To address the technical limitations of traditional pavement performance prediction models in capturing temporal features and analyzing multi-factor coupling, this study proposes a Bayesian Optimization Dual-Layer Feature Fusion Model (BO-DLFF). The framework integrates heterogeneous data streams from embedded strain sensors, temperature/humidity monitoring nodes, and weigh-in-motion (WIM) systems, combined with pavement distress detection and historical maintenance records. A dual-stage feature selection mechanism (BP-MIV/RF-RFECV) is developed to identify 12 critical predictors from multi-modal sensor measurements, effectively resolving dimensional conflicts between static structural parameters and dynamic operational data. The model architecture adopts a dual-layer fusion design: the lower layer captures statistical patterns and temporal–spatial dependencies from asynchronous sensor time-series through Local Cascade Ensemble (LCE) ensemble learning and improved TCN-Transformer networks; the upper layer implements feature fusion using a Stacking framework with logistic regression as the meta-learner. BO is introduced to simultaneously optimize network hyperparameters and feature fusion coefficients. The experimental results demonstrate that the model achieves a prediction accuracy of R2 = 0.9292 on an 8-year observation dataset, effectively revealing the non-linear mapping relationship between the Pavement Condition Index (PCI) and multi-source heterogeneous features. The framework demonstrates particular efficacy in correlating high-frequency strain gauge responses with long-term performance degradation, providing mechanistic insights into pavement deterioration processes. This methodology advances infrastructure monitoring through the intelligent synthesis of IoT-enabled sensing systems and empirical inspection data.
15 October 2024 in International Journal of Pavement Engineering
The rutting depth is one of the important indicators for evaluating the performance of road use. Accurately predicting the Rutting Depth Index (RDI) can timely detect and solve the problem of road damage, ensuring driving safety. This article constructs a multi influencing factor dataset for RDI prediction by integrating road detection, climate, and traffic flow data from the AB expressway in Shaanxi Province, China. by comparing the prediction results of four models, namely Random Forest(RF), XGBoost, CatBoost, and LightGBM, it was found that XGBoost and CatBoost has the highest accuracy on the test dataset. To further improve the predictive precision of above model, we utilized the Particle Swarm Optimization (PSO) technique to meticulously adjust the hyperparameters of both XGBoost and CatBoost. The experimental results show that using the PSO-XGBoost and PSO-CatBoost models to predict RDI, the R2 scores reached 0.8029 and 0.8135, respectively. Compared with the XGBoost and CatBoost models, the accuracy improved by 0.0655 and 0.022, respectively. This indicates that the application of PSO significantly enhances the predictive accuracy of both XGBoost and CatBoost models for road rutting prediction, with the PSO-CatBoost model demonstrating superior performance. Consequently, it provides valuable data for road maintenance strategies.
23 October 2025 in Environment, Development and Sustainability