Research On Mapping Algorithm Based On Laser 2D-SLAM

When an intelligent mobile robot moves in an unknown environment,it needs to sense the surrounding environment and determine its position in the environment.Therefore,it must have the ability to build an environmental map and autonomous localization.Simultaneous Localization and Mapping(SLAM)is a key technology used to solve the problem of robots building environmental maps and autonomous localization in unknown environments.It plays an important role in robots' autonomous navigation.SLAM is a research boom in the field of robotics in recent years,and many different solutions have emerged.In this paper,the ROS robot system is used as a platform to research the laser SLAM problem based on the scanning matching method of the indoor unknown environment.The main work is as follows:(1)The LM(Levenberg-Marquardt)method is used to improve scan matching of the laser SLAM and reduce localization error.Firstly,based on the map model used in this paper,the scan matching method based on Occupancy Grid Map is derived.For the phenomenon that the local approximation is not accurate enough and may cause nonconvergence when using the Gauss-Newton method to solve scan matching,this paper uses the LM method to alleviate this problem.Finally verifying the effectiveness of the LM method through experiments.(2)An improved SDF(Signed Distance Function)map is proposed to optimize the limit of map accuracy on Occupancy Grid Map.By studying the traditional HectorSLAM algorithm and 2D-SDF-SLAM algorithm,this paper proposes a WSDF(weighted SDF)map.WSDF map has higher accuracy,and the idea of weighting is used to make the map more smooth and effective.Finally,combined with the LM optimization method,an improved SLAM method based on WSDF map is proposed.(3)Set up the simulation experiment environment used in this paper,including ROS robot system,Gazebo simulation software and Rviz visualization software.Experiments are carried out through ROS robot system and the above softwares,and then the improved SLAM method and the traditional Hector-SLAM method are compared and analyzed.The experimental results that the improved SLAM method has better location and mapping effects.