Conflict-Based Search for Multi-AGV Path Planning in Intelligent Storage System

The rapid advancement of smart storage systems has intensified the demand for simultaneous collision-free path planning of multiple Automated Guided Vehicles (AGVs), posing significant challenges in ensuring collision-free and efficient operations. This paper constructs grid maps based on typical storage environments and formulates the multi-AGV path planning problem as graph-based search problem: Multi-Agent Path Finding (MAPF). To address conflicts among AGVs, we adopt the Conflict-Based Search (CBS) algorithm and use priority conflict strategy to optimize the efficiency of high-level search. In the low-level single AGV path planning, we introduce an obstacle density heatmap and design a partitioned search strategy. By dynamical heuristic weight and incorporating penalty terms, the smoothness, efficiency, and safety of path planning are significantly improved. Simulation experiments results demonstrate that our proposed improved CBS algorithm can effectively handle multi-AGV path planning in storage environments, achieving conflict-free path planning.