| The development of Industry 4.0 has promoted the transformation of assembly manufacturing mode,leading to the emergence of large-scale personalized production.For this task of flexible assembly requirements and huge production scale,it is difficult for both human and robot to independently and effectively complete.In order to overcome this challenge,human-machine collaborative assembly(HRCA),which gives full play to the flexibility of human operation and the automation ability of robots,shows great potential.However,in order to facilitate deployment and operation,current cooperative robots do not fully consider the utilization of shared workspace.Therefore,in the research of human-machine collaborative assembly,this thesis adopts a new type of collaborative robot,Supernumerary Robotic Limbs(SRL),to improve the utilization rate of shared workspace,and completes the research of human-SRL collaborative assembly by solving two problems in collaborative assembly: 1)How to determine the task executor of each step in collaborative assembly;2)How to avoid collision caused by improving the utilization of shared workspace.The main work of this thesis includes:1.Aiming at the problem of determining the task executor in collaborative assembly,this thesis takes work efficiency and task competence as the allocation basis,and human physiological and psychological fatigue as the allocation constraints,and establishes a multi-objective optimization model of collaborative task allocation for human-SRL assembly.The improved NSGA-II is used to encode the parallel gene of task process and allocation result to obtain the non-dominated set,and then evaluate the set according to the evaluation index to select the best result to execute.Finally,the effectiveness of the solution was verified through an actual case of gearbox assembly.2.Aiming at the problem of collision between SRL and operators in collaborative assembly,this thesis simplifies human limbs as cylindrical spatial moving objects and proposes an artificial repulsive field model for cylindrical moving obstacles and an artificial gravitational field model for the desired direction of the target point with path.Kinect captures the posture of the human upper limb and calculates the joint velocity under the upper limb constraint,so as to establish the potential field to generate the motion trajectory of the end of the SRL,and determines the final collision-free joint trajectory of SRL through the inverse kinematics solution and the collision detection between the human upper limb and SRL,which is executed by SRL to complete the assembly task.Finally,experimental verification was conducted on the impact of factors such as the radius of action,spatial position,direction of motion,and directional parameters of obstacles on the generated path3.The establishment of the software and hardware platform of SRL and the experimental verification of the human-SRL collaborative assembly.In order to verify the effectiveness of task planning and safety cooperation and the feasibility of human-SRL collaborative assembly,this thesis built the software and hardware system of SRL.A personalized assembly experiment for the pull type assist mechanism is designed,and the task allocation model is established to obtain the task planning plan.Under the guidance of the task allocation results,the human and SRL use the path generated by the improved artificial potential field to safely complete the collaborative assembly task at a distance of 5cm-20 cm from the person,which verifies the effectiveness of the path planning algorithm and the feasibility of the human-SRL collaborative assembly.The effective verification of the research on human-SRL collaborative assembly in the actual system in this thesis lays a technical foundation for the application of SRL robots in the field of collaborative assembly. |
PDF Full Text Request