Fundamental understanding of micro-mechanical behaviors in bitumen, including phase separation, micro-friction, micro-abrasion, etc., can help the pavement engineers better understand the bitumen mechanical performances at macroscale. Recent researches show that the microstructure evolution in bitumen will directly affect its surface structure and micro-mechanical performance. In this study, the bitumen microstructure and micro-mechanical behaviors are studied using Atomic Force Microscopy (AFM) experiments, Phase Dynamics Theory and Molecular Dynamics (MD) Simulation. The AFM experiment results show that different phase-structure will occur at the surface of the bitumen samples under certain thermodynamic conditions at microscale. The phenomenon can be explained using the phase dynamics theory, where the effects of stability parameter and temperature on bitumen microstructure and micro-mechanical behavior are studied combined with MD Simulation. Simulation results show that the saturates phase, in contrast to the naphthene aromatics phase, plays a major role in bitumen micro-mechanical behavior. A high stress zone occurs at the interface between the saturates phase and the naphthene aromatics phase, which may form discontinuities that further affect the bitumen frictional performance.

Efficient and accurate road distress detection is crucial for infrastructure maintenance and transportation safety. Traditional manual inspections are labor-intensive and time-consuming, while increasingly popular automated systems often rely on computationally intensive devices, limiting widespread adoption. To address these challenges, this study introduces MobiLiteNet, a lightweight deep learning approach designed for mobile deployment on smartphones and mixed reality systems. Utilizing a diverse dataset collected from Europe and Asia, MobiLiteNet incorporates Efficient Channel Attention to boost model performance, followed by structural refinement, sparse knowledge distillation, structured pruning, and quantization to significantly increase the computational efficiency while preserving high detection accuracy. To validate its effectiveness, MobiLiteNet improves the existing MobileNet model. Test results show that the improved MobileNet outperforms baseline models on mobile devices. With significantly reduced computational costs, this approach enables real-time, scalable, and accurate road distress detection, contributing to more efficient road infrastructure management and intelligent transportation systems.

The aggregate skeleton significantly influences the deformation resistance of asphalt mixtures. A stable skeleton has good interlocking effects and load transfer ability to resist deformation of mixtures during loading. Therefore, the methodology to accurately evaluate the morphology of aggregate skeleton based on the mechanism of load transfer inside mixtures is of great value to predict the mixture stability. The objective of this study is to predict and validate mixture stability from the perspective of evolutional morphologies of the skeleton. The methodology has three main steps, which are as follows: (1) the initial morphology of the aggregate skeleton is characterized and evaluated; (2) the topological and non-topological evolutions of the skeleton morphology during external loading are characterized; and (3) the strain energy of the mixture and the stress distribution on contact regions in the skeleton are used to analyze and validate the prediction of the mixture's stability. Simulations of uniaxial displacement-controlled tests of three cylindrical specimens drilled in a field-compacted test track were conducted, and 7.5, 15, 22.5, and 30 s were selected as time points in the simulation to analyze mixture stability. Results indicate that a mixture has a better load-bearing capacity when its initial skeleton contains more chains with higher evaluation indices, which proves the reliability of the stability prediction using the proposed method.