The Research Of Iocation Technology And Path Planning Based On Visual Navigation AGV

With the rapid development of industrial automation industry,the traditional manual factory logistics can not adapt to the need of moden automated production,and so more and more researchers begin to study intelligent logistics system.Automatic Guided vehicle(AGV)has been widely used in intelligent logistics system due to high intelligence,environmental adaptability,reliability and safety.C.ompared with other navigation AGV,Visual Guidance AGV has great advantages in road planning,safe obstacle avoidance and high-precision positioning.Positioning technology and path planning technology in visual navigation AGV are the key to improve efficiency.Therefore,visual positioning technology and path planning technology was studied in this paper,and the main contents are as follows:(1)Through analyzing the development status of visual navigation AGV and the key technologies of visual navigation,the focus of this paper is visual positioning technology and path planning technology are deeply researched in this paper.(2)The system structure of AGV and kinematics model of two-wheel differential velocity were established,which laid a mathematical foundation for visual positioning,and also the visual sensor was selected and debugged according to the working principle of visual navigation.(3)Compared the advantages and disadvantages of the matching algorithm.ORB was selected as the most suitable visual matching algorithm for this paper.Using convolutional neural network algorithm for image segmentation,the interested area image are extracted,and only the pure image of the scene remained.At the salme time image feature is extracted through the visual dictionary algorithm in order to improving the matching speed.Then using TF-IDF algorithm to calculate the similarity,and combining the PNP algorithm and ICP algorithm to calculate the relative position between image matching.Finally,the RANSCA and nonlinear least square method are used to optimize the estimated relative pose.The feasibility of the algorithm is verified by experiments.(4)The advantages and disadvantages of artificial potential field algorithm are analyzed in local path planning.And the velocity repulsion field is introduced to solve the problem of dynamic obstacle avoidance,So a new gravity function and a repulsive force function are introduced to solve the problem of target inaccessibility,and also the virtual target point is added to guide AGV out of the minimum region.In the global programming,an improved ant colony algorithm with potential field is proposed to solve the blind search and time-consuming problems of point ant colony.And the search direction of ant colony in ant colony algorithm is provided according to the improved artificial potential field algorithm.The results of simulation experiment show that compared with the single artificial potential field method and ant colony algorithm,The improved ant colony algorithm with potential field has a good effect on path planning in automated workshop environment.In this paper,visual location algorithm and path planning algorithm are proposed.And the results show that both have high stability and accuracy.