Traffic-induced noise pollution is a significant environmental issue, driving the development of advanced noise-reducing pavement materials. Semi-dense graded asphalt mixtures (SDAMs) present a promising compromise, offering enhanced acoustic properties compared to conventional dense-graded asphalt mixtures while maintaining superior durability to porous asphalt mixtures. However, the mechanism underlying the relationship between the energy dissipation characteristics and noise reduction effects of such mixtures remains unclear, which limits further optimization of their noise reduction performance. This study designed and prepared semi-dense graded noise-reducing asphalt mixtures SMA-6 TM, SMA-10 TM, and SMA-13 TM (SMA TM represents noise-reducing SMA mixture) based on traditional dense-graded asphalt mixtures SMA-6, SMA-10, and SMA-13, and conducted tests for water stability, high-temperature performance (60 °C), and low-temperature performance (−10 °C). Subsequently, energy loss parameters such as loss factor and damping ratio were calculated through dynamic modulus tests to characterize their energy dissipation properties. The mechanism linking the energy dissipation characteristics of semi-dense graded asphalt mixtures to noise reduction was investigated. Finally, the noise reduction effect was further verified through a tire free fall test and a close-proximity (CPX) method. The indoor test results indicate that the semi-dense mixtures exhibited a trade-off in performance: their dynamic stability was 11.1–11.3% lower and low-temperature performance decreased by 4.2% (SMA-13 TM) to 14.1% (SMA-6 TM), with moisture stability remaining comparable. Conversely, they demonstrated superior damping, with consistently higher loss factors and damping ratios. All mixtures reached peak damping at 20 °C, and the loss factor showed a strong positive correlation (R² > 0.91) with energy dissipation. Field results from a test section showed that the optimized SMA-10 TM mixture yielded a significant tire–road noise reduction of 3–5 dB(A) relative to the SMA-13, while concurrently meeting key performance criteria for anti-water ability and durability. This study establishes a link between the energy dissipation in SDAM and their noise reduction efficacy. The findings provide a theoretical framework for optimizing mixture designs and support the wider application of SDAM as a practical noise mitigation solution.

Fault diagnosis for embedded sensors in asphalt pavement faces significant challenges, including the scarcity of real-world fault data and the difficulty in identifying compound faults, which severely compromises the reliability of monitoring data. To address these issues, this study proposes an intelligent diagnostic framework that integrates a Discrete Wavelet Transform (DWT) with a staged, attention-based Long Short-Term Memory (LSTM) network. First, various fault modes were systematically defined, including short-term (i.e., bias, gain, and detachment), long-term (i.e., drift), and their compound forms. A fine-grained fault injection and labeling strategy was then developed to generate a comprehensive dataset. Second, a novel diagnostic model was designed based on a “Decomposition-Focus-Fusion” architecture. In this architecture, the DWT is employed to extract multi-scale features, and independent sub-models—a Bidirectional LSTM (Bi-LSTM) and a stacked LSTM—are subsequently utilized to specialize in learning short-term and long-term fault characteristics, respectively. Finally, an attention network intelligently weights and fuses the outputs from these sub-models to achieve precise classification of eight distinct sensor operational states. Validated through rigorous 5-fold cross-validation, experimental results demonstrate that the proposed framework achieves a mean diagnostic accuracy of 98.89% (±0.0040) on the comprehensive test set, significantly outperforming baseline models such as SVM, KNN, and a unified LSTM. A comprehensive ablation study confirmed that each component of the “Decomposition-Focus-Fusion” architecture—DWT features, staged training, and the attention mechanism—makes an indispensable contribution to the model’s superior performance. The model successfully distinguishes between “drift” and “normal” states—which severely confuse the baseline models—and accurately identifies various complex compound faults. Furthermore, simulated online diagnostic tests confirmed the framework’s rapid response capability to dynamic faults and its computational efficiency, meeting the demands of real-time monitoring. This study offers a precise and robust solution for the fault diagnosis of embedded sensors in asphalt pavement.