Multi-AGV Simulation System Based on Bipartite Graph, Space-Time A, and Conflict Search

• Published in: Chinese Automation Congress (Online) pp. 3767 - 3772
• Main Authors: Liang, Yihao, Liu, Mingrui
• Format: Conference Proceeding
• Language: English
• Published: IEEE 25.11.2022
• Subjects: Bipartite graph; conflict search; Job shop scheduling; KM algorithm; matching problem; Multi-AGV; Path planning; Processor scheduling; Real-time systems; Search problems; smart logistic; space-time A; Visualization
• ISSN: 2688-0938
• DOI: 10.1109/CAC57257.2022.10056052

Smart logistic industry prefers to use multiple AGVs for package dispatch. AGVs work in a pre-designed map of dispatching center, and are unified scheduling and planning by a high-performance centralized computing equipment. To proper function, these AGVs need to complete collision-free path planning in the map to reach their destination, and need to make a decision to return to a package entrance to pick up the package after completing the delivery. In this paper, we use optimal matching based on bipartite graph as the model of AGVs' returning decision, and KM algorithm to solve the optimal matching iteratively using graph theory. For the collision-free path planning, we use space-time A* algorithm based on time stamp to search for conflicts in the space-time domain, and a binary search tree based on conflict search is used to find the collision-free optimal matching results as the upper layer of space-time A*. Additionally, we have integrated the two algorithms and placed them in a simulation system to visualize and evaluate the effects of the two methods on solving the problem. The algorithm shows good collision-free planning ability and real-time performance.