Research And Implement Of Multi-AGV Path Planning Algorithm

Automated Guided Vehicle(AGV)path planning is an important part of Automated warehouse.With the popularization of digital and intelligent storage systems,more and more AGVs have been applied to the storage systems to participate in the moving and handling of materials in the warehouse,thus improving the efficiency and reducing the transportation cost of enterprises.For the multi-AGV path planning problem caused by the path conflict between AGVs,it is necessary to combine the site environment of the actual operation and the business logic of warehouse handling,so as to improve the efficiency of mobile handling task as much as possible.The main research work done in this paper includes: combining the characteristics of fire-fighting warehouse environment,using raster map modeling,designing the corresponding map information storage structure,and completing the construction of electronic map;For the search algorithm of single AGV path planning,the A *algorithm based on Manhattan distance is selected as the basic algorithm of path optimization,and the turn factor is introduced to optimize it.For multi-AGV path planning,Conflict Based Search(CBS)algorithm is adopted to divide the multi-AGV path planning into two levels for solution.The lower level of the algorithm adopts the optimized A* algorithm to solve multiple single AGV path planning problems,while the upper level of the algorithm USES constraint tree nodes to solve the Conflict between multiple AGV paths.On the basis of CBS algorithm,by designing multiple threads to optimize the execution process in parallel and improve the computing efficiency,this optimization design is applied to CBS algorithm based on focused search.It verifies and compares the parameters of the optimized A* algorithm and the traditional A* algorithm,the improved CBS algorithm and the traditional CBS algorithm.Through Python's matplotlib library,the trajectory of multi-AGV path planning is simulated.The test results show that the improved A* algorithm reduces the number of turns by 17% compared with the traditional A* algorithm at the same path cost.Compared with the traditional CBS algorithm,the parallel CBS algorithm reduces the time of path planning by 33.9%.Under similar scenarios such as automatic fire-fighting warehouse,the multi-AGV path planning algorithm studied in this paper can plan the path in a shorter time,and the number of path turns is lower,which has certain research significance and practical value.