Research On Multi-AGV Path Planning And Conflict Avoidance In Automated Warehouse System

The explosive development of e-commerce,express delivery and other industries puts higher requirements on the operational efficiency of wareh ousing and logistics.Multi-AGV automated warehouse systems can effectively improve the operating mode of traditional warehouses with low efficiency,high rate of erroneous orders,and poor flexibility of transfer tools.The high drawbacks of manpower dependence,and the realization of logistics automation and informationization,are the hot research directions with high returns because of their outstanding contribution to improving land utilization and reducing labor costs.This paper mainly optimizes and implements the algorithm of path planning and collision avoidance for key issues of multi-AGV automated warehouses.First of all,to study the navigation methods,combined with the actual situation of the automated warehouse and the accuracy,cost and other considerations,UHF RFID technology will be applied to the guidance of AGV.Studying various map modeling methods and abstracting the automated warehouse environment into topological electronic maps,the continuous dynamic multi-objective problem is reduced to a discrete multi-objective problem,which greatly reduces the amount of data transmission and space occupation,and analyzes the adjacency matrix by contrast.Adjacent to the time and space complexity of the table,combined with the number of map points and lines,the adjacency table is used as a map storage data structure.Secondly,the characteristics of commonly used single-source shortest path algorithms—Dijkstra algorithm,A-STAR algorithm,Bellman-Ford algorithm,and SPFA algorithm are analyzed.Since the map created in Chapter 1 is a right weight graph,Dijkstra and A-STAR algorithms are considered as alternative algorithms..By introducing the exponential decay factor as a weight into the cost function of the A-STAR algorithm,the A-STAR algorithm is improved so that the search speed can be accelerated when the distance between the current point and the target point is long,and the index term is close to 1 when the target point is near,ensuring that the target point is reachable.Under the same conditions,Dijkstra's algorithm,A-STAR algorithm and improved A-STAR algorithm are simulated to prove that all three algorithms can obtain the optimal solution.The improved A-STAR algorithm has the least search point.To ensure the real-time performance of automated warehouse system,As a single AGV path planning algorithm.Then,the priority rules for order tasks and AGV task execution are defined.By adding a penalty term that is proportional to the frequency of use of the edge to the weight of the edge,the crowded road section of the multi-AGV is evacuated,and the total running time of the multi-AGV system is reduced.The linear function,the exponential function and the Sigmoid function are used as punishment items for the three groups respectively.Multi-task and multi-AGV parallel experiments,according to the simulation results,the Sigmoid function as a punishment item when the map load balance,there is no large-scale congestion,so it is used as a map weight optimization function.Finally,according to the characteristics of the warehouse environment and the AGV operation mode of the cargo-to-people,this paper presents four possible conflict situations in the environment,and designs a three-level parallel collision avoidance mechanism,including offline prediction avoidance,setting lock zone and passive two-stage obstacle avoidance strategy.Effectively ensure the reliability of the AGV system.