From crash reports to safer roads: a multimodal framework integrating vision-language models and street view analysis.

Exploring novel surrogate safety indicators measuring conflict riskiness and severity: a case study in Sacramento, United States.

Traffic congestion has become one of the most pressing challenges in modern urban environments. Rapid population growth, the surge in private vehicle ownership, and limited infrastructure capacity have all contributed to overcrowded roads and inefficient traffic flow. Traditional traffic management systems, which rely heavily on fixed-time signal control, fail to adapt to dynamic traffic conditions. This often results in prolonged waiting times at intersections, unnecessary fuel consumption, increased air pollution, and a higher risk of accidents. To address these issues, this project proposes a Smart Traffic Management System powered by deep learning, artificial intelligence, and IoT-based sensors. The system continuously monitors, analyses, and regulates traffic in real time. Data is collected from surveillance cameras, IoT sensors, and GPS-enabled devices, providing a holistic view of traffic conditions. Convolutional Neural Networks (CNNs) are employed for vehicle detection, counting, and classification, while Long Short-Term Memory (LSTM) networks enhance traffic flow prediction. By integrating adaptive traffic signal control, the system prioritizes emergency vehicles, reduces congestion, and improves overall road safety. Beyond efficiency, the system aims to create a smoother and more sustainable commuting experience. By reducing idle times and optimizing fuel usage, it directly contributes to lowering carbon emissions and improving air quality. Commuters benefit from shorter travel times, while cities gain a smarter infrastructure capable of evolving with changing demands. Ultimately, this project envisions a future where technology and human needs merge seamlessly—building smarter, greener, and safer cities that enhance the quality of everyday life.

Road traffic injuries remain a leading cause of death and disability among children and adolescents worldwide, particularly in low- and middle-income countries (LMICs), where rapid motorisation and limited trauma care capacity increase vulnerability. In this study, we aimed to characterise long-term patterns and potential future trajectories of the burden of road traffic injuries among children and adolescents aged 0-19 years in LMICs, using estimates from the Global Burden of Disease (GBD) 2023 study. From the GBD 2023 database, we extracted incidence, prevalence, mortality, and disability-adjusted life years (DALYs) for road traffic injuries across 129 LMICs and stratified them by age, sex, and gross national income. We assessed temporal patterns using estimated annual percentage change and joinpoint regression. Further, we used decomposition analysis to illustrate the relative contributions of population growth, age structure, and epidemiological change to the disease burden. We used autoregressive integrated moving average (ARIMA) models for exploratory and scenario-based projections of future trends. Between 1990 and 2023, the overall burden of road traffic injuries among children and adolescents in LMICs declined across DALYs, mortality, incidence, and prevalence. Declines were most pronounced in upper-middle-income and more modest in low-income countries. Motor vehicle-related injuries accounted for the largest share of DALYs across income groups. Males and older adolescents showed higher estimated rates and slower declines. Decomposition analysis indicated that population growth was the primary driver of the increasing burden in low-income countries, whereas epidemiological improvements were primarily observed in upper-middle-income countries. Exploratory extrapolations of ARIMA suggested that DALYs and mortality might continue to decline, while incidence and prevalence might stabilise or increase modestly under unchanged historical trends. Based on estimates from GBD 2023, the burden of road traffic injuries among children and adolescents in LMICs has declined over the past three decades, despite substantial differences across different income, age, and sex groups. These findings should be interpreted as estimated patterns rather than directly observed epidemiological changes. Strengthening road safety, trauma care, and prevention strategies, particularly in low-income settings, is essential to reduce inequality and mitigate the burden of road traffic injuries in children and adolescents.

Colonial historic city centers represent a paradigmatic challenge for modern road safety, as they are characterized by narrow streets originally designed for carriage and pedestrian traffic. This research presents ADAS-TSR, a deep learning-based advanced driver assistance system for vertical traffic sign detection with voice alerts, specifically designed for the Historic Center of Ayacucho, Peru, which is located at 2761 m a.s.l. An original dataset comprising 2250 images with 2450 instances corresponding to 14 sign classes according to Peruvian regulations was constructed. The dataset was captured under real operational conditions, including deteriorated, partially occluded, and vehicle impact-deformed signage. A comprehensive multi-model benchmark experiment was conducted, comparing four CNN-based detectors (YOLOv8m, YOLO11n, YOLO26n, YOLO26s) and one transformer-based detector (RT-DETR-l) spanning both classical and state-of-the-art architectures released through January 2026. YOLO26s achieved the best overall performance, with an mAP@0.5 of 0.994 and mAP@0.5:0.95 of 0.989 while using only 9.5 M parameters. YOLO11n matched the performance of YOLOv8m with 10× fewer parameters (2.6 M vs. 25.9 M). Uncertainty analysis revealed that modern architectures exhibit significantly higher prediction confidence (mean > 0.90) compared to YOLOv8m (0.82), and fairness analysis confirmed equitable detection across all 14 classes (Gini < 0.002). A voice alert system with five priority levels and rule-based temporal filtering for detection stabilization was implemented. Validation across five urban circuits spanning 14.11 km demonstrated a detection rate of 94.7% with a 73% reduction in redundant alerts.

Road safety threats and occupational health risk associated with psychoactive substance use and dependence risk among commercial drivers: WHO ASSIST evidence from Ghana.

Driver behavior remains a leading factor in road accidents, yet existing monitoring systems typically rely on single data modalities such as facial expressions or speech alone, limiting their reliability and contextual awareness. This work proposes a comprehensive driver behavior monitoring system using multimodal AI, uniquely integrating video, audio, and vehicle speed telemetry — an approach that remains underexplored in existing literature — to predict driver emotions and behaviors in real time. The system analyzes facial cues to detect visual anomalies, processes audio inputs to infer emotional states, and incorporates speed telemetry to provide additional behavioral context. This fusion of modalities is designed to improve classification accuracy and reduce false positives compared to unimodal approaches. Performance evaluation is conducted using benchmark datasets for both video-based and audio-based emotion recognition, with comparative analysis between individual and combined modalities. By addressing the challenges of multimodal integration and real-time processing, this research contributes a novel and effective framework for intelligent driver assistance systems, advancing the goal of enhanced road safety through predictive behavioral intervention. Additionally, this research is being extended to incorporate an intermediate-fusion-based multimodal decision framework, wherein top predictions from image, audio, and vehicle telemetry are jointly fed into a decision model (e.g., Random Forest or SVM) for improved context-aware warning generation. This approach addresses prior limitations of ensemble-style fusion and better aligns with the goals of true multimodal AI.

Vehicle traction control is the cornerstone for enhancing driving performance, ensuring road safety, and enabling intelligent driving. For off-road vehicles, variable road roughness imposes varying degrees of vertical dynamic loads on the wheels and affects changes in vehicle traction. Through simulation experiments, this paper found that the slip rate control performance of vehicle traction control systems traditionally designed for static loads exhibited significant variations when dynamic loads were introduced. To address this issue, vertical dynamic wheel load data under different vehicle speeds and road classifications were subjected to clustering, leading to the determination of three as the optimal number of clusters. The integration of the immune whale algorithm for the design of corresponding traction control systems for different dynamic load categories, and their subsequent comparison with traditional controllers, revealed a significant enhancement in slip rate control performance. Under Class I dynamic loads, the control performance on all road surfaces achieved the best results, reaching a maximum of 63.4%. Consideration of the dynamic variations in dynamic loads caused by changes in vehicle speed led to the design of a global controller capable of adapting to different dynamic load conditions, based on the aforementioned classification controller. The final simulation verification confirmed the maintained excellent control performance of this adaptive controller under various operating conditions.

ObjectiveTo develop and validate a computational framework that infers individualized attention strategies and latent distraction states to support personalized modeling of multitasking behavior and intervention.BackgroundDriver distraction from in-vehicle systems is a growing safety concern. However, the level of distraction is often latent and varies significantly across individuals. Existing models typically overlook these differences, limiting their effective use for personalized interventions.MethodWe introduce a Partially Observable Semi-Markov Decision Process (POSMDP) to model hidden attentional dynamics and attention allocation decisions. Using behavioral data, including glance behavior, velocity, and pupillometry, from a high-fidelity driving simulator with 18 participants, we estimate personalized reward functions that reflect each driver's subjective valuation of secondary task utility versus safety cost.ResultsThe method accurately infers distraction states and recovers participant-specific utility weights governing the trade-off between secondary task benefit and driving cost. Compared to a well-established 2-s glance rule, it improves detection of distraction events and reveals individual variability in attention strategies. Some drivers exhibit highly conservative profiles, while others assign greater value to secondary tasks, even under high distraction. Counterfactual simulations show how perceived task importance could modulate visual attention behavior across individuals.ConclusionOur POSMDP-based framework provides an interpretable and individualized account of driver attention allocation, capturing both latent states and behavioral variability.ApplicationThis model enables the development of personalized, risk-sensitive driver assistance systems that adapt to individual attention strategies, enhancing road safety through context-aware, graded interventions.

Speed management plays a critical role in road safety; however, conventional speed limits are determined based on characteristics such as geometry and traffic volume. Limited consideration is given to the structural condition of pavements and surface distress. This study proposes an integrated mechanistic–quantitative framework that links pavement distress and road safety indicators to the selection of speed limits. A flexible pavement section on Highway No. 80 in Iraq is analyzed as a case study. Mechanistic pavement analysis using KENPAVE is employed to estimate critical strains based on field traffic data and Equivalent Single-Axle Loads (ESALs). The rate of failure is estimated by comparing ESALs and the allowable load repetitions. Safety-related constraints are then derived to quantify hydroplaning risk, braking performance through stopping sight distance, and the vertical shock criterion. The results indicate that the existing pavement structure is marginal, with a high probability of fatigue failure and sensitivity to rutting under increased traffic loads. The integrated safety analysis yields a critical wet-weather speed of approximately 67–70 km/h, while localized settlements exceeding 10 mm require speed reductions of 50–60 km/h to maintain vehicle stability. The proposed framework demonstrates that pavement conditions directly influence safe speed, providing a rational basis for safety-oriented speed management.

Road safety has become an increasingly important concern and the integration of Advanced Rider Assistance Systems and Advanced Driver Assistance Systems plays a crucial role in preventing accidents. This work proposes a computer vision pipeline to automatically detect hazardous road anomalies—loose gravel, potholes, and puddles—from a motorcycle-mounted camera, targeting real-time operation on embedded edge devices. A hybrid dataset of 28764 annotated images was created by combining real-world photos, Blender-rendered synthetic scenes, and AI-generated images to improve diversity and coverage. Multiple state-of-the-art object detectors were trained and benchmarked, including the YOLOv5/7/11/12 families and the transformer-based RT-DETR architecture. While the RT-DETR model achieved the highest precision overall, its computational complexity and heavy resource requirements limited its suitability for real-time deployment on low-cost embedded platforms. Conversely, the YOLOv11n model demonstrated the best accuracy–efficiency trade-off, reaching mAP@0.5 = 0.872 at 320$\times$320 with 0.045 s/frame on a Jetson Nano, while lighter variants remained viable on Raspberry Pi boards. Across classes, gravel was the most reliably detected, and operating points around a confidence threshold of $\tau \approx 0.31$ yielded balanced F1 scores up to 0.82. Although results show that automatic road-condition monitoring on affordable hardware is feasible, the prototype has not yet undergone on-road field trials. It does not include an integrated rider alert module or energy-use assessment. These gaps define the immediate roadmap for deployment.

In recent years, automotive weight reduction has attracted considerable attention due to its benefits in fuel consumption, emissions, material usage, and vehicle dynamics. For unsprung masses, these effects are particularly pronounced, directly influencing vehicle stability, maneuverability, and road safety. Conventional engineering optimization is typically based on static load cases; however, such "simple" optimization is insufficient for safety-critical components operating under real service conditions.In practice, automotive components are exposed to dynamically varying, stochastic loads originating from road excitation, and their failure is therefore predominantly governed by fatigue rather than static strength. Current engineering optimization tools do not yet enable direct optimization with respect to fatigue life. To address this limitation, a dynamic factor is introduced to represent time-dependent loading effects within the optimization framework. The optimization problem is reformulated with the explicit constraint that the original safety factor must not decrease, ensuring that the expected service life of the component is preserved.The results indicate that, although the achievable mass reduction is smaller than that obtained by purely static optimization, it remains significant while maintaining fatigue-related safety margins. The applied approach is restricted to geometry modifications compatible with conventional manufacturing, ensuring industrial relevance.

This study examines temporal changes from 1990 to 2021 in the burden of vertebral fractures (VFs) attributable to motor vehicle road injuries (MVRIs), with a particular focus on age- and sex-specific patterns in China and India. These national trends are compared with global patterns to better understand population distribution characteristics and injury mechanisms underlying this public health challenge. Data were obtained from the 2021 Global Burden of Disease (GBD) study. Crude rates and age-standardized rates (ASRs) of incidence, prevalence, and years lived with disability (YLDs) for MVRI-related VFs were estimated. Joinpoint regression was applied to assess temporal trends, while age-period-cohort (APC) modeling was used to disentangle the independent effects of age, calendar period, and birth cohort. From 1990 to 2021, the global age-standardized incidence rate (ASIR) of MVRI-related VFs declined by 48.8%, with an average annual percentage change (AAPC) of -1.839% (95% confidence interval [CI], -1.869 to -1.808). In contrast, China showed no significant reduction in ASIR (AAPC = -0.478%, 95% CI, -0.531 to -0.426), whereas India demonstrated minimal variation over the study period (AAPC = -0.013%, 95% CI, -0.054 to 0.028). Regional analyses revealed heterogeneous drivers of disease burden. In China, period effects during 2000-2021 were strongly associated with elevated risk among males aged 20-40 years, likely reflecting hazardous driving behaviors, while cohort effects were most prominent among individuals born between 1980 and 1990. Conversely, individuals older than 60 years experienced an increasing burden, potentially related to osteoporosis and rapid motorization. Across all regions, males consistently exhibited higher ASIRs, age-standardized prevalence rates (ASPRs), and YLD rates than females, with the greatest sex disparities observed among younger males in China. The persistently high burden of MVRI-related VFs in China, which diverges from declining trends observed in countries with a high sociodemographic index (SDI), highlights the need for targeted prevention strategies. Interventions should prioritize behavioral risk reduction in younger male populations and address age-related biomechanical vulnerability in older adults. In India, strengthening road safety enforcement and trauma care infrastructure remains essential. These findings underscore the heterogeneous demands for road injury prevention in China and India and provide evidence to support more effective allocation of public health resources.

Scoping review on motorcycle crashes patterns, risk factors, and potential in setting policy priorities in the gulf cooperation council countries (GCC).

From work environment to roadway: A narrative review on organizational psychology’s role in road safety

From thoughts to crashes: A multilevel analysis of driver psychology, moral disengagement, and road safety

The growing use of video surveillance systems requires reliable communication networks capable of supporting high-quality real-time streaming. This study evaluates the performance of a wireless network designed for CCTV transmission using the NS-3 simulation platform. The proposed network follows a star topology and consists of four communication links with different distances and bandwidth capacities to examine their behavior under varying traffic loads. This work is part of a broader research effort aimed at designing an efficient communication network to support road safety monitoring and traffic surveillance in mountainous urban environments. It also seeks to provide technical insights that may assist local authorities in developing reliable monitoring infrastructures and improving traffic management. The evaluation focuses on key Quality of Service (QoS) metrics, including throughput, delay, jitter, and packet loss. The results show that higher-capacity links maintain stable throughput and low delay under increased traffic load, while longer-distance links begin to experience congestion as the offered traffic approaches their capacity limits. These findings highlight the importance of link capacity in network performance and confirm that a properly designed star-based wireless infrastructure can effectively support traffic surveillance systems in geographically challenging environments.

Reducing private car use and increasing active travel is essential for transport decarbonisation and addressing public health crises of road traffic injuries, physical inactivity and air pollution. Low Traffic Neighbourhoods have emerged as a key intervention, particularly in London, United Kingdom, to create better environments for walking, wheeling and cycling by restricting through traffic on residential streets. While evidence suggests that Low Traffic Neighbourhoods reduce car use, increase walking and cycling, and improve road safety, their implementation has been politically contentious and has elicited a wide range of public reactions. This paper presents findings from the qualitative strand of a wider 3.5-year mixed-methods study of Low Traffic Neighbourhoods in London. Qualitative data were collected to explore the lived experiences of Low Traffic Neighbourhood residents, with a focus on how residents - including disabled residents - perceive and navigate the schemes. Participants were selected from among those living in or on the road adjacent to four selected Low Traffic Neighbourhoods, to ensure a diversity of views on the schemes, and diverse demographic characteristics were represented. Using 61 go-along interviews and 7 focus groups, we explore how Low Traffic Neighbourhoods influence residents' experiences and perceptions of travel. Our findings show that residents' attitudes towards Low Traffic Neighbourhoods often shape their reported experiences: those who are initially supportive tend to notice and highlight positive impacts, while opponents are more likely to report no change or negative impacts. Overall, participants observed increased walking and cycling, improved perceptions of road safety, and reduced noise and air pollution within Low Traffic Neighbourhoods. However, concerns were raised by some disabled residents about longer journey times and accessibility problems. Notably, many residents living on boundary roads perceived an increase in traffic and pollution, although quantitative data on the impacts of Low Traffic Neighborhoodson boundary roads remain mixed. Our findings highlight the importance of considering residents' lived experiences in scheme evaluations. While Low Traffic Neighbourhoods contribute to climate and health objectives, their implementation should be guided by a just transition framework to ensure benefits are equitably distributed. Our study's main limitation is that it contributes to an evidence body of research on Low Traffic Neighbourhoods being predominantly from London. Future research should extend beyond London, explore a wider range of schemes and attend to experiences of further marginalised population groups. This article presents independent research funded by the National Institute for Health and Care Research (NIHR) Public Health Research programme as award number NIHR135020.

The causal effects of removing Hook-Turn regulation on road safety

This study aims to evaluate the effectiveness of acoustic stimulation in enhancing driver's vigilance, improving driving performance, and preventing inattentive driving on expressways. While numerous studies have investigated the effects of acoustic stimuli on drivers' attention, the influence of different types of stimuli on sustained attention and driving behavior remains unclear. Particularly, the habituation effect to the stimuli during driving has not been investigated. In this study, several types of acoustic stimuli-monotone sounds, verbal messages, and emotional sounds-are examined as potential countermeasures for vigilance decrement. Their effects on inattentive driving and the stability of driving behavior are assessed and compared with a control condition using a driving simulator (DS) experiment in which each participant was asked to continuously operate a DS for 20min under each experimental condition. The experimental results with 30 participants reveal that: (1) acoustic stimulation initially produces an awakening effect, but its effectiveness tends to decline over time due to habituation; and (2) among the three types of acoustic stimuli tested, emotional sounds have a stronger and more sustained effect on maintaining driver alertness, showing less susceptibility to habituation than monotone or verbal stimuli. These findings suggest that emotional acoustic stimuli may serve as a promising basis for the development of in-vehicle or infrastructure-based systems aimed at preventing inattentive driving and improving road safety.