Research On Lead Recognition And Motion Control Of AGV In Complicated Environment

With the continuous development of automation and intelligent industry in today's society,AGV(Automated Guided Vehicle)plays a more and more important role in material transmission and other derivative functions.Among them,visual guidance in AGV guidance mode has gradually become one of the hottest perception modes,But in fact,it has not been commercialized and applied on a large scale.In the actual operation of AGV,it is inevitable to encounter such phenomena as overexposure or shadow area,and the guide wire may be not obvious or even damaged,which may interfere with the identification of the guide wire.Therefore,in order to improve the robustness of visual guided AGV,the shortcomings of the existing technology are solved from the software level,This paper focuses on the problem that it is difficult to accurately extract the guide wire of vision guided AGV in complex working environment,and explores the application of anti-interference vision technology in the field of AGV;At the same time,the calculation of navigation deviation and lateral motion control are also studied and verified in order to improve the motion quality of AGV.It provides a new method and idea for path tracking control of vision guided AGV in image processing and motion control.The main contents of this paper are as follows:(1)This paper studies the algorithm of AGV leader recognition,analyzes the technical difficulties in different stages and the shortcomings of the existing technology,and divides the process of leader recognition and extraction into four steps: image preprocessing,edge extraction,leader selection and navigation deviation calculation.In order to solve the problem of non illumination region and non illumination region in two-dimensional image,a fast algorithm is designed.In this paper,in the process of guide line extraction and selection,firstly,the color region is selected according to the interval of H,s and V components of different colors,so as to obtain binary image;then,the ant like colony algorithm is used to preliminarily detect the image edge,so that the algorithm can extract all the lines that meet the conditions on the premise of real-time;and then the guide line width feature is introduced in the process of guide line selection Then,according to the forward looking value,the longitudinal position of the preview point is calculated,and the specific coordinates of the preview point are determined from the recognized guide line by using the image moment,which is different from the absolute center of the image to get the navigation deviation.Finally,the algorithm is validated in different typical scenes,and the results show that the image preprocessing and lead extraction algorithm used in this paper has high anti-interference ability and recognition rate.(2)The lateral motion control of AGV is deeply studied.Firstly,the desired yaw rate is obtained according to the optimal preview control theory and kinematics model,and then the front wheel angle is calculated by using the variable universe fuzzy sliding mode controller according to the deviation between the actual and expected yaw rate.In the calculation process,the traditional sliding mode controller is optimized,and the variable universe fuzzy control is used to adjust the switching gain coefficient in time,thus becoming a variable universe fuzzy sliding mode controller based on preview.Finally,the motion control algorithm designed in this paper is verified based on Simulink and Car Sim co simulation platform,and compared with PID control and traditional fuzzy sliding mode control.The results show that the variable universe fuzzy sliding mode control algorithm based on preview has high control accuracy and fast convergence.Compared with fuzzy sliding mode control,the algorithm can adapt to more working conditions,has faster response speed,higher control accuracy and adaptive ability.(3)After the design of the control algorithm part,this paper introduces the construction of the AGV system function test platform,the test scene and the result analysis.The hardware part introduces the hardware composition and the function of each module of the AGV vehicle test platform,and the software part introduces the communication mode,infrastructure and the specific algorithm implementation of each module.Finally,on the basis of verifying the function of each part of the control algorithm separately,the comprehensive performance test of real vehicle is carried out in typical complex scenes relying on AGV vehicle test platform.The test results show that the perception algorithm adopted in this paper is feasible and effective,and has strong anti-interference performance,better motion quality and faster computing performance,It can meet the requirements of vision guided AGV in various complex environments.