Significant regional environmental differences result in varied patterns of pavement temperature changes. To enhance the cross-regional adaptability of temperature prediction models, transfer learning (TL) was introduced into the random forest (RF) model to improve its generalization capability. Firstly, meteorological data on air temperature, solar radiation, relative humidity, and wind speed were collected from different regions. Pavement temperatures at various depths were also monitored over a long period. Secondly, prediction models were constructed using the RF method. The prediction performance of the models was evaluated. Thirdly, the RF model was optimized using TL with a feature enhancement strategy. Finally, the optimized model was validated using data from other regions not included in the initial training set. The results indicated that although the RF model achieved good prediction accuracy within individual regions, its performance declined when applied across different regions. After optimization through TL with feature enhancement, the model’s prediction accuracy in target regions was significantly improved. Specifically, the mean squared error was reduced from 44.91 to 14.88, the mean absolute error from 4.91 to 2.78, and the coefficient of determination increased from 0.75 to 0.92. Further validation revealed that the determination coefficient exceeded 0.94 and the mean absolute error remained below 2.3 °C at all depths. In summary, the transfer learning approach based on the random forest model demonstrates strong adaptability to different regions. It effectively addresses the issue of reduced prediction accuracy caused by regional differences and provides a reliable method for accurate pavement temperature prediction across multiple regions.

The mechanical response of asphalt pavement under vehicular loading is an essential reference for crack-resistant pavement design. However, limited research focuses on the mechanical response measurement of asphalt pavement using a large-particle-size crushed stone base treated with fly ash and slag powder. Therefore, two types of asphalt pavements were constructed. The first type of asphalt pavement uses a large-particle-size crushed stone base treated with fly ash and slag powder, where the slag powder uses granulated blast furnace slag powder. The second type uses a conventional cement-stabilized crushed stone base and serves as a reference structure. Firstly, the strain gauges and temperature sensors were installed during the construction of asphalt pavements. Secondly, the mechanical response of the pavement was tested at different speeds and service time conditions. Then, sensitivity analysis and three-factor analysis of variance (ANOVA) were carried out. Finally, the prediction equations were developed. The results show that the longitudinal strain pulse of the asphalt layer exhibited a “compression–tension–compression” characteristic. For the transverse strain pulse of the asphalt layer, the base layer’s transverse and longitudinal strain pulses were only shown as “tensile” characteristics. The vehicular speed significantly affected the strain values for the base and asphalt layers, showing a decreasing trend with increasing speed. For the asphalt layer, the strain values showed an increasing trend with increasing temperature; for the base layer, the strain values showed a decreasing trend with increasing service time. The type of base layer had a significant effect on the strain value. Compared with the conventional base layer, the large-particle-size crushed stone base treated with fly ash and slag powder had lower strain at the base layer and a lower position of the asphalt layer, which could better prevent bottom-up fatigue cracking. Finally, the strain prediction model of the pavement under the speed and temperature (service time) was fitted to obtain a model that can predict the mechanical response of the pavement under different operating conditions. The findings of this research can provide a reference for the design of asphalt pavement using a large-particle-size crushed stone base treated with fly ash and slag powder.

The reliability of the mechanical property values adopted for current pavement structural design remains low. Therefore, this study investigated the nonlinear characteristics for the tensile and compressive properties of cement-stabilized macadam (CSM) and an asphalt mixture (AM) under different gradation types and loading rates. And a multi-factor value model was developed for both. The results show that the tensile and compressive stress–strain behavior of both CSM and AM exhibited bilinear characteristics consistent with the bi-modulus theory (elasticity with different moduli in tension and compression). The strength, elastic modulus, and Poisson’s ratio followed a power function relationship with increasing loading rates, stabilizing at values beyond 0.1 MPa/s. The skeleton-dense gradation demonstrates the most favorable mechanical performance. For semi-open-graded AM with a void ratio of 8~16%, the mechanical parameters exhibit relatively high rates of change. Among the influencing factors, gradation type had the most significant impact on the mechanical parameters, especially on the elastic modulus. In general, tensile mechanical parameters were more sensitive to changes than compressive ones. R_c/R_t was most strongly affected by gradation type. Accordingly, a quantitative value model was established to describe the variation in tensile and compressive mechanical parameters of typical asphalt pavement materials, which vary with air void ratio, loading rate, binder content, and temperature. The findings provide a reference for the prediction of pavement structure design parameters considering the difference in compression and tension.