Research On Laser Radar Based Simultaneous Localization And Mapping Method

Simultaneous Localization and Mapping(SLAM)is the key technology to achieve autonomous localization of mobile robot in unknown environment.The core idea of SLAM is to use the current map which has been already established for robot localization,and then to update the current map with the new pose of mobile robot.The main research contents and achievements of this paper include:1.A local search based CSM algorithm is developed to deal with the SLAM front-end.A 2d-SLAM system is established by the developed algorithm with inputs from a 2d laser.Experiments show that the developed method can build map accurately under indoor environment,and has better performance than Karto-SLAM(a kind of template matching based SLAM method)in real-time.2.A height value matching based CSM algorithm is proposed to deal with the SLAM problem in outdoor environment.The 3d laser is chosen as the perception device and a grid map which stores height value is designed to express the environment.The effectiveness of the proposed algorithm is verified by the off-line experiments with 3d laser data of our school,and the main factors which affect the accuracy of the results are discussed.By comparing with the GPS data,the localization error of the proposed matching algorithm is discussed.3.The application of back-end optimization method is studied to deal with the cumulative error in SLAM.By combining pose adjustment and CSparse toolkit,the sparse storage and automatic optimization of the pose map is achieved,and the accuracy of the map is improved.