Multi Robot Cooperative Map Construction Based On UWB And LIDAR

With the progress of society and the development of science and technology,robots have been widely used in life.If mobile robots want to perform tasks in unfamiliar environments,they must be supported by accurate positioning information and reliable environment maps.In a large and complex environment where GPS signals are scarce and infrastructure is not installed,the existing positioning technology is difficult to achieve accurate positioning of mobile robots.Single robot system has poor performance in range perception,anti-interference performance and work efficiency,and can not efficiently and accurately achieve map construction.To solve the above problems,this paper proposes a method of multi robot cooperative map building based on UWB and lidar to ensure the accurate positioning of mobile robots and the requirements of accurate map building in large and complex environments.Mobile robots can obtain the surrounding environment information through laser sensors to achieve accurate pose estimation.Multiple robots can expand the perception area through mutual cooperation,obtain more abundant sensor information,and achieve accurate map construction.The research mainly includes two aspects: one is robot localization based on laser direction endpoint;Second,multi robot collaborative map building.The main work is as follows:(1)The traditional laser scanning matching method does not always correctly establish the corresponding relationship between point clouds,which is prone to wrong matching points,resulting in poor positioning accuracy.To solve this problem,a laser scanning matching method based on directional endpoints is proposed.Extract the endpoints of a straight line in the environment as feature points,and obtain the relative posture relationship of the mobile robot at adjacent moments through feature matching between the endpoints.The extracted endpoint features may have errors,resulting in incorrect matching between endpoints.To further improve the accuracy of point cloud registration,the straight line direction is used as the endpoint direction to eliminate mismatched points.On this basis,the iterative closest point(ICP)method is used to further optimize the matching results of the direction endpoints to obtain a more accurate posture of the mobile robot.(2)The wheeled odometer can accurately estimate the robot’s posture in a short time,but the error in posture estimation will accumulate over time.Using the distance information of the UWB and the relative position and attitude information of the odometer in a short time,a nonlinear optimization problem is constructed,and the relative position and attitude estimation among multiple machines is solved by combining the rough search with the optimal solution.UWB fluctuates greatly in the non line of sight situation,resulting in large errors in the relative pose estimation results among multiple machines.To solve this problem,a closed-loop consistency check algorithm is proposed to eliminate outliers,which effectively improves the accuracy of relative pose estimation among multiple machines.(3)In order to further improve the accuracy of relative pose estimation among multiple robots,the pose of the mobile robot is used as the vertex to construct UWB distance constraint,odometer relative transformation constraint,relative pose constraint among multiple robots and laser closed-loop constraint,and further optimize the movement trajectory of multiple robots through graph optimization algorithm.In order to improve the efficiency and success rate of the closed-loop detection,the sub-map closed-loop detection method based on AKAZE(Accelerated-KAZE)is adopted in this project,which effectively improves the efficiency of the closed-loop detection and the success rate of point cloud matching.On this basis,combined with laser scanning data,multi-robot cooperation is realized to build environment map.