Research On Simultaneous Location And Mapping Method Based On Lidar And IMU

Autonomous driving and mobile robot related technology research is a popular research direction in recent years.The ability of simultaneous localization and mapping is an important prerequisite for the realization of unmanned system.Because of its high accuracy and all-weather operation,lidar is widely used as the main sensor in SLAM technology.The combination of lidar and IMU is an effective method to improve the overall accuracy.In this dissertation,multi-line lidar is used as the main data acquisition sensor,and the data provided by IMU are integrated to study robot positioning and environment modeling.In this dissertation,the principle of lidar and IMU sensors is introduced,and the data processing method of point cloud distortion caused by lidar motion is studied.The matching principle of two frames of point cloud data is described in detail by combining the current mainstream point surface matching method.By designing a lidar point cloud data acquisition system,The data collected by lidar and IMU sensor are saved,and the format of data preservation is introduced.To solve the problem of insufficient information fusion and utilization between lidar and IMU sensors,an optimization method using factor graph tightly coupled lidar and IMU is proposed in this dissertation.On the basis of point-surface matching,the point-surface matching residual is used as the factor of lidar,and the IMU pre-integration residual is used as the IMU factor.The factor graph is used to fuse the two factors to optimize the effect of the graph construction.The important point cloud frames are passed into the factor graph as key frames to optimize the factor graph to improve the optimization efficiency of the factor graph.In this dissertation,the sliding window is used to limit the number of optimized frames in the factor graph,and the method of marginalization should be used to retain the constraint relation of the old frames removed from the sliding window,so as to ensure that the optimization speed is improved without reducing the optimization accuracy.The results show that compared with other algorithms,the graph building method using factor graph optimization has better effect and less error.After using sliding window optimization,the speed is improved by 12%,which is an important basis for realizing the system unmanned.Use in improvement based on factor graph optimization to build graph method,for the loopback map,merge Scan Context loopback detection method to enhance the building figure precision,when detected loopback optimize the position of the laser radar,the pose as a loopback detection factor into the design in this dissertation combines the IMU with laser radar factor of factor diagram,As the third factor,the optimization effect of factor graph is improved.The experiment shows that the mapping accuracy is improved and the matching speed between two frames is decreased,but the real-time positioning and mapping are not affected,which can meet the needs of the actual application of real-time positioning and mapping.