Research On Autonomous Navigation System Of Substation Inspection Robot

Substations shall regularly repair electrical equipment to ensure safe operation.With the development of electrical system,the traditional text analysis is quickly replaced by robot survey.Private tourism is the main technology of robot exploration.Comments on the introduction of self statements in this document,the research background of selfexploration existing algorithms for substation location accuracy is not high and the path planning,planning problem with long,flowing light and backlight,on the basis of the combination of localization algorithm,improved and extended algorithm improve the search plan according to the experimental platform in the field of navigation system is put forward by the temporary itself,experimental verification algorithm.The main work contents are as follows:1.This paper describes the research status of autonomous navigation of substation inspection robot at home and abroad,and introduces the robot motion model,camera model,IMU model,camera distortion processing and IMU preintegration theory.2.During the positioning of the robot,the VINS-Mono system only uses a single optical flow tracking to process data in the front vision,which will lead to corner points with large errors and inaccurate positioning accuracy.In this paper,forward optical flow tracking and reverse optical flow tracking are combined to improve the positioning accuracy.Aiming at the problem that the marginal residual information in the objective optimization function runs slowly in the back-end nonlinear optimization of the VINSMono system,when the information is marginalized,the parts except the camera pose are first marginalized,and then the camera pose is marginalized.A new information matrix is constructed by grouping to rise the marginalization speed.The simulation results show that the positioning accuracy and operation speed of the algorithm are improved.3.In terms of path planning,aiming at the problem that A~* algorithm has many inflection points and the planned path length in the global path planning,the extended domain search method of 7×7 is firstly used to search the path,and then the path points are classified according to the found paths,and then the discriminant is used to screen out the qualified paths.Finally,the third order Bessel curve is used to smooth the path.In this paper,the A~* algorithm before and after the improvement is compared and tested in different environments.The experimental results show that the A~* algorithm before and after the improvement significantly shortens the path length and reduces the number of path inflection points.4.Build inspection robot autonomous navigation hardware circuit,software design based on ROS operation,structures,inspection robot experiment platform,the inspection robot in a real environment using the experiment platform for the improvement of algorithm performance testing,test results show that the improved algorithm is proposed in this paper has good localization and path planning effect,The autonomous navigation function of the robot is realized.Figure [58] Table [10] Reference [81]...