In this work, to clarify the impact of electric bicycle drivers’ risky driving behavior on driving safety, we used multiple regression analysis methods combined with a questionnaire survey of residents of the city of Guilin, China. We studied the impact of the two dimensions of safety knowledge and safety attitude on risky driving behavior, and identified the differences in the impact of these two dimensions from the perspective of personal characteristics. Through modeling analysis, we found that “responsible attitude” and “group behavior attitude” explain 62.4% of the variation in aggressive behavior; 48.5% of the variation in negligent behavior is caused by “age”, “safety knowledge” and “responsible attitude”; and 52% of the variation in violations is caused by “age”, “violation attitude” and “group behavior attitude”. The results show that “group behavior attitude” affects the occurrence of aggression; that safety knowledge has a significant negative impact on unintentional negligence but has no significant effect on deliberate violations and aggression; and that the difference in risky driving behavior is mainly manifested in “age”, “gender”, “violation” and “accident experience”.

In the intricate urban road scenarios, the algorithm for vehicle speed measurement based on image sensors encounters challenges of low precision. Often, a confluence of video and radar solutions is employed, impeding large-scale applications in the engineering domain. By amalgamating deep learning algorithms and camera imaging principle, a speed measurement algorithm was crafted that integrates scene and object features. Initially, the YOLOv5s object detection model was employed, and refined the model’s backbone and loss function to elevate the precision of vehicle detection. Subsequently, in conjunction with the RAFT algorithm and camera imaging principle, the geometric relationship between road scenes and images was established, deriving vehicle speed based on the coordinates of vehicle’s feature points. Furthermore, a vehicle dataset was curated to train and assess the improved YOLOv5s model. Finally, employing an experimental vehicle traversed the designated video detection section at five constant speeds, to validate the efficacy of the speed measurement algorithm. Experimental results illustrate that, in comparison to the original model on the self-established dataset, the improved YOLOv5s demonstrates enhancements of 1.7% and 1.4% in mAP@0.5 and mAP@0.5:0.95, respectively. The mean absolute error and root mean square error for all speed measurement values are 1.95 km/h and 2.38 km/h, respectively. In summary, this algorithm is applicable to vehicle speed measurement tasks in complex traffic environments, showcasing better precision.

The outbreak of multiple disaster sites during the coronavirus disease 2019 (COVID-19) pandemic has presented challenges due to varying access time intensity, population density, and medical resources at each site. To address these issues, this study focuses on 13 districts and counties in Wuhan, China. The importance of each research area is analyzed using the improved PageRank and TOPSIS algorithms to determine the optimal site selection plan. Additionally, a particle swarm algorithm is used to construct an emergency material dispatching model that targets both distribution and site selection costs to solve the multi-distribution center dispatching problem. The results suggest that constructing 10 distribution centers can satisfy the demand for epidemic prevention and control in Wuhan city while saving costs associated with site selection and material distribution. Compared to the previous optimal solution, the distribution and site selection costs under the optimal solution decreased by 27.9% and 17.82%, respectively. This approach can serve as a basis for dispatching emergency materials during public health emergencies.

Road-net was regarded as a big dynastic system, in which operation is affected and restrained by lots of factors. Among of them, lane changing has peculiar characteristic and takes greatly effect to traffic flow in jam condition. A kind of lane changing model was constructed to analyze two types of lane changing based on the game theory in jamming. The maneuver of driver's lane changing was described according to traffic information and experiences, and a model of driver's cooperation was introduced based on time and load security. At last, this paper analyzed the mechanism of lane changing in jam condition, carried out a theory deduction from time and space angle, gave some valuable examples, and got the intent point of lane changing

Taking the 2003–2005 highway accident data surveyed from Hada Highway and Hashuang Highway into consideration, rear-end collisions are divided into two patterns: rear-collisions and stationary object collisions. The principal analysis identified the impact level of different factors on two types of rear-end collisions. Then, significant impact factors are detected and identified by charting analysis and clustering research. The final investigation result indicates that fatigue, excessive speed, unsafe following distance, congestion et al, are main impact factors on rear collision accidents. However, excessive speed, fatigue, illegal parking, et al, have a significant effect on the occurrence of stationary object collisions, which provide a reference for highway management departments.

Accurate vehicle type recognition in low-light environments remains a critical challenge for intelligent transportation systems (ITSs). To address the performance degradation caused by insufficient lighting, complex backgrounds, and light interference, this paper proposes a Twin-Stream Feature Fusion Graph Neural Network (TFF-Net) model. The model employs multi-scale convolutional operations combined with an Efficient Channel Attention (ECA) module to extract discriminative local features, while independent convolutional layers capture hierarchical global representations. These features are mapped as nodes to construct fully connected graph structures. Hybrid graph neural networks (GNNs) process the graph structures and model spatial dependencies and semantic associations. TFF-Net enhances the representation of features by fusing local details and global context information from the output of GNNs. To further improve its robustness, we propose an Adaptive Weighted Fusion-Bagging (AWF-Bagging) algorithm, which dynamically assigns weights to base classifiers based on their F1 scores. TFF-Net also includes dynamic feature weighting and label smoothing techniques for solving the category imbalance problem. Finally, the proposed TFF-Net is integrated into YOLOv11n (a lightweight real-time object detector) with an improved adaptive loss function. For experimental validation in low-light scenarios, we constructed the low-light vehicle dataset VDD-Light based on the public dataset UA-DETRAC. Experimental results demonstrate that our model achieves 2.6% and 2.2% improvements in mAP50 and mAP50-95 metrics over the baseline model. Compared to mainstream models and methods, the proposed model shows excellent performance and practical deployment potential.