Research And Practice On Indoor Mobile Robot Localization And Navigation Based On ROS

Mobile robot localization and navigation technology is an important research di-rection in the field of robotics,which endows robots with the ability to move autono-mously in the environment.Mobile robot localization and navigation technology de-veloped based on ROS（Robot Operating System）can achieve cross-platform reuse on different robot platforms,providing more possibilities and practicality for the localiza-tion and navigation applications of mobile robots.The implementation of mobile robot localization and navigation requires the robot to determine its position in the environ-ment and plan a path to a specified target location in that environment.Designing a more accurate and efficient localization and navigation system based on the actual ap-plication scenario is the key issue that needs to be addressed in the application of mobile robots.This thesis takes the indoor mobile robot localization and navigation algorithm based on ROS as the research direction,aiming to improve the indoor localization ac-curacy of robots and reduce navigation costs.A method of obtaining the pose of the mobile robot based on ORB＿SLAM3（A simultaneous localization and mapping algo-rithm based on extracting image feature points）and a mobile robot navigation solution based on indoor environment localization are proposed,and experimental verification is conducted.The specific research contents of this thesis are as follows:1.A method of obtaining the robot’s pose based on ORB＿SLAM3 is proposed.This method improves ORB＿SLAM3 algorithm,using RGBD camera as information collection sensor,publishes ROS topic communication protocol-compliant robot posi-tion information during the movement of ROS mobile robot.Experiments show that the reference values obtained by this algorithm are more accurate than those obtained by wheel odometer,wheel odometer and IMU fusion,and can effectively improve the positioning accuracy of robot in unknown environment.2.A method of correcting the wheel odometer error based on ORB＿SLAM3 is proposed.The ratio of the robot displacement reference value obtained based on ORB＿SLAM3 to that obtained based on the wheel odometer is used as the linear ve-locity correction parameter,and the robot’s linear velocity is adjusted using this param-eter.The corrected linear velocity is then substituted into the odometer measurement model to correct the wheel odometer displacement error.Experiments show that this method can reduce the measurement error of the corrected wheel odometer and improve the localization accuracy of the odometer measurement model.3.A safe path planning algorithm based on key points is proposed,named A^plusalgorithm.This algorithm introduces the radius of the circumscribed circle outside the robot’s occupied area as the minimum safe distance between the robot and obstacles,improves the path search conditions of the A*algorithm,and extracts the key points of the path.Finally,a key-points path that can maintain a safe distance from obstacles is planned on the 2D grid map.Experiments show that,on the basis of ensuring the path security,the algorithm has the advantages of shorter path length and fewer route inflec-tion points compared with A*algorithm,which can effectively improve the practica-bility of the route planning algorithm.4.A motion control algorithm based on key points is proposed.Based on the key-points path planned by A^plus algorithm,this algorithm uses multi-sensor fusion locali-zation to calculate the pose difference between the robot and key points,and then pub-lishes motion control commands according to the pose difference to control the robot to move along the key-points path to avoid obstacles and reach the path endpoint.The algorithm has been shown to be more efficient than existing fixed-point navigation al-gorithms in indoor static environments,with fewer adjustments to orientation angle and speed calculations and shorter navigation times.