A Continuous Path Planning Method for Multi-AGV Tasks Based on an Improved ECBS Algorithm

This thesis addresses the Multi-Agent Path Finding Problem (MAPF) in the context of multi-AGV (Automated Guided Transport Vehicle) systems operating under continuous task scenarios. An enhanced version of the Conflict-Based Search (ECBS) algorithm is proposed, aimed at overcoming the limitations of the traditional ECBS algorithm in handling continuous multi-tasks within large-scale environments. The improvements involve optimization strategies such as task spreading, path truncation, and task articulation. Initially, the study reviews the strengths and weaknesses of existing multi-agent path planning algorithms, emphasizing the shortcomings of popular approaches when dealing with continuous tasks. The proposed algorithm combines the benefits of low-level path conflict resolution with high-level global planning, striking a balance between planning efficiency and solution quality through a suboptimal search strategy. This addresses the inability of the traditional ECBS algorithm to manage continuous multi-path point tasks. Experimental results demonstrate the improved algorithm's adaptability, high efficiency, and solution quality in dynamic environments. The research offers valuable theoretical and practical insights for solving multi-AGV continuous task path planning problems in complex scenarios.