Research On Mobile Robot SLAM And Obstacle Avoidance Technology In Indoor Glass Scene

With the in-depth development of artificial intelligence technology,the role of mo-bile robots in various fields of social life has become more and more prominent.Among them,SLAM technology is one of the most important technologies in the field of mobile robot research.The important feature of SLAM technology is ”simultaneous”,that is,po-sitioning and map construction are carried out at the same time.Through this technology,Combined with path planning algorithm,mobile robot can realize autonomous movement.Regardless of whether it is robot positioning or map construction,external sensors need to be used to perceive the surrounding environment in real time.Lidar is widely used in mobile robots due to its high accuracy,fast data acquisition speed and wide scanning range.However,when lidar is scanning transparent objects,it is unable to return correct data information to the robot due to the physical characteristics of light pulses.This will cause errors in the mobile robot's perception of the environment,which will affect the normal operation of the entire robot system.Aiming at the particularity of transparent glass,a glass detection method combining lidar and ultrasonic data is proposed.The core of this method is the distance at which ultrasonic waves can get the correct glass feedback,while lidar cannot.And the wrong distance obtained by lidar is much larger than the correct distance of ultrasonic,so the glass is detected based on the difference between the two.This method can only detect static glass.For this reason,a dynamic glass detection method based on neural network(mainly referring to automatic glass doors)is used.This method can accurately obtain the position and contour information of the glass.Through the research of these two detection methods,It laid the foundation for the subsequent mapping and autonomous movement of mobile robots.In order to realize the map construction of mobile robots in the glass environment,three commonly used LIDAR SLAM algorithms were analyzed and compared.According to the simulation results,the Gmapping algorithm is selected as the basic SLAM algorithm,On the basis of the glass detection algorithm,a method of replacing the wrong data of the lidar is proposed.Thereby improving the Gmapping-SLAM algorithm,so that it can build the map correctly in the glass scene.Considering the obstacle avoidance requirements of mobile robots in indoor envi-ronments,this paper combines the advantages of artificial potential field method and dy-namic window method,and uses dynamic window artificial potential field method as the algorithm of local path planning.On this basis,combined with the global path planning algorithm to form a hybrid path planning,realize the autonomous movement of the mobile machine from the starting point to the target point,Corresponding simulation experiments were done on the Gazebo simulation platform to verify the effectiveness of the algorithm.Finally,this topic is based on the Robot Operating System(ROS),The Turtlebot2 robot is equipped with various sensors such as lidar,ultrasonic,and vision to build a mobile robot system platform.Relevant experimental verifications are carried out in real glass scenes.The experimental results show that the mobile robot system can normally compose pictures,avoid obstacles and move autonomously in indoor scenes containing glass.