics of the Stewart platform was analyzed,the inverse solution algorithm of the leg was determined,and the inverse solution model of Simulink was built according to the inverse solution algorithm.The structure of Stewart platform was simplified,and the model of Stewart platform was built with Sim.Mechanics.The sensors and exciter were installed on the model.(3)The dynamic performance index of PID controller was discussed,and the time and absolute error integral(ITAE)index were used as the target function.The advantages and disadvantages of several existing methods for optimizing PID controllers were compared,and the design of PID controllers based on particle swarm optimization(PSO)was selected.(4)The single leg of Stewart platform was studied.Particle swarm optimization and fuzzy control algorithm were used to solve the parameters of PID controller.The change curves of leg displacement under two sets of parameters were compared.The parameters obtained by the particle swarm optimization were taken as the parameters of the PID controller of the Stewart platform.The data exchange between the Stewart platform and the PID controller was realized through the UDP communication,and according to the control principle of PID controller,the real-time semi-physical simulation system of Stewart platform is built.The expected displacement of the leg and the actual displacement were compared to verify the correctness of the entire system.In this thesis,the xPC real-time simulation platform was set up to avoid the delay of the controller under the Windows system,and improved the real-time performance of the system.Secondly,the Stewart platform was built with SimMechanics,and SimMchanics can be modeled directly in Simulink,and the efficiency of modeling was improved.Then,the maximum overshoot of the controller optimized by the particle swarm optimization(PSO)PID controller was 2%lower than that of the fuzzy control algorithm,at the same time,the time to reach steady state was also reduced by 0.4s.It not only improved the stability of the system,but also improved the speed of the system to steady state.Finally,by comparing the expected displacements and actual displacements of the leg,we can see that the Stewart platform can run at the expected displacements and the whole simulation system can be verified.