Research And Implementation Of Simultaneous Localization And Indoor Mapping And Path Planning Based On Lidar

In recent years,mobile robot navigation has been applied to unmanned vehicles(UGVs),unmanned ships(USVs),and unmanned aerial vehicles(UAVs).Of the many functions required for robot navigation,the three most important components are positioning,maps,and path planning.This topic takes the background of indoor mobile robot navigation as the research and implementation of SLAM and path planning alg orithm based on single-line laser radar.Firstly,the commonly used coordinate system in SLAM and the mutual transformation between coordinate systems are introduced.The transformation mainly includes the conversion of the observation data from laser radar into the world coordinate system for the construction of maps.Next,the odometry movement model and the sensor(lidar)observation model are introduced.Finally,the two-dimensional map model is compared and analyzed,and combined with the path planning algorithm,the grid map is selected as the map expression model of the subject,and the grid map construction algor ithm is deduced theoretically.Aiming at the simultaneous positioning and map construc tion technology of indoor mobile robots,two probabilistic SLAM algorithms,EKF_SLAM and Fast SLAM,are introduced respectively.Theoretical derivation and simulation experiments of the above two SLAM algorithms based on probability are presented.Then SLAM based on scan matching is introduced.The algorithm performs theoretical analysis and simulation experiments on the Hector_SLAM algorithm using a high scanning frequency laser radar.The three SLAM algorithms are compared and analyzed from the computational complexity of the algorithm and the simulation results.For the problem of path planning of mobile robots based on known maps,this topic uses A* algorithm as the global path planning algorithm and DWA algorithm as the local path planning algorithm.The advantages and disadvantages of the Dijsktra algorithm,the BFS algorithm,and the A* algorithm are conceptually compared.Because the A* algorithm has both the advantage s of the shortest path of the Dijsktra algorithm and the advantage of the fastest programming speed of the BFS algorithm,the subject has decided to use the A* algorithm.As a global path planning algorithm.Then through modeling the two-wheeled differential mobile robot,the specific flow of DWA local path planning algorithm is introduced.In order to verify the effectiveness of the above algorithm,a Turtlebot2 robot(Kuboki chassis)was used as an experimental platform to carry out experimental analysis in an indoor environment.Firstly,the indoor same environment is constructed by Gmapping and Hector_SLAM.The accuracy of the two SLAM algorithms is compared and analyzed.The results show that the Hector_SLAM algorithm based on the high scan frequency Li DAR has better construction accuracy.Then based on the two-dimensional environment map constructed by Hector_SLAM,the path planning experiment was conducted,and the validity of the path planning algorithm was verified through autonomous navigation to the s pecified target point.