Lidar Indoor SLAM Method

With the development of artificial intelligence and robotics,various kinds of robots have gradually entered into millions of households and have continuously improved people's living standards.Mobile robots have become the focus of robot research.With the rise of unmanned and service robots,mobile robots have become the focus of current discussions and research.The ability to achieve autonomous positioning and navigation in an unknown environment is the key to implementing an intelligent robot.For the problem of SLAM(positioning and map construction)of mobile robots in indoor unknown environments,this paper uses a four-wheeled mobile robot equipped with a single-line laser radar developed by the laboratory as a platform to deeply study the point cloud ICP matching algorithm and the extended Kalman filter algorithm(EKF)2D SLAM method for indoor robots.Firstly,an indoor mobile robot model is built,including the definition of coordinate system,environment model,environmental landmark dynamic model,sensor model and robot kinematics model.At the same time,the structure of laser radar and the method of obtaining distance measurement data are introduced.Secondly,the processing of point cloud data is introduced.Taking into account the position and pose uncertainty,landmark correlation and system convergence during robot movement,point cloud data is converted into image data,and image processing algorithms are used to perform point cloud data processing.Extraction of line features and corner features,and the working principle of the Kalman filter and the specific process of the EKF-SLAM are highlighted.Thirdly,the SLAM method based on the grid map and the improved ICP algorithm is introduced.The improved ICP algorithm and the inter-frame ICP matching of the lidar point cloud are proposed.Then the process of obtaining the transformation matrix based on the singular value decomposition is introduced.Finally,the global is introduced.The creation and updating of the map,which focuses on the process of matching the current frame with the map.Finally,based on the experimental platform of mobile robots,the method verification was carried out and the current mainstream open-source SLAM algorithm was compared.The experimental results confirmed the effectiveness of the indoor linear landmark extraction and matching algorithm and the EKF-SLAM algorithm.