基于联邦学习的跨区域隧道安全监测模型构建方法及系统
By constructing cross-regional local monitoring nodes and using federated learning methods, the problem of sharing tunnel operation data was solved, enabling the efficient construction and updating of cross-regional tunnel safety monitoring models and improving the applicability and stability of the models.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CCCC SECOND HIGHWAY CONSULTANTS CO LTD
- Filing Date
- 2026-06-22
- Publication Date
- 2026-07-17
AI Technical Summary
Tunnel operation data from different regions and operating units are difficult to share directly. Samples are scattered, abnormal samples are scarce, cross-regional model generalization ability is insufficient, data security requirements are high, and existing model updates are difficult to adapt to the evolution of tunnel operation status and the development of defects.
Cross-regional local monitoring nodes are constructed to access multi-source heterogeneous tunnel operation monitoring data. After preprocessing, the data is trained locally, and parameter updates are generated through federated learning. After security processing, the data is uploaded to the central server. The central server performs quality-aware weighted aggregation and personalized calibration to form a global model, which is then distributed to the local nodes.
This system enables cross-regional tunnel operation safety monitoring models to operate within their respective domains, reducing the security risks associated with cross-domain data sharing, improving the model's applicability and stability, adapting to the characteristics of different regions and tunnel types, and enhancing the model's generalization ability and update efficiency.
Smart Images
Figure CN122412967A_ABST