Research On Human-Dual Arm Robot Cooperative Motion Planning Based On Dynamic System Method

The cooperative operation of the dual-arm robot has many advantages,including larger working space,stronger load capacity,and suit for many applicable scenarios,but it still has challenges when facing with a complex and changing environment.Human-machine collaboration integrates the advantages of human problem-solving flexibility and adaptability to environmental changes,as well as the durability and motion accuracy of robot work.It is suitable for flexible production lines with compact layout and high accuracy,which is in line with the production miniaturization and precision demand.However,the current dual-arm coordinated control method is still not perfect,and the efficient collaboration between humans and dual-arm robots is a major challenge.In this process,how to realize the real-time motion trajectory compensation of both arms according to environmental changes is the core issue,and it is also the basis for completing coordination tasks.To achieve the collaborative operation of robots and humans,and to be sufficiently robust to environmental changes,two key issues need to be solved:(1)designing a coordinated control strategy for a dual-arm robot;(2)realizing real-time robots in a changing environment re-planning and improving the anti-interference ability of the robot to the dynamic environment.In view of the above problems,this paper proposes a two-arm planning strategy based on dynamic systems.The method of motion planning is studied for the coordinated motion process of the same and asynchronous behaviors.Based on the man-machine cooperation model of the dual-arm robot,the cooperative control strategy of the dual-arm robot is studied and simulated.For the independent task and collaborative task of the robot,different test scenarios were designed and their effectiveness was verified on the built experimental platform.The main research work of this article is as follows:1.Build a human-machine collaborative interaction system based on dual-arm robots.The UR5 robotic arm,Robotiq gripper and Opti Track camera are used to build a hardware system.Based on the open source system platform ROS,the system software architecture is established to provide control instructions and raw sensor data for the upper layer.The kinematics model of the UR5 robot is established,which provides a basis for the subsequent construction of a dual-arm workspace model and robot motion control.2.dynamic systems is trained to generate robot movements.The impedance control method is used to realize the compliant drag of the robot,complete the collection of teaching data and train a globally progressively convergent dynamic system.A dynamic system method is used to generate the robot's motion,analyze the impact of the disturbance on the robot's next action and plan a new path for the robot's real-time motion.Verification experiments of the constructed dynamic system in the simulation space proves that the constructed dynamic system has both convergence and traceability.3.The strategy for coordinated control of two arms under a dynamic system is designed.The six joint angles of the robot are equally divided according to the joint limit,and then arranged and combined.The end position set of the robot is used as training data,and the dual-arm working space is modeled by the Gaussian mixture model.The control laws of the real object dynamic system and the virtual object dynamic system are designed.The real object dynamic system is used to control the synchronous and asynchronous behavior of the robot arm.The virtual object dynamic system is used to predict the trajectory of the actual object and obtain the interception of the virtual object in the work space.generating the motion of virtual object.4.The stability and convergence of the two-arm coordinated control strategy based on the dynamic system method proposed are verified.In the built human-robot cooperation operating system,experiments on the cooperation between human and single-arm/dual-arm were designed to achieve the dynamic coordinated motion and target object tracking,while the two arms are collaborating without collision,accurately,smoothly and synchronically.The experimental results prove the effectiveness of the algorithm proposed in this paper,and clarify the possibility of applying the constructed dynamic system theory to actual robots.