Academic Journal of Computing & Information Science, 2025, 8(10); doi: 10.25236/AJCIS.2025.081002.
Jingni Ma1, Junli Feng1
1School of Information Science and Engineering, Chongqing Jiaotong University, Chongqing, 400074, China
With increasing bridge service life and traffic loads, real-time structural health monitoring faces challenges from high-dimensional, non-stationary, and heterogeneous data. This paper proposes a Spatio-Temporal Multi-Relational Graph Attention Network (ST-MRGAT), which constructs spatio-temporal graphs from bridge sensor nodes and incorporates structural mechanics priors to capture complex spatial dependencies. The model employs enhanced gated temporal convolution to capture short-term dynamics and long-term evolution, and multi-relational graph attention to unify local cross-section and overall bridge responses. Experiments on real-world data show that ST-MRGAT significantly outperforms baseline models in multi-step predictions of deflection and strain, maintaining minimal error growth and demonstrating high accuracy, stability, and effective multi-task feature sharing, validating its robustness and generalizability for bridge structural health monitoring.
Bridge Health Monitoring; Spatiotemporal Prediction; Multi-Relational Graph; Graph Attention Network
Jingni Ma, Junli Feng. ST-MRGAT: Multi-Relational Spatio-Temporal Graph Attention for Multi-Task Prediction of Bridge Deflection and Strain. Academic Journal of Computing & Information Science (2025), Vol. 8, Issue 10: 8-13. https://doi.org/10.25236/AJCIS.2025.081002.
[1] Rizzo P, Enshaeian A. Challenges in bridge health monitoring: A review[J]. Sensors, 2021, 21(13): 4336.
[2] Hochreiter S, Schmidhuber J. Long short-term memory[J]. Neural computation, 1997, 9(8): 1735-1780.
[3] Chung J, Gulcehre C, Cho K H, et al. Empirical evaluation of gated recurrent neural networks on sequence modeling[J]. arXiv preprint arXiv:1412.3555, 2014.
[4] Lea C, Vidal R, Reiter A, et al. Temporal convolutional networks: A unified approach to action segmentation[C]//European conference on computer vision. Cham: Springer International Publishing, 2016: 47-54.
[5] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[J]. Advances in neural information processing systems, 2017, 30.
[6] Box G E P, Pierce D A. Distribution of residual autocorrelations in autoregressive-integrated moving average time series models[J]. Journal of the American statistical Association, 1970, 65(332): 1509-1526.
[7] Wu Z, Pan S, Chen F, et al. A comprehensive survey on graph neural networks[J]. IEEE transactions on neural networks and learning systems, 2020, 32(1): 4-24.
[8] Li Y, Yu R, Shahabi C, et al. Diffusion convolutional recurrent neural network: Data-driven traffic forecasting[J]. arXiv preprint arXiv:1707.01926, 2017.
[9] Yu B, Yin H, Zhu Z. Spatio-temporal graph convolutional networks: A deep learning framework for traffic forecasting[J]. arXiv preprint arXiv:1709.04875, 2017.
[10] Wu Z, Pan S, Long G, et al. Graph wavenet for deep spatial-temporal graph modeling[J]. arXiv preprint arXiv:1906.00121, 2019.
[11] Guo S, Lin Y, Feng N, et al. Attention based spatial-temporal graph convolutional networks for traffic flow forecasting[C]//Proceedings of the AAAI conference on artificial intelligence. 2019, 33(01): 922-929.