With the development of modern industry and the progress of science and technology,robot technology has penetrated into many fields of society.But the traditional single manipulator is restricted by the environment.With the increasing demand for industry,medical and service,People hope that robots can perform more complex and fine works like humans.Therefore,the study for dual manipulator system technology is needed.ROS(Robot Operation System)has a more convenient communication mechanism and a more friendly development environment compared with many traditional robot development systems.Therefore,this paper is based on ROS to build an experimental platform for dual manipulator system and the dual manipulator collision detection and motion planning methods are also studied.First of all,the kinematic of UR5 manipulator is modeled in this paper.Two methods,numerical method and analytical method,are used for performance comparison analysis.The singularity of UR5 is also analyzed.Secondly,based on ROS,the simulation platform of dual manipulator system is built and displayed in Rviz.What's more,the connection between the manipulator and the upper computer is established,the message type of the communication between ROS nodes is selected,the ROS main program library Move It!,which can realize the motion planning of the manipulator is also configured.Based on Move It!,the motion planning experiment is carried out.Then consider the possible collision between two manipulators when they move,in this paper,the collision detection algorithm is studied.Firstly,study the AABB and OBB algorithms in the hierarchical bounding box method.Considering the specific space structure of the dual manipulator system,the ball capsule bounding box is used to simplify the model of the dual manipulator system,and the collision detection method is proposed.The collision avoidance of dual manipulator system based on open source collision detection library FCL is realized.Next,the motion planning algorithm of the manipulator is studied,and the PRM algorithm based on sampling principle is studied and verified by simulation.Considering some unnecessary collision detection costs in the PRM algorithm,the improved PRM algorithm,Lazy PRM algorithm,is used to improve the planning efficiency by using the method of delay collision detection.The performance of the two algorithms is compared and analyzed based on the open source motion planning library OMPL.Dual manipulator or multi manipulator cooperative robot are creating a new model of intelligent network manufacturing,realizing perfect coordination,coexistence and sharing of machines and people,forming an era of sharing economy. |