YuMi Robot Human-machine Collaboration Research Based On Motion-sensing Camera

YuMi robot based on the body feeling camera man-machine collaboration is a combination of information through the body movements and gestures,remote control robot to complete the homework with man-machine collaboration system.Mainly through the establishment of robot arms tasks,remote communication,robot camera collection system,combined with human action recognition based on skeleton tracking algorithm and the gesture recognition algorithm based on MLP,completed his arms collaboration and work target detection,man-machine follow-up,gesture recognition and finish homework to make robot autonomous in isolation.In many dangerous environments,robots are needed to replace people.Due to safety problems such as difficulty in close operation,inestimable consequences are produced.However,the existing man-machine cooperation system,due to the limitations of cooperation accuracy,security and other issues,generally stays in the simulation of remote sensing control,unable to meet social needs.Therefore,this paper proposes a YuMi robot human-machine collaboration system based on a somatosensory camera,with the main contents as follows:(1)Based on YuMi robot arms cooperation of the platform system,according to the workpiece and the target information in the work environment,combining machine vision with industrial robots,robot arms to establish collaboration and image acquisition system.At the same time,Socket communication is used to realize the remote man-machine cooperative follow-up control which can achieve zero contact,zero casualties and ensure the safety of man-machine cooperation.(2)Aiming at the difficulty of remote operation in man-machine cooperation,this paper proposes a method of using somatosensory and?deep learning to obtain human body information,so as to operate the robot system.Based on the motion sensor camera,the skeleton tracking algorithm is used to obtain the human motion information and the node coordinates,so get the human motion recognitionas to achieve the man-machine follow-up control.At the same time,MLP multi-layer perceptron is used for human gesture recognition,the results of the gesture number prediction are sent to the robot,so that the robot can independently obtain the corresponding target information and complete the task.Compared with the traditional control method of cooperative system,the method of motion sensing and gesture recognition proposed in this paper has higher recognition accuracy,simpler and more reliable operation.(3)Aiming at the problems of poor accuracy and low efficiency in the application of man-machine collaboration system,the YuMi robot of man-machine collaboration system based on the somatosensory camera proposed in this paper USES the connection and transmission of data between the server end and the client end to establish multiple robot clients with a working accuracy of 0.02mm,which enables one operator to operate multiple robot clusters.Compared with the traditional man-machine collaboration system,the working accuracy and efficiency are higher.