Research On AGV Positioning Technology Based On Lidar

In recent years,the freight volume of the e-commerce shopping festival has been increasing,which has made the logistics industry develop rapidly.Intelligent logistics and unmanned logistics will be the future development direction of the logistics industry.As an important tool in the warehousing and logistics industry,AGV has experienced many technical iterations from traditional positioning using auxiliary devices such as twodimensional codes,magnetic stripes,and reflective panels to using AGV-carried sensors to achieve autonomous positioning.Among them,the lidar-based AGV is the most widely used form,but in actual use,the lidar-based AGV positioning still has problems such as global relocation failure and poor positioning accuracy.This paper proposes solutions to the problems of global relocation failure and poor positioning accuracy of lidar AGV in actual use,which improves the positioning ability of AGV in the warehouse environment and makes warehouse AGV positioning more efficient and intelligent.The main research contents of this article are as follows:Aiming at the problem of global relocation failure when the AGV system is restarted or the connection to the server is interrupted,the environmental map construction method and lidar data preprocessing technology are studied.The real-time scanning data of the lidar is used to perform image matching with the existing map to construct line feature matching.Algorithm,completed the engineering implementation and positioning test of the algorithm.For the problem of poor AGV positioning accuracy,the existing particle filter positioning algorithm and scan matching algorithm principles are analyzed,and an improved particle filter positioning algorithm that combines scan matching methods is proposed.The coarse global pose is determined by the particle filter positioning algorithm,which is used as the initial value of the iterative closest point algorithm,and the accurate global pose is output after multiple iterations of optimization,and the accuracy test is completed in the warehouse simulation environment.Aiming at the localization of artificial road signs on local maps,the existing manual road sign positioning method is improved,the space structure of artificial road signs is optimized,the number of road signs required for positioning is reduced,and the positioning algorithm is adapted to improve the artificial road signs.The positioning is completed by simultaneously constructing two types of positioning information,vector and center point,on a single road sign,and the algorithm test is completed in the simulation system.In the Gazebo simulation environment and ROS robot operating system,a simulation warehouse was built to simulate the real environment.And tested the above three algorithms in this environment,the simulation experiment results have reached the expected goal,verifying the effectiveness of the solution in this paper.