Path Planning Algorithm Design For Multi-AGVs In The Multi-task Scenario Of Logistics Center

In recent years,the rapid growth of express delivery volume has brought huge challenges to the logistics industry.The outbreak of Covid-19 epidemic has put forward higher requirements on the logistics system.The timely delivery of masks,medicines,etc.,and the protection of basic daily necessities are inseparable from the support of the logistics industry.Warehousing throughput is very important in the entire logistics system.The traditional manual sorting method can no longer meet the requirements of a large number of items,high throughput,and reduced contact.In order to solve these problems,modern logistics warehouses have introduced Automated Guided Vehicles(AGV)to complete a large number of tasks.This paper focuses on the real-time path planning of multiple AGVs in the multi-task scenario of logistics center,with the goal of improving the efficiency of logistics center cargo transportation and meeting the real-time requirements in practical applications,and the following aspects of work were mainly completed:For the problem of multiple AGV path planning in a single task scenario,this paper proposes a segmented path planning algorithm that limits the execution time.The algorithm does not need to plan conflict-free paths for all AGVs at one time,but only needs to plan the movement of the next time step for the AGV,so as to achieve the goal of fast path planning.In terms of conflict handling,the algorithm uses safe intervals to handle conflicts.By establishing a safety interval on each node,it is confirmed whether the AGV needs to wait during the search path.A large number of comparative experiments prove that the algorithm reduces the task completion time of the algorithm while keeping the success rate basically unchanged.For the problem of multi-AGV path planning in the continuous multi-task scenario,this paper proposes a two-stage multi-AGV path planning algorithm based on jump points.In the offline phase,all jump points are calculated,and in the online phase,path planning is performed according to tasks and conflicts are resolved.In view of the problem of AGV congestion under large-scale tasks,a congestion control strategy is proposed,and heuristic weights are introduced to avoid the excessively high path overlapping due to the repeated selection of some points.If the congestion degree exceeds the threshold,the algorithm changes the target point of the corresponding AGV to a temporary buffer point,avoiding congestion to a certain extent.The analysis of the comparative experiment proves that the algorithm can reach a greater maximum throughput in the same scenario,improve the transportation efficiency of goods and meet the real-time requirements of practical applications.