Private Data Sharing in EMU Maintenance: A Method Study Based on Federated Learning

• Published in: 2022 3rd International Conference on Computer Science and Management Technology (ICCSMT) pp. 137 - 143
• Main Authors: Yang, Jiaying, Huang, Lei, Wang, Ying
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.11.2022
• Subjects: Computational modeling; Data models; Data privacy; Data sharing; EMU maintenance; Federated learning; Information sharing; Keywords: Federated learning; Maintenance engineering; Private data; Training
• DOI: 10.1109/ICCSMT58129.2022.00036

In the maintenance of (Electric Multiple Unit) EMU, the operation of intelligent operation and maintenance model needs the support of data sharing from multiple subjects. However, in the context of strict safety supervision in the railway industry, the data security and privacy protection requirements of each subject have formed a huge challenge to data sharing. Therefore, this paper proposes a new method of EMU maintenance data sharing based on federated learning, which solves the problem that some data subjects cannot share data due to the private data. In this paper, a case analysis of the method is carried out to realize the multi-subject joint training of a decision tree prediction model under the premise of protecting the privacy data in the real maintenance scenarios.