Design And Implementation Of Aided-Navigation Module For Inspection Robot Based On Multi-sensor Fusion

In recent years,as an intelligent solution to replace the traditional artificial inspection,intelligent inspection robot has been installed in hundreds of substations.Compared with the traditional artificial inspection method,intelligent inspection robot in substation reduces the work intensity of technical personnel,reduces the risk and labor cost of manual work in highrisk environment of substation,and solves the problems which the inspection results related to inspection personnel's technical level.The premise of substation inspection robot to complete inspection work is accurate navigation.At present,the mainstream substation intelligent inspection robot carries high-precision lidar sensors for navigation.This method has high navigation accuracy,but it can not identify the types of obstacles,which seriously affects the inspection performance and efficiency of inspection robot.Therefore,this paper designs a aided-navigation module based on multi-sensor fusion for intelligent inspection robot in substation.In this module,the information of lidar sensor,visible light sensor and inertial measurement unit sensor is fused,which makes the inspection robot realize the autonomous understanding of the environment and improve the navigation accuracy of inspection robot.The main work of this paper is summarized as follows:1.Requirement analysis of navigation assistant module for inspection robot in substation.Combined with the real environment of the substation,the working environment of the inspection robot in the substation is analyzed.At the same time,combined with the shortcomings of the traditional navigation mode,the necessity and importance of the navigation of the inspection robot in the substation are determined,and finally the design requirements of the aided-navigation module are determined;2.Design of multi-sensor fusion algorithm for aided-navigation module.In the algorithm part,the multi-sensor asynchronous information matching algorithm is proposed for the time asynchronous multi-sensor information,the 2D lidar 3D imaging method is proposed for the construction of 3D point cloud map using 2D lidar,the color point cloud map construction is proposed for the construction of color 3D point cloud map,a segmentation method of color point cloud map based on ENet is proposed for the construction 3D point cloud map with environment semantic information;3.Completed the software and hardware implementation of the aided-navigation module.In the aspect of software,the algorithm design of aided-navigation module is realized with the help of ROS system.In the aspect of hardware,the selection of key components and the determination of installation position of aided-navigation module are completed through experiments.Through the test in the real environment of the substation,the test shows that the inspection robot in the substation can effectively understand the environment by using the information provided by the aided-navigation module,which enriches the function of the inspection robot in the substation and improves the intelligent degree of the inspection robot.