Incorporating mobile robots into the production shop-floor helps realize the concept of smart production, and it is considered one of the approaches to enhance manufacturing and operational efficiency and effectiveness by academics and industrial practitioners. This paper develops a cyber-physical robotic mobile fulfillment system (CPRMFS) for tool storage in smart manufacturing. The purpose is to enable Just-in-Time material transfer on the production shop-floor during manufacturing. A decentralized multi-robot path planning adopts graph neural networks (GNN) in the new proposed CPRMFS. We compare multiple classification algorithms for the mobile robots' action prediction, including proposing a spatial-temporal graph convolutional network (ST-GNN) under these circumstances. We also extend the research with the enhanced conflict-based search path planning algorithm. Compared with the existing literature, ST-GNN, under the enhanced conflict-based search, could obtain higher accuracy with an average value of 90% under different scenarios. The practical applicability of the proposed system with the further consideration of ST-GNN is further explained as a reference for manufacturing practitioners who looked out on a confrontation of introducing the mobile robot solutions in their manufacturing site with the goal of enhancing the operation processes.
Bibliographical noteCopyright © 2023 Elsevier Ltd. This accepted manuscript version is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License [https://creativecommons.org/licenses/by-nc-nd/4.0/].
- Cyber-physical production system
- Graph neural networks
- Robotic mobile fulfillment system
- Smart manufacturing