LAUSR

A great challenge in modern automated logistics is to coordinate multi-AGV system safely and efficiently. Current strategies which detect collisions and deadlocks mainly by nodes are not feasible in some scenes. With the aim of avoiding collisions and preventing deadlocks regardless of the layout of warehouses and the size of AGVs, in this letter we explore a coordination strategy combining nodes and grids, where nodes are physical locations of the environment and grids are tiny and independent spaces resulted from a virtual division of the environment. The proposed strategy guarantees the safety of the system and prevents 75%-100% deadlocks compared to traditional strategies. A fast grid calculating method is also presented to ensure real-time control of the multi-AGV system. The proposed strategies are validated by simulations and applications. Experiment results demonstrated that our approach can avoid all potential collisions and prevent more than 75% deadlocks, increasing the overall efficiency greatly.