Research And Design Of Trackless AGV System Based On Global Vision

Automated Guided Vehicle(AGV),as a new type of intelligent logistics transport equipment,has been widely used in many industries such as warehousing,manufacturing and service industries.With the continuous development of automation technology and computer vision,AGV operating environment,positioning and tracking technology and guidance control based on vision sensor are the current research hotspots of AGV.This thesis combines the current research hotspots and considers the general use of surveillance cameras in the actual operating environment.Taking the three-wheel omnidirectional AGV as the research object,it mainly studies the positioning and guidance method of the AGV under global vision,and realizes the trackless automatic guidance of the AGV.The main research work of this paper is as follows:(1)The overall design of the system.According to the selected global vision guidance method,the overall structure of the AGV system is designed,and the working principle and the hardware equipment used in the system are selected and briefly introduced.The software structure is designed according to the system functions.By analyzing the positioning accuracy of the constructed global map,it is prepared for the subsequent positioning guidance.(2)Research on AGV real-time positioning algorithm.First,the continuous inter-frame difference method is introduced.The traditional moving object detection algorithm has the disadvantage of weak anti-interference in the static state and the presence of multiple moving objects.And the unfixed center of the AGV area detected by the cascade classifier will cause positioning accuracy problems.An AGV positioning and tracking method based on the combination of cascaded classifiers and calibration color mark is proposed.That is,the AGV area is obtained by training the cascade classifier,the initial positioning of the calibration color mark is performed in this area,and the detection area is continuously updated to obtain realtime positioning.(3)A path tracking control scheme in trackless map is designed to realize the function of automatic guidance.First,aiming at the problem of global map positioning accuracy,a global map coordinate prediction based on the least square regression model is proposed to reduce the target point positioning error.Through the kinematic analysis of the three-wheel omnidirectional AGV,the path tracking control method combining the dynamic rotation of the turning point and the differential control path correction is adopted to realize the dynamic tracking and guidance of the AGV in the trackless map.(4)Based on the proposed AGV localization and guidance algorithms,a system platform with functions such as global map real-time display,parameter setting,path design,manual control,automatic guidance,operation status and operation trajectory display was built.A reasonable test scheme is designed,including the test of the system and the analysis of the system operation results.The feasibility of the system is verified,and the positioning and autonomous guidance requirements of AGV are basically realized.