Font Size: a A A

Research On Structure Improvement And Visual Recognition Control Of Multi-unit Inspection Robot

Posted on:2021-07-06Degree:MasterType:Thesis
Country:ChinaCandidate:K GaoFull Text:PDF
GTID:2492306353453734Subject:Mechanical Manufacturing and Automation
Abstract/Summary:PDF Full Text Request
Robots play an increasingly important role in social life and production.High-voltage transmission lines need to be inspected regularly,and the environment of high-voltage lines is very complex.Therefore,the application design of transmission line inspection robot has great significance to replace human inspection and ensure the efficiency of inspection.Patrol robots need to cross power fittings and inspect the line when working on the line.They need the coordination of mechanical structure,sensor system and control method to complete the work.This paper improves the mechanical structure of the existing multi-unit line-patrolling robot,designs the visual recognition and control system of power fittings,and carries out the experimental verification of the prototype to ensure that the line-patrolling robot can complete the basic line-patrolling work.The research work of this paper is as follows.(1)Improve the unreasonable mechanical structure of the robot and design a new walking clamping mechanism.Through the analysis of the mechanical structure of the existing inspection robot,the unreasonable structure which causes the problems of walking on line and obstacle surmounting is found out.The arm structure and parallelogram structure of the robot are improved,and the rationality of the quadrilateral structure is analyzed.A partitioned traveling and clamping mechanism is designed.The robot’s one-sided traveling mechanism is changed into two-sided traveling mechanism.The improved traveling robot is modeled and simulated in three dimensions.(2)The obstacle recognition of transmission line is carried out,and the visual recognition of obstacle is needed by analyzing the environment of transmission line and the work of robot.Firstly,obstacle recognition is compared with geometric feature based method and classifier based method,and SVM classifier method is adopted to realize obstacle recognition.The HOG features of the main power fittings are extracted,and the SVM method is used to train the classifier.The recognition of the shock hammer,suspension insulator strings and other fittings is completed.At the same time,the steering obstacle identification of the robot is carried out by putting forward the experimental method of gray value ratio.i-solution problem of stiffness in the workspace is also analyzed.(3)The control system of the improved line-patrolling robot is designed,and the mathematical model of the mechanical transmission part of the robot is established.The action of the robot crossing different fittings and the execution state of the corresponding motor are planned,and the obstacle-crossing control based on behavior tree is proposed to complete the control program of the robot motion.(4)The experimental platform was built by assembling the new walking clamping mechanism and the robot,and the obstacle surmounting experiment of the robot was carried out.The problem of modifying the control program through the prototype experiment and choosing reasonable parameters to complete the experiment.
Keywords/Search Tags:Multi-unit line-patrolling robot, split-and-tie walking and clamping mechanism, visual recognition, motion control, obstacle-crossing
PDF Full Text Request
Related items