Scholarlink

<p>To address challenges such as low efficiency in highway slope deformation monitoring, insufficient precision in geological hazard susceptibility evaluation, and hyperparameter optimization of prediction models, this study developed an integrated monitoring-prediction framework combining the Small Baseline Subset Interferometric Synthetic Aperture Radar (SBAS-InSAR) Technology Integrating Persistent Scatterers Points with a&#xa0;hybrid algorithm incorporating game theory, Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise, Wavelet Transform Coherence, and an enhanced Multi-Head Differential Transformer mechanism. By processing 45&#xa0;scenes of Sentinel-1A satellite data, deformation time-series of the Benxi-Kuandian Expressway slopes during 2023–2024 were obtained and validated through global positioning system measurements. Deformation rates were integrated with eight other disaster-inducing factors (elevation, slope aspect, rainfall, etc.) to establish a&#xa0;composite evaluation model using game theory-optimized fuzzy analytic hierarchy process and improved CRITIC method. Results revealed spatially uneven deformation patterns, with southern slopes exhibiting maximum annual subsidence of −45.35 mm/a. High susceptibility zones clustered in southern slopes were governed by deformation rates, slope gradients, rainfall, and lithology. The proposed ICEEMDAN-WTC-Itransformer model achieved optimal prediction performance, demonstrating maximum root mean square error of 0.85 mm and maximum absolute error of 0.79 mm, significantly outperforming conventional models. This framework enables precise slope stability monitoring and prediction through multi-source data fusion and algorithmic synergy, providing robust technical support for highway geological hazard prevention.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Susceptibility Assessment and Deformation Prediction of Highway Slope Geological Hazards Based On Game Theory and ICEEMDAN-WTC-Itransformer

  • 期刊论文

  • 发表日期：

    2026年1月29日

查看全文
  • Yuqi Su,
  • Ruren Li,
  • Mengchen Li

摘要

To address challenges such as low efficiency in highway slope deformation monitoring, insufficient precision in geological hazard susceptibility evaluation, and hyperparameter optimization of prediction models, this study developed an integrated monitoring-prediction framework combining the Small Baseline Subset Interferometric Synthetic Aperture Radar (SBAS-InSAR) Technology Integrating Persistent Scatterers Points with a hybrid algorithm incorporating game theory, Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise, Wavelet Transform Coherence, and an enhanced Multi-Head Differential Transformer mechanism. By processing 45 scenes of Sentinel-1A satellite data, deformation time-series of the Benxi-Kuandian Expressway slopes during 2023–2024 were obtained and validated through global positioning system measurements. Deformation rates were integrated with eight other disaster-inducing factors (elevation, slope aspect, rainfall, etc.) to establish a composite evaluation model using game theory-optimized fuzzy analytic hierarchy process and improved CRITIC method. Results revealed spatially uneven deformation patterns, with southern slopes exhibiting maximum annual subsidence of −45.35 mm/a. High susceptibility zones clustered in southern slopes were governed by deformation rates, slope gradients, rainfall, and lithology. The proposed ICEEMDAN-WTC-Itransformer model achieved optimal prediction performance, demonstrating maximum root mean square error of 0.85 mm and maximum absolute error of 0.79 mm, significantly outperforming conventional models. This framework enables precise slope stability monitoring and prediction through multi-source data fusion and algorithmic synergy, providing robust technical support for highway geological hazard prevention.