Information diffusion prediction is a fundamental yet challenging task in social networks. Existing approaches typically focus on either microscopic or macroscopic prediction, but few effectively integrate both perspectives. Moreover, current models often overlook the dynamic nature of cascades and struggle to capture both diffusion patterns simultaneously. This article addresses this limitation by proposing a novel multiscale information diffusion prediction model that unifies microscopic and macroscopic prediction based on graph attention networks (GATs) and hypergraph neural networks (MS-HGNN). Specifically, MS-HGNN integrates user social features from GATs and global cascade features from HGNN to predict the next affected user. A gated recurrent unit then generates a sequence of predicted users until reaching a virtual terminal user, enabling cascade size estimation. To enhance macroscopic prediction accuracy, we embed the model within a reinforcement learning framework and optimize it using policy gradient methods. Experimental results on four real-world datasets demonstrate that MS-HGNN consistently outperforms state-of-the-art baselines. On average, it achieves approximately 3%–4% improvements in Hits@k and 10%–30% improvements in MAP@k for microscopic prediction across datasets, while also reducing cascade size estimation error in macroscopic prediction.

Accurate prediction of public opinion trends during public health emergencies is crucial for understanding public attitudes and enabling proactive responses. Existing methods frequently exhibit inadequate prediction accuracy and elevated computational complexity in long-term forecasting. The proposed model is an enhanced time series transformer model that incorporates three key innovations. First, a sparse probabilistic attention mechanism reducing spatial complexity from$\mathbf{O(L^{2})}$to$\mathbf{O(LlnL)}$. Second, a progressive sequence decomposition architecture that explicitly separates trend and seasonal components. Third, a global attention distillation technique to mitigate error accumulation in autoregressive prediction. Experiments on a COVID-19 Weibo dataset containing over 780 000 posts demonstrate that the model accurately predicts trends up to seven times the input sequence length. The model outperforms existing methods by over 20% in terms of mean squared error (MSE) and mean absolute error (MAE). For a prediction length of 720, the model achieves an MSE of 0.457 and an MAE of 0.373, effectively capturing key fluctuation patterns and peak timings. The findings establish a substantial technical basis for public health management early-warning systems.

The decision behavior of the leading vehicle driver significantly impacts the safety of the connected mixed platoon, especially on the longitudinal decision during emergency braking events. However, driving abilities vary among leading vehicle drivers. Predictive-forward-collision-warning (PFCW) is an emerging technology that enhances driving safety by providing traffic information beyond the driver’s line of sight. But the influence of PFCW on their decision behavior remains unclear. Therefore, it is necessary to investigate the leading driving competence of heterogeneous driver groups under the influence of PFCW. This study established a test platform for the connected mixed platoon using driving simulation technology. Subsequently, it developed a connected human–machine interface incorporating PFCW functionality and recruited 36 participants to experiment. The experiment aimed to analyze decision behaviors during the emergency braking event involving preceding vehicles. The latent profile analysis model was employed to identify heterogeneous driver groups. Moreover, the cumulative prospect theory was used to characterize drivers’ decision behaviors. The results indicated that PFCW encouraged drivers to maintain a larger distance from the preceding vehicle and to exhibit smoother braking behavior. However, the group of older, skilled drivers who were subjectively concerned about the driving status of following vehicles exhibited higher driving risk as the leading vehicle driver when PFCW was absent. Overall, this study enhances the understanding of longitudinal decision behavioral performance among heterogeneous driver groups in the presence of PFCW. It offers insights for subsequent customization of the selection criteria for leading vehicle drivers, and the tailored design of PFCW based on the characteristics and behavioral preferences of these heterogeneous driver groups.

In this paper, the potential of using a head-up display (HUD) in the connected environment to improve a vehicle’s running comfort and pedestrian safety is tested, by providing warning information to drivers in advance. To achieve this objective, driving simulation technology is used to construct the connected environment and develop the HUD, and the effectiveness of the system is then tested. Specifically, thirty-four participants were recruited to conduct driving simulation experiments in six scenarios: three warning display types (Baseline/Head-down display/Head-up display) combined with two weather conditions (clear weather/foggy weather). The effects of the three different warning display types on braking risk-avoidance strategy were studied by comparing the drivers’ performance during the perception and decision stage (position of accelerator-pedal release, position of first braking), the risk-avoidance manipulation stage (maximum deceleration, braking distance) and the risk-avoidance result stage (minimum collision distance, position of minimum speed). The influences of weather conditions and driver attributes were also considered. When the HUD warnings were activated, drivers started to decelerate further away from pedestrians, with a more stable and moderate deceleration process and a greater safety margin between the vehicle and the pedestrians. Using HUD warnings in foggy conditions improved drivers’ perception and decision abilities, this study confirmed the great benefits that HUD warnings in the connected environment can bring to traffic safety, especially under risky situations and inclement weathers. The research results provide a reference for the more humanized and rationalized optimization design of these warning systems.

The driver-assisted navigation system is an invaluable tool. However, in intricate scenarios, drivers frequently commit navigation errors. To mitigate this issue, this study focuses on the F-type intersection with the highest incidence of deviations as the research subject. Road scenarios are replicated, and driver behavior data is collected through driving simulator technology. Speed and speed standard deviation are indicators for investigating the influence of driveway distance (DD), navigation prompt timing (NPT), and driver attributes on driving efficiency and safety stability using a generalized linear mixed model (GLMM). Findings reveal that excessively large or small driveway distances and navigation messages that are either premature or delayed negatively affect driving efficiency and safety stability. Consequently, it is recommended to adhere to a driveway distance range of 15-30m, accompanied by the prompt mode of {-300m, -150m, Confirmation}. Furthermore, although no random effects of driver attributes were identified, it is essential to recognize that driver attributes heavily influence their driving behavior in complex road scenarios. This study lays the foundation for optimizing the design of road facilities and navigation systems.

The change of space environment at the tunnel entrance which the light changes from bright to dark will lead to frequent traffic accidents. To explore whether the connected vehicle environment will improve the driving performance and traffic safety at the tunnel entrance, this paper first of all designed a driving simulation experiment and provided driving aids through the Human-Machine Interface (HMI). Then 35 drivers were invited to take a driving simulation under two different driving conditions: traditional driving environment (HMI-OFF) and connected vehicle environment (HMI-ON). Finally, the effectiveness and safety of HMI can be evaluated from two aspects of driver response efficiency and deceleration process, and further explore the causes of driving behavior changes by Cox model. The results show that the driver’s response efficiency and deceleration behavior are both improved in the connected vehicle environment. The warning information provided by in-vehicle HMI enables the driver to know the road conditions ahead earlier and adjust the speed through more gentle acceleration so that the vehicle can enter the tunnel at a more stable speed. At the same time, there are significant differences in the brake-movement time of different driving ages, which the brake-movement time increases with the driving age. This indicates that experienced drivers usually adopt a soft way to adjust the speed, so that the vehicle speed change process is more stable on the premise of ensuring driving safety, so as to avoid the potential danger brought by drastic speed change. In this paper, a validity evaluation method based on survival analysis is proposed for tunnel entrance scenarios, which provides a reference for the realization of connected vehicle evaluation.

Glance behavior is significant because whether and how the driver is scanning and observing the driving scene is closely related to driving safety. This paper aims to improve the accuracy of glance behavior modeling and realize the spatiotemporal representation and visualization of glance behavior. Forty subjects were recruited to perform a freeway driving task using a driving simulator. The vehicle data were collected by the simulator. Drivers’ gaze points were collected by an eye tracker. The prior knowledge on gaze points obtained through a statistical analysis were provided for K-means (KM) to form a semi-supervised K-means (SSKM), which classifies gaze points into different fixation zones. The classification results were compared with the results of KM. Furthermore, a clustering center-based piecewise aggregate representation (CCPAR) was proposed to characterize glance behavior. Maneuvers identification was taken as a case to evaluate the proposed method. The k-nearest neighbour (KNN) based on the similarity of CCPAR identified driving maneuvers into lane-keeping, left lane change, and right lane change. The identification results were compared with the results of the Hidden Markov Model (HMM). The average classification accuracies of KM and SSKM were 55.28% and 94.75%, respectively. The accuracies of maneuvers identified by CCPAR-KNN and by HMM were 87.50% and 85.83%, respectively. The results indicate that SSKM and CCPAR are feasible for glance behavior modeling. SSKM eliminates the randomness of initial cluster center selection and improves the accuracy of gaze points classification. CCPAR is intuitive and convenient to describe and visualize the spatiotemporal characteristics of glance behavior.

According to many studies, motor vehicle collisions lead to many deaths and disabilities, and bring about huge economic losses to societies and individuals around the world. If risky scenarios could be recognized and the driver could be timely warned, the frequency of traffic accidents could be reduced effectively, and traffic safety could be improved. This article summarizes the current state of the research on safety warnings under connected vehicle environments-from risk recognition algorithms, collision avoidance systems, dangerous event identification and notification methods, real-time safety warning system and effects of various types of warning information notifications. At the end of this article, the conclusions of this work are presented, and possible future directions for safety warning research under connected vehicle environments are discussed. This article represents the current research status of safety warnings under connected vehicle environments to some extent, which can provide references for future safety warning research in terms of framework, methods and technologies, etc.

Beamlet transform has been widely used for extracting line features from images, which is an excellent multiscale geometric analysis method. However, it has a major drawback that it always performs too slowly due to very much redundant computation. In many application fields, the speed of the original beamlet transform is almost unbearable. To cure the problem, beamlet transform is improved by introducing geometrical flow, which utilises image semantic information in the process of generating beamlets. Besides, to further speed up the algorithm, interesting factor is presented to reduce recursively partitioned boxes. As a result, lots of computation time is saved. Experiments are conducted on various crack images and the results show that the proposed method runs significantly faster than the original beamlet transform. Cracks in an image are detected accurately. Moreover, the proposed method is robustly enough since the performance is hardly affected by crack shape and background texture.