Research On Motion Planning And Tracking Control For Manipulators To Avoid Obstacles In Dynamic Environments

Manipulators are widely used in the field of human social production.Now,the level of automation and intelligence of manipulators are poor,necessitating manual participation accomplish essential tasks while working,limiting the application of manipulators on the one hand and reducing the safety of humanmachine collaboration on the other hand.Exploring a universal obstacle avoidance approach of smooth dynamic motion planning for manipulators is of considerable benefit in solving the above difficulties.A grid mapping-model was built to simplify the collision detection procedure based on the concept of boundary discretization.A LSLPRM(Local Semi Lazy Probabilistic Roadmap)method was proposed by introducing the stratified sampling method,the local idea and the potential cost into the PRM(Probabilistic Roadmap)method.The LSLPRM method was applied to the dynamic obstacle avoidance of manipulators,to explore a dynamic obstacle avoidance method of path planning for manipulators with strong applicability,high planning efficiency and high quality.In order to improve the motion performance of manipulators in an obstacle avoidance process,A DPSO(Double Population Particle Swarm Optimization)algorithm was proposed by introducing the idea of swarm into the PSO(Particle Swarm Optimization)algorithm,in order to overcome the problem of poor optimization quality of which.The trajectory of obstacle avoidance dynamically constructed by the PIMSMO(Polynomial Interpolation with Multi-Segment and Multi-Order)method was optimized by DPSO,and to explore a dynamic obstacle avoidance method of smooth trajectory planning for manipulators with high execution efficiency and low impact.In order to make the manipulator track the obstacle avoidance curve more accurately,by designing RBFNN(Radial Basis Function Neural Network)and fuzzy controller,a FSMCORBFNN(Fuzzy Sliding Mode Control of RBFNN)was proposed,to explore the trajectory tracking control method of manipulators with low chattering and excellent control effect.The main research work and results of this paper are as follows.(1)Path planning for manipulators to avoid obstacles in dynamic environments have been studied.Aiming at the problem that PRM algorithm is difficult to be used the obstacle avoidance of path planning for manipulators in dynamic environments,which resulting in the sharp decline or even failure of planning efficiency,A LSLPRM algorithm was put forward in this paper by constructing a grid mapping model,using the semi lazy collision-detection method and local idea,combining with the layered sampling method,and studying the dynamic application strategy of PRM method.The grid mappingmodel was constructed to obtain the mapping relationship between task space and joint space,so as to simplify the collision-detection process.By using the semi lazy collision-detection method and local strategy,the path collision-detection strategy and lazy collision-detection strategy of sampling points were adjusted,the times of path collision-detection and single collision-detection were reduced,and the time consumption of the algorithm was also reduced.By constructing the potential energy function between the manipulator’s link and the obstacle,the cost function of the graph-search algorithm was adjusted,and it was explored the application of LSLPRM algorithm in the obstacle avoidance of path planning in dynamic environments for manipulators.(2)Trajectory planning for manipulators to avoid obstacles in dynamic environments have been studied.The cubic spline was used to interpolate the joint position nodes.The optimization objective mathematical model was established.The constraint conditions of the joint were analyzed.The particle of PSO algorithm was given the attributes of discoverer and follower,and its speed attribute is eliminated.On this basis,PSO was improved,and a DPSO algorithm was proposed to improve the population diversity and the efficiency of which.In this paper,the optimization performance of PSO algorithm was improved,and the obstacle avoidance curve with high smoothness and short running time was obtained by using the improved PSO algorithm.In addition,a PIMSMO method was proposed to overcome the continuity problem of generation of trajectory truncation-point in dynamic environments.(3)Trajectory tracking for manipulators to avoid obstacles in dynamic environments have been studied.The RBFNN was used to approximate the unknown nonlinear function of kinetics model for manipulators.By designing a fuzzy controller for the fuzzy constant in the fuzzy switching term,the sliding mode switching gain was adjusted online,trying to reduce the chattering of the system.The FSMCRBFNN law of the system is obtained by adding the above two.On this basis,the Lyapunov stability and finite time convergence of the system were analyzed;In this paper,the purpose of reducing chattering of sliding mode system is realized,and the trajectory tracking of joint obstacle avoidance curve was completed with high control effect.