| Live working is used to ensure the reliability of the power supply,but it is very dangerous for the operator.Therefore,the live working robot was bom to replace the operator.In recent years,research on live working robots has made great progress.And even live working robots for fastening transmission line bolts have been developed.However,loading/unloading lead-in wire of high voltage disconnector still requires the operator to complete by hand.So it is necessary to develop a live working robot and a robotic remote control system for loading/unloading lead-in wire of high voltage disconnector.In order to aceomplish remote eontrolling of a eomplex loading/unloading robot which has 22 servo motors,2 cylinders,a demolition device and a torque wrench,a remote control system is designed in this paper.It can communicate with the loading/unloading robot,control the movement of the robot.It can real-time display of the robot status,real-time playback of the video,identify nuts and spatial positioning between nut and demolition device as well.The main works of this thesis are as follows:1.The communication function and the remote control robot function of the remote control system are implemented.The programming of the Modbus/TCP server address table is completed on the basis of easy access to the client.The Modbus/TCP client and the robot motion control program is completed.When the remote control system software button is pressed or the joystick is pushed,the Modbus/TCP client is asked to send a control command to the motor to remotely control the motion of the robot.The remote control system can control the robot to complete the actions such as moving the lifting platform,establishing the equipotential,hanging the robot,pushing the robot to move,clamping the line,aligning the nut,breaking the nut,pushing the bolt,aligning the hole,tightening the nut and so on.The RS485 serial communication between the remote control system and the power supply battery was also implemented.2.The identification function of the nut based on deep learning and the automatic centering function between the nut and the demolition deviee are implemented.Based on Darknet,the training of the YOLOv3 pre-training model is completed.The nuts are identified based on the OpenCV and YOLOv3 models.Based on the nut prediction results and the monocular vision ranging technology,the rough positioning between the nut and the demolition device is implemented,and the automatic centering between the nut and the demolition device is implemented.3.Real-time display of robot status and operating environment is comleted.Based on multi-threading technology,the remote control system reads the robot status data packet,parses the message,and displays these things such as the battery p ower,current wind speed,and robot position,speed,torque,and alarm in real time,And real-time playback of 8 cameras in the work environment has also been implemented.Through the system test between the remote control system and the robot,all functions of the remote control system inclucling real-time transmission a nd display of robot status data,real-time playback of video,control of robot motion through remote control system interface buttons or joysticks have b een tested.The test results show that the wireless communication distance of the remote control system can be up to 20m,while the video transmission delay is less than 0.5s,and the response of the robot to the control command is less than 0.1s.The remote control system has complete functions and shows good performances in real-time transmission.The rough positioning between the nut and the demolition device which is used to implement automatic centering meets the automatic control requirements,indicating that the ground control system has a certain degree of intelligence. |
PDF Full Text Request