A hierarchical federated learning framework for collaborative quality defect inspection in construction

• Published in: Engineering applications of artificial intelligence 2024-07, Vol.133, p.108218, Article 108218
• Main Authors: Wu, Hai-Tao, Li, Heng, Chi, Hung-Lin, Kou, Wei-Bin, Wu, Yik-Chung, Wang, Shuai
• Format: Article
• Language: English
• Subjects: Construction quality inspection; Construction robot; Crack; Federated learning; Quality defect
• ISSN: 0952-1976; 1873-6769
• DOI: 10.1016/j.engappai.2024.108218

Recent advancements in robotics and deep learning (DL) have made it possible to implement robots in civil infrastructures' quality defect inspection. Robots can reduce human inspectors' workloads and enhance inspection results' reliability by collecting data and automatically identifying quality defects from the raw data. However, current methods for training DL models rely on centralized strategies that require the aggregation of defect data (e.g., uploading to a cloud server), posing concerns about data privacy and security. Thus, this study proposes a three-fold federated learning (FL) framework for training DL models collaboratively, without the need to share local data among construction robots. The framework is specifically applied to image-based crack segmentation, critical for ensuring infrastructures’ safety and serviceability. A lightweight DL model is developed to enable easy implementation on resource-constrained construction robots and to reduce communication costs during federated training. Experimental results show that the proposed FL method outperforms traditional centralized methods. The critical contribution of this study is the hierarchical FL framework, which enables construction robots to leverage big data in a privacy-preserving manner. •A cloud-edge-robot hierarchical federated learning (FL) method is proposed.•Robots collaboratively train the model for construction quality defect inspections.•A lightweight crack segmentation model, termly, CrackNet, is proposed.•CrackNet achieves comparable performance with less parameters.•FL outperforms centralized training methods and typical FL methods.