Research On AGV Positioning And Navigation Strategy For Warehouse Environment

With the continuous promotion of modern logistics industry,intelligent storage has become an indispensable part of modern logistics.Automated Guided Vehicle(AGV),as an important part of intelligent storage,has a direct impact on its operating efficiency due to its location and path tracking performance.Therefore,how to quickly and accurately determine the position of the AGV,and how to accurately track the target path and quickly reach the target point according to its position has become a hot issue in AGV technology research.In this thesis,we take differential steering AGV as the research object,and the positioning method and path tracking strategy of AGV are studied.This thesis establishes a global positioning system in the warehouse environment with the reflector as a priori feature,and establishes the AGV coordinate system based on it.At the same time,the kinematic model and dynamic model of differential steering AGV are derived under the coordinate system.In order to further improve the accuracy and speed of AGV positioning,the prediction equation is established by the odometer model,and the observation equation is established by the original polar coordinate data of the laser reflector,which is linearized by the unscented Kalman filter method,and finally a positioning strategy based on the fusion of multi-source sensor information is realized.This method minimizes the accuracy loss of linearization processing and ensures efficient output of global positioning data.The evaluation function of the DWA algorithm is improved for the traditional Dynamic Window Approach(DWA),which has too many turns and tends to deviate from the path with large curvature turns.At the same time,the external disturbances from side shifts,slips,ground friction changes and internal disturbances caused by inaccurate internal modeling are observed and compensated in real time by the Extended State Observer(ESO).Thus,the path tracking error of the AGV is reduced and the convergence rate of the AGV is accelerated after encountering disturbance.The algorithm proposed in this thesis is verified based on MATLAB and Robot Operating System(ROS)tool named gazebo,a physics experiment platform for the warehouse system is built.Through simulation and experimental comparison,the performance of positioning and path tracking under various working conditions is tested.The result shows that the AGV can achieve fast and accurate positioning,and can track the given path stably and accurately.