Research On AGV Navigation Based On Lidar Positioning And Visual Redundancy Confirmation

Automated Guided Vehicle(AGV),as a key equipment in the industrial logistics automation system,is widely used in the automatic handling of goods.This dissertation takes the fork-type AGV as the research object,and studies the positioning algorithm based on lidar and the visual redundancy confirmation algorithm based on artificial road signs.The main research contents are as follows:1)According to the application scenarios and working environment of fork-type AGV,the overall structure of fork-type AGV and its navigation scheme are analyzed.The kinematics model of AGV is established,the kinematics equations are given,and the various modules and functions of the navigation system are analyzed.2)On the basis of the given AGV control architecture,study the SLAM(Simultaneous Localization and Mapping)navigation algorithm based on lidar.First,the mathematical model of the odometer and lidar is established,and the laser SLAM scheme is given.Aiming at the cumulative error and stability problem in the solution of the point cloud matching algorithm,an improved point cloud matching algorithm based on the LM(Levenberg-Marquardt)algorithm is proposed,and finally loop detection is used to achieve global optimization,so as to realize the positioning of AGV.3)In order to distinguish multiple machines with the same laser point cloud characteristics in the factory environment,the visual redundancy confirmation algorithm based on artificial road signs is applied to ensure that the AGV actually arrives at the designated machine.A specific artificial road sign is designed for visual confirmation for each machine,and the artificial road sign visual recognition algorithm is studied and proposed based on logarithmic polar coordinate transformation,which is used for redundant confirmation of having arrived at the specific machines based on lidar guidance,and for improving the reliability of laser navigation.4)The proposed lidar positioning algorithm was implemented by program,and the AGV laser positioning and artificial road sign recognition experimental platform was built,and the proposed lidar positioning algorithm and visual redundancy confirmation algorithm were experimentally verified.Through collecting and analyzing experimental data,the correctness and stability of the proposed laser positioning algorithm and road sign recognition algorithm are verified.