Research On Simultaneous Localization And Mapping Of Mobile Robot Based On Lidar

The mobile robot is a typical product of the intelligent era,which represents the forefront of the development of emerging technologies and has been widely used in various disciplines.Simultaneous localization and mapping(SLAM)is the basic technology for mobile robot to realize its intelligence,and autonomous navigation and path planning is the basis of robot autonomy.For mobile robots'application,this thesis will study the LiDAR based SLAM technology and its implementation.Firstly,the SLAM problem is described.The mobile robot motion model,the odometer model and the sensor observation model are studied,which lays a foundation for the derivation of the SLAM algorithm.At the same time,the expressions of three commonly used environment maps of SLAM are analyzed and compared.For the problem of the SLAM,this thesis first introduces the basic principles of filter-based,scan-based,optimization-based SLAM algorithm,focusing on three typical SLAM solutions based on particle filter,scan-based matching and graph-based optimization.The scheme,while discussing its theories,provides detailed mathematical derivations and working flow.By the means of constructing the environment simulation model in Gazebo,the three main algorithms are simulated and verified,and their advantages and disadvantages and analysis of important parameter that affect the experimental results are analyzed and compared.For the path planning problem of mobile robots,this paper divides it into two parts:the local path planning and the global path planning.With regard to the global path planning problem,the paper introduces the principles of Dijkstra algorithm and A~*algorithm firstly,and uses MATLAB simulation tool to realize the comparison verification and analysis of the algorithm.Using the global path as the basis of the local planning,the TEB algorithm is adopted for the local path planning problem.Then the principle of the local TEB algorithm is explained and the simulation experiment is carried out by using the Stage model.This part realizes the path planning function of the mobile robot navigation system,which provides the basis for the robot's autonomous navigation.In order to verify the effectiveness of the algorithm,a mobile robot platform based on ROS with lidar as the main navigation sensor is built.The architecture and layered modules of the system are analyzed,which helps the implementation of the SLAM and path planning algorithms of the proposed navigation system.The system expands the dynamic obstacle layer based on the static map layer constructed,and builds a complete map that can be updated in real time.It can be used to dynamically avoid obstacles and complete the set goal point navigation tasks.