Analysis of pavement deterioration is critical for road maintenance. Many section-based pavement performance evaluation methodologies have been investigated to determine the deteriorating tendency from a macro perspective. However, little research shed light on the refined deterioration analysis for single distress, which is valuable for daily and preventive maintenance. This paper proposed a deep-learning-based tracking framework to construct a large-scale continuous observation data set for every distress. A deep learning model is applied to detect six types of distress automatically. Then we adopted the spatial clustering method to match the pavement images in the same scene. Finally, image feature matching and perspective conversion methods are adopted to track the distress in the same scene. Using the data collected from the bus driving recorder, we have realized the daily observation of over 270 kilometers of the urban road network. More than 14,000 pavement distress have been continuously tracked, proving this framework’s effectiveness. In addition, the features of pavement deterioration are further discussed. The results show that heavy rain will significantly accelerate road surface deterioration. Under its influence, an intact pavement may suddenly deteriorate into serious potholes within a day. The established continuous pavement distress tracking dataset is significant for distress-level performance prediction research.

In this paper, a gated recurrent unit–deep neural network (GRU-DNN) model integrated with multimodal feature embedding (MFE) is developed to evaluate the real-time risk of hazmat road transportation based on various types of data for contributing factors. MFE was incorporated into the framework of a deep learning model in which discrete variables, continuous variables, and images were uniformly embedded. GRU is a pre-trained sub-model, and the DNN is able to directly use the relative structure and weights of the GRU, improving the poor classification and recognition results due to insufficient samples. Additionally, the model is trained and validated based on hazmat road transportation database consisting of 2100 samples with 20 real-time contributing factors and four risk levels in China. The accuracy (ACC), precision (PR), recall (RE), F1-score (F1), and areas under receiver-operating-characteristic curves (AUC) of the proposed model and other commonly used models are compared as performance measurements in numerical examples. Finally, Carlini & Wagner attack and three defenses of adversarial training, dimensionality reduction and prediction similarity are proposed in the training to improve the robustness of the model, alleviating the impact of noise and error on small-sized samples. The results demonstrate that the average ACC of the model reaches 93.51% and 87.6% on the training and validation sets, respectively. The prediction of accidents resulting in injury is the most accurate, followed by fatal accidents. Combined with the RE of 89.0%, the model exhibits excellent performance. In addition, the proposed model outperforms other widely used models based on the overall comparisons of ACC, AUC, F1 and PR-RE curve. Finally, prediction similarity can be used as an effective approach for robustness improvement, with the launched adversarial attacks being detected at a high success rate.

Deep learning has significantly improved the recognition efficiency and accuracy of ground-penetrating radar (GPR) images. A significant number of weight parameters need to be specified, which requires lots of labeled GPR images. However, obtaining the ground-truth subsurface distress labels is challenging as they are invisible. Data augmentation is a predominant method to expand the dataset. The traditional data augmentation methods, such as rotating, scaling, cropping, and flipping, would change the GPR signals’ real features and cause the model’s poor generalization ability. We proposed three GPR data augmentation methods (gain compensation, station spacing, and radar signal mapping) to overcome these challenges by incorporating domain knowledge. Then, the most state-of-the-art model YOLOv7 was applied to verify the effectiveness of these data augmentation methods. The results showed that the proposed data augmentation methods decrease loss function values when the training epochs grow. The performance of the deep learning model gradually became stable when the original datasets were augmented two times, four times, and eight times, proving that the augmented datasets can increase the robustness of the training model. The proposed data augmentation methods can be used to expand the datasets when the labeled training GPR images are insufficient.

The new generation of smart highway (NGSH) has become an irresistible global trend to improve transport efficiency and safety. The exploration of the features and framework for NGSH can guide us to upgrade the current highway system. This paper summarizes the fundamental features of the NGSH from the perspective of the interactive evolution of automobile industry and road transport. In line with the popularity of automated and connected vehicles, the primary technical features of the NGSH are proposed as (I) complete elements sensing, (II) cyber-physical systems, (III) cooperative vehicle-infrastructure applications, and (IV) 5th generation mobile communication technology. The corresponding physical framework and data flow are introduced, in which three data attributes (data accuracy, dimensionality, and freshness) are highlighted to describe the data requirements for various scenarios. The development path of the NGSH is further discussed in terms of the different vehicle automation levels. The characteristics of five levels of NGSH are identified from R1 to R5. Different combinations of NGSH level and vehicle automation level lead to distinct system functions. Several urgent problems in the current stage are pointed out in terms of system compatibility, standard specification, and information security. This paper provides new insights for sustainable and reproducible highway reformation, drawing some implications for NGSH design.

Pavement roughness serves as a crucial indicator for evaluating road performance. However, traditional measurement methods, such as laser detection vehicles, are limited to providing roughness values for a single profile, failing to capture the overall pavement condition comprehensively. To address this limitation, this study utilized high-precision light detection and ranging technology (LiDAR) to acquire three-dimensional point cloud data for a 25 km road section in Shanghai. Road elevations were extracted from different lateral survey lines. Subsequently, variance analysis and the Kruskal–Wallis non-parametric test were conducted to evaluate the differences in the lateral distribution and longitudinal variability of the pavement roughness. The findings revealed significant differences in the international roughness index (IRI) among the survey lines within the road section. Moreover, the observed variations in the lateral distribution of pavement roughness were influenced by the characteristics of the road section itself. Roads exhibiting discrete roughness patterns displayed a higher likelihood of significant detection disparities. Additionally, it was discovered that the discrepancy between the detection length and the actual road length introduced volatility in repeated detection results, necessitating a limitation of this discrepancy to 30 m. Consequently, it has been recommended to consider the lateral distribution of pavement roughness and to regulate the detection length in road performance evaluations to enhance reliability and facilitate more accurate maintenance decision making. The study highlights the importance of incorporating comprehensive assessment approaches for pavement roughness in road management practices.

Pavement condition prediction plays a vital role in pavement maintenance. Many prediction models and analyses have been conducted based on long-term pavement condition data. However, the condition evaluation for road sections can hardly support daily routine maintenance. This paper uses high-frequency pavement distress data to explore the relationship between distress initiation, weather, and geometric factors. Firstly, a framework is designed to extract the initial time of pavement distress. Weather and geometric data are integrated to establish a pavement distress initiation dataset. Then, the time-lag cross-correlation analysis methods were utilized to explore the relationship between distress initiation and environmental factors. In addition, the logistic regression model is used to establish the distress initiation prediction model. Finally, Akaike information criterion (AIC), Bayesian information criterions (BIC), and areas under receiver operating characteristic curves (AUC) of logistic regression models with or without time-lag variables are compared as performance measurements. The results show that pavement distress initiation is susceptible to weather factors and location relationships. Daily total precipitation, minimum temperature, and daily average temperature have a time delay effect on the initiation of the pavement distress. Distress initiation is negatively correlated with the distance from the nearby intersection and positively correlated with adjacent distresses. The weather factors, considering the time-lag effect, can improve the model performance of the distress initiation prediction model and provide support for emergency management after severe weather.