Research On AGV Path Planning Algorithm Based On Visual Labeling

With the development of computer science and automation technology,flexible manufacturing system is applied to enterprise production practice more frequently.As an intelligent and indispensable means of transport tool in flexible manufacturing system,AGV(Automated Guided Vehicles)plays a pivotal role in the production and transportation of materials.AGV can not only reduce human intervention,improve transportation efficiency,and saves transportation cost,easy to expand,has a very important practical significance to enhance the core competitiveness of enterprises.The path planning problem is an important problem in the practical application of AGV system.In this paper,the AGV path planning algorithm is analyzed in combination with the actual logistics and transportation needs of contemporary enterprises.A set of visually-based Multi-AGV Path Planning System is designed and developed.First of all,the paper introduces the overall framework of AGV path planning system,the whole system is divided into console software and AGV hardware system,both through the wireless local area network connection to exchange data.AGV hardware system is mainly divided into sensing layer,computing layer and response layer,and a brief introduction is done to the visual tag design and image processing identification process,to provide AGV basic position and angle information for the path planning algorithm.Secondly,we analyze and compare various electronic map modeling methods,and use topological modeling to model the warehouse environment and complete the digital work of AGV operating environment.The single AGV path planning algorithm is the core of the software part of the console.Combined with the system stability and the extensibility of the algorithm,the Dijkstra algorithm is used as the shortest path search algorithm.Then,we discuss the problems existing in the actual operation of multiple AGV systems,and analyze two kinds of multi-AGV path planning algorithms based on time window and directed graph.The advantages and disadvantages of the two methods are compared and analyzed.According to the practical application requirements,the method of directional graphs restricts the electronic map model by setting the direction of all the sections,using the stop waiting strategy to avoid the potential conflict between the multiple AG Vs,so that it can cycle and order the transportation tasks in the warehouse.Finally,the main design of the console software is introduced.The software is divided into modules such as electronic map,task management and monitoring.The object-oriented C++ language is developed on the Visual Studio 2010 platform to realize the various functions required for each module.With the help of MFC to draw man-machine interface,through the simulation experiment to complete the algorithm test,verify the effectiveness and practicality of the algorithm.