Study On Autonomous Navigation Technology Of Mobile Robot Based On Lidar And Binocular Stereo Vision

Nowadays,more and more mobile robots have been widely used in aerospace,emergency rescue,warehousing logistics and other fields.In practical applications,how the mobile robots completes the pose estimation,the construction of environmental maps and the autonomous navigation on this foundation have become the key technologies.Therefore,the technical problem of simultaneous localization and mapping(SLAM)emerged as the times require.The common solutions of slam are vision slam and lidar slam,in recent years,the fusion of vision and lidar sensors to solve the SLAM problem has become a research hotspot,which can solve the problems existing in practical applications,therefore this paper takes the odometer module of the visual SLAM algorithm as the research object,and then combines the improved visual odometer with the lidar SLAM to build the map,and finally realizes the autonomous navigation of the mobile robot.The main research contents are as follows:Firstly,in view of the problem of the feature distributing uneven of odometer module in visual slam,the feature points extracted from the image can be evenly distributed by improving the division depth of quadtree nodes.A comparative experiment on the algorithm show that the improved algorithm increase the uniformity of the feature distributing and positioning accuracy with effect.Secondly,in view of the problem of the traditional Gmapping-SLAM algorithm is prone to exist the discontinuity when constructing environment map,the visual odometer is introduced and the observation model of lidar and vision fusion is proposed.Experiments proves that the improved Gmapping-SLAM algorithm can solve the problem of discontinuous and incomplete mapping caused by the mobile robot slipping.It is verified through experiments that the improved Gmapping-SLAM algorithm can solve the problem of discontinuous and incomplete mapping caused by the mobile robot slipping.Thirdly,research on the path planning algorithm of mobile robot,the heuristic function of the global path planning algorithm—A-star algorithm is improved,and several different heuristic functions are compared and analyzed through experiments,and finally the navigation control accuracy and local The effectiveness of the path planning algorithm for obstacle avoidance has been verified.Lastly,the autonomous navigation system of mobile robot is built,the position accuracy of visual odometer is verified by trajectory accuracy experiment.the navigation control accuracy of mobile robot and the effectiveness of obstacle avoidance of the local path planning algorithm are verified.This paper uses lidar and stereo vision to realize the autonomous navigation system of mobile robots,and conducts detailed research on the three parts of the system:positioning,mapping and path planning.A set of autonomous navigation system for mobile robots with good safety and high precision is obtained.