Due to the widespread use of unmanned aerial vehicles (UAVs) in remote sensing, there are fully developed techniques for extracting vehicle speed and trajectory data from aerial video, using either a traditional method based on optical features or a deep learning method; however, there are few papers that discuss how to solve the issue of video shaking, and existing vehicle data are rarely linked to lane lines. To address the deficiencies in current research, in this study, we formulated a more reliable method for real traffic data acquisition that outperforms the traditional methods in terms of data accuracy and integrity. First, this method implements the scale-invariant feature transform (SIFT) algorithm to detect, describe, and match local features acquired from high-altitude fixed-point aerial photographs. Second, it applies “you only look once” version 5 (YOLOv5) and deep simple online and real-time tracking (DeepSORT) to detect and track moving vehicles. Next, it leverages the developed Python program to acquire data on vehicle speed and distance (to the marked reference line). The results show that this method achieved over 95% accuracy in speed detection and less than 20 cm tolerance in vehicle trajectory mapping. This method also addresses common problems involving the lack of quality aerial photographic data and accuracy in lane line recognition. Finally, this approach can be used to establish a Frenet coordinate system, which can further decipher driving behaviors and road traffic safety.

Road infrastructure has significant effects on road traffic safety and needs further examination. In terms of traffic crash prediction, recent studies have started to develop deep learning classification algorithms. However, given the uncertainty of traffic crashes, predicting the traffic risk potential of different road sections remains a challenge. To bridge this knowledge gap, this study investigated a real-world expressway and collected its traffic crash data between 2013 and 2020. Then, according to the time-spatial density ratio (Pts), road sections were assigned into three classes corresponding to low, medium, and high risk levels of traffic. Next, different classifiers were compared that were trained using the transformed and resampled feature data to construct a traffic crash risk prediction model. Last, but not least, partial dependence plots (PDPs) were employed to interpret the results and analyze the importance of individual features describing the geometry, pavement, structure, and weather conditions. The results showed that a variety of data balancing algorithms improved the performance of the classifiers, the ensemble classifier superseded the others in terms of the performance metrics, and the combined SMOTEENN and random forest algorithms improved the classification accuracy the most. In the future, the proposed traffic crash risk prediction method will be tested in more road maintenance and design safety assessment scenarios.

In order to create a practical road-resistance function for work zones under different lane occupation conditions, the expected speed of vehicles was calibrated in the work zone simulation model based on measured data, and simulation models were constructed for the closed half lane and the closed inside lane under different rates of trucks. Based on the statistical theory, the influence of significance of traffic volume and truck ratios for road resistance was analyzed, and a suitable truck ratio was found for the work zone. By using the optimal nonlinear fitting theory, the practical road-resistance function for work zones under different lane occupation conditions was constructed. The results showed that the road resistance is significantly affected by the traffic volume and rate of trucks. Under the same truck ratio, the road resistance linearly increases slowly when the traffic volume is less than the critical traffic volume and rapid increases irregularly when it is greater than the critical traffic volume. Under the same traffic load, the road resistance of the work zone increases with the increase in the rate of trucks, and the difference is not obvious when the traffic volume is less than the critical traffic volume, and increases gradually when it is greater than the critical traffic volume. Through the goodness of fit test and the homogeneity of variance test, the road-resistance function constructed in this paper has high goodness of fit. The practical road-resistance functions constructed in this study could be used to guide the diversion of the rebuilt/expanded highway to ensure traffic safety. Further, the study provides a theoretical basis for the construction of intelligent highway work zones.

Due to the different driving characteristics of different vehicle models, inappropriate mandatory lane changes (MLCs) by heavy vehicles at interchanges often lead to serious traffic accidents. Therefore, this paper focuses on the impact of road geometric design on the MLC duration of heavy trucks by using full time-domain trajectory data. Specifically, we use the generalized additive time-varying Cox model to establish the MLC duration model of heavy trucks at interchanges, then analyze the combined influence of geometric elements. The results show that the consistency index of the model is 0.9, indicating that it has advantages in building models in complex environments. The length of the deceleration lane, ramp type, and curve radius have a significant impact on the validity and duration of MLCs. This finding provides a theoretical and methodological reference for the safety analysis of interchange areas and the refinement of road geometric design.

The traditional method of designing freeway geometric characteristics may not well consider traffic safety and evaluation problems associated with vehicle, driver, road, and environmental features. This paper looks into accident-prone compound curves and puts forward a traffic safety evaluation method that can comprehensively assess various influencing factors. This method integrates driving simulation and Virtual Reality (VR) technology. Methodically, this paper first used three-dimensional (3D) design software to build the digital model and spatial scene model of the compound curves from the aspects of geometric structure, spatial characteristics, terrain information, and so on. Next, drivers were invited to conduct a series of driving simulation experiments upon the human–computer interaction safety experience platform, and driver physiological data and vehicle driving information collected. Last but not least, the mean values of heart rate changes, steering wheel angle changes, and driving trajectory changes were derived to synthesise the comprehensive traffic safety evaluation framework. Based on the analysis of the results of the orthogonal test, select the road plane design indicators that have a significant impact on the traffic safety evaluation, and carry out regression analysis. The research shows that the novel traffic safety evaluation method integrating with VR technology can comprehensively consider various influencing factors, and designers can dynamically adjust the design metrics according to the traffic safety level.