Accid Anal Prev [JOURNAL]
Sun
200 papers
Transformer-based multi-agent traffic simulation for autonomous vehicle testing in shared urban road segments.
Wei S, Ni Y, Sun J
Accid Anal Prev · 2026 Jul · PMID 41996737 · Publisher ↗
Microscopic traffic simulation plays a crucial role in the development and testing of autonomous driving systems. However, accurately reproducing traffic participant behavior in shared urban road segments remains challen...
Microscopic traffic simulation plays a crucial role in the development and testing of autonomous driving systems. However, accurately reproducing traffic participant behavior in shared urban road segments remains challenging due to their flexible movement characteristics, particularly for bicycles that potentially interfere with motorized vehicles. This study presents a novel data-driven simulation model that integrates a Transformer-based neural network architecture, trained through imitation learning, with a Markov Decision Process (MDP) formulation. By leveraging a Transformer-based multi-agent policy, the model jointly controls the behaviors of road users, effectively capturing complex multi-agent interactions. The proposed similarity reward function enables comprehensive capture of both trajectory and behavioral features. The MDP-based training ensures consistent and realistic traffic behaviors over long-term simulations. Validation experiments demonstrate the model's effectiveness with a Mean Distance Error (MDE) of 2.123 m over 9.6-second simulations, closely matching real behavioral distributions and achieving an F-1 score of 0.865 for interference scene reproduction. Our proposed scene-centric Transformer policy demonstrates superior computational efficiency, operating 3.8 to 5.4 times faster than agent-centric models, with an inference time of 5.98 ms for 20 agents, meeting real-time processing requirements. Comparative analysis reveals that our MDP-based Transformer model significantly outperforms non-MDP alternatives, reducing MDE by 15.8-45.6% and improving interference behavior reproduction F-1 scores by 10.2-18.7%. Furthermore, validation across diverse road segments demonstrates the model's adaptability to varied urban road environments. This approach enhances the realism of microscopic traffic simulation, improving the reliability of simulation platforms for autonomous vehicle testing.
Personalizing In-Vehicle warnings: A causal machine learning approach to optimizing workload and risk perception.
Peng Z, Huo Y, Wang C … +2 more , Easa SM, Zhao X
Accid Anal Prev · 2026 Jul · PMID 41990544 · Publisher ↗
The effectiveness of in-vehicle Connected Information (CI) is often limited by uniform warning strategies that overlook the interaction among warning design, traffic context, and driver state. This study establishes a ca...
The effectiveness of in-vehicle Connected Information (CI) is often limited by uniform warning strategies that overlook the interaction among warning design, traffic context, and driver state. This study establishes a causal machine learning framework to quantify how CI modality and lead time influence driver workload (WL) and perceived risk (PR) across multiple traffic scenes. A 3 × 3 × 5 factorial driving-simulator experiment collected multimodal physiological, behavioral, and self-reported data from 52 drivers. Double Machine Learning with Causal Forests (DML-CF) was applied to identify both average and heterogeneous causal effects. The results show strong context dependence and clear nonlinear patterns in CI effectiveness. Moderately early warnings in the range of 5.5 to 6.5 s combined with congruent dual-modality cues consistently reduced WL and PR. Delayed or single-modality cues frequently increased cognitive demand and perceived hazard. Conditional treatment effect (CATE) analyses revealed substantial heterogeneity and identified driver subgroups with distinct physiological and behavioral characteristics. For instance, certain individuals exhibited elevated WL due to heightened arousal even when exposed to multimodal early warnings that were beneficial for most drivers. These findings indicate that optimal CI design requires continuous alignment between warning parameters, situational uncertainty, and the driver's state. The evidence provides a quantitative causal basis for developing adaptive CI systems capable of tailoring information delivery to enhance safety and human-machine interaction.
AI Conflict Observer: Conflict severity and scenario identification for intersection based on Ensemble Transformer.
Luo G, Wang X, Zang J … +4 more , Zhou Y, Zou Y, Wang C, Ni D
Accid Anal Prev · 2026 Jul · PMID 41990543 · Publisher ↗
The application of perception and edge computing technologies provides extensive data support for intersection conflict analysis. However, traditional threshold-based and semantic rule-based conflict analysis methods str...
The application of perception and edge computing technologies provides extensive data support for intersection conflict analysis. However, traditional threshold-based and semantic rule-based conflict analysis methods struggle to address the challenges posed by intersection heterogeneity and data diversity. This study integrates the kinematic features of trajectory data from eight different Holographic Intersections. Based on the proposed definitions for conflict severity and scenarios, 5,339 typical conflict events were labeled as the training dataset. Subsequently, four Transformer encoders with different combinations of heads and layers were trained, then integrated using the Weighted Voting method. The Ensemble Transformer was selected as the benchmark to construct the AI Conflict Observer (AICO) model. For the tasks of classifying conflict severity and scenarios, the Weighted F1 Scores of AICO reached 0.846 and 0.902, respectively. To validate the model's generalization performance, this study conducted case studies using conflict events from a 4-leg intersection and a 3-leg intersection sourced from different datasets as ground truth. A total of 560 and 136 conflict events were identified through threshold-based preliminary screening and manual verification, respectively. The conflict recognition results of the AICO model were then compared with the ground truth. The results indicate: 1) At the 4-leg intersection and the 3-leg intersection, AICO achieved classification accuracies of 89.97% and 86.13% for conflict severity, and 88.71% and 90.97% for conflict scenario, respectively. 2) Regarding spatial distribution, the kernel density values of conflicts identified by AICO were highly consistent with the ground truth. 3) For temporal distribution, AICO's results demonstrated a high degree of goodness-of-fit. The R values for common and serious conflicts were 0.927 and 0.987 at the 4-leg intersection, and 0.934 and 0.945 at the 3-leg intersection, respectively. 4) In terms of conflict severity distribution, there was no significant difference between AICO and the ground truth in identification of TTC. However, significant differences were observed in conflict duration identification (p < 0.05). AICO employed a more conservative strategy, tending to assign longer durations. The AICO model overcomes the limitations of traditional threshold and rule-based methods. It can generalize to different types of signalized and unsignalized intersections and achieve batch conflict identification. The model can be applied to near-real-time analysis of intersection operational risks, extraction of critical risk scenarios, providing decision support for intersection improvement, governance, and management.
Empirical modeling of vehicle-bicycle interactions on high-speed rural roads: passing lateral distance and speed.
Claros B, Chitturi M, Bill A … +1 more , Noyce DA
Accid Anal Prev · 2026 Jul · PMID 41990542 · Publisher ↗
Driving behavior and interactions with bicyclists on rural roads have not been quantified and modeled extensively. Naturalistic bicycling data for 1,991 passing events were collected on a rural two-lane roadway (55 mph,...
Driving behavior and interactions with bicyclists on rural roads have not been quantified and modeled extensively. Naturalistic bicycling data for 1,991 passing events were collected on a rural two-lane roadway (55 mph, 88 kph speed limit) to quantify how opposing traffic and vehicle platooning influence passing lateral distance, speed, and aerodynamic forces. Results indicate that opposing traffic significantly reduces passing lateral distance by an average of 2.0 ft (61 cm) and decreases speed by an average of 2.3 mph (3.7 kph). Platooning leads to progressively reduced passing distance and speed among following vehicles. The reductions reflect limited available space and increased risk for bicyclists when opposing vehicles are present. The estimated aerodynamic lateral forces created by passenger vehicles were well below tolerable safety limits for bicyclists. To surpass tolerable limits, passenger vehicles would have to pass at a lateral distance of 0.9 ft (27 cm) at a speed of 55 mph (88 kph). Lateral distance and speed were found to be independent at a disaggregate level. Leading vehicles' lateral distance followed a Log-normal distribution and speed followed a Weibull distribution. Theoretical joint probability density functions were developed for leading and following vehicles with and without opposing traffic. Pairwise differences among lead and follower vehicles were similar and resembled a Normal distribution. The developed joint probability density functions can be used for calibration and validation of driving simulators, or development of autonomous and artificial intelligence driving models. Results contribute to developing safer design guidance and risk mitigating strategies for bicyclists.
The nonlinear impact of road safety policy implementation on the severity of road traffic crashes: A fusion of deep learning and Bayesian random parameter methods.
Shen R, Ma L, Yan X … +1 more , Li X
Accid Anal Prev · 2026 Jul · PMID 41990541 · Publisher ↗
Road crashes remain a primary focus in the field of traffic safety research due to their potentially severe societal and individual consequences. While many studies have focused on environmental and collision-related fac...
Road crashes remain a primary focus in the field of traffic safety research due to their potentially severe societal and individual consequences. While many studies have focused on environmental and collision-related factors that influence traffic accident fatalities, research on the impact of road safety policing, particularly its spatiotemporal heterogeneity, remains to be explored. To bridge this critical gap, we analyzed road traffic accident data from seven Australian states, along with data on the strength of each state's road safety policies. We propose a hybrid modeling strategy to examine the complex relationship between policy interventions and road traffic accident severity. This approach integrates interpretable machine learning techniques with Bayesian random parameter method, enabling the identification of high-dimensional predictors and the exploration of nonlinear relationships. Our findings indicate that the effect of road safety policy implementation on accident severity is inherently nonlinear. Specifically, our analysis suggests that higher numbers of licensed drivers and the presence of seatbelt cameras are associated with a reduction in fatal collisions, while incidents resulting solely in property damage may increase. Rigorous testing confirms the reliability of our models, offering a robust foundation for future research on traffic policy interventions.
An exploration of active safety technologies for commercial vehicles: Research on on-road driving risk identification based on alarm and other multi-source data.
Xu B, Lu J, Dong C
Accid Anal Prev · 2026 Jul · PMID 41966718 · Publisher ↗
In accident analysis and prevention, passive vehicle safety technologies ensure the lower limit of driving safety, while active safety technologies determine the upper limit. This study aims to provide suggestions for th...
In accident analysis and prevention, passive vehicle safety technologies ensure the lower limit of driving safety, while active safety technologies determine the upper limit. This study aims to provide suggestions for the active safety management of commercial vehicles by identifying high-risk on-road scenarios. Firstly, taking regular-route passenger buses as the research object, based on multi-source data fusion technology, this study integrates driving alarm data, vehicle trajectory data within 5 min before the alarm, driver video data within 10 s before the alarm, and driving record video data to extract key features and construct a driving risk identification variable set. Secondly, Pearson correlation coefficient and variance inflation factor (VIF) are used sequentially to conduct collinearity tests and eliminate redundant variables. Considering that the variables include both continuous and discrete heterogeneous data, the K-prototype hybrid clustering method is adopted, and the optimal number of clusters (K = 4) is finally determined. Thirdly, an integrated method of 'multi-source heterogeneous data fusion-hybrid variable clustering-Ordered Logit modeling-SHAP interpretability analysis' is constructed. In an effort to explore active safety technologies, this study attempts to map the identified driving patterns to ordinal risk levels based on key vehicle kinematic parameters. Subsequently, the Ordered Logit model is applied to quantitatively analyze the marginal effects of significant variables. Finally, combined with the variable distribution characteristics of the clustering results and SHAP interpretability analysis, the core features and key incentives of the four risk levels are systematically characterized, and targeted active safety management suggestions are generated with the assistance of Large Language Models (LLMs). This study intends to provide certain insights for the research on vehicle active safety and offer references and suggestions for the dynamic monitoring and management of commercial vehicles.
Differences in pedestrian injury severities between contrasting and non-contrasting attire in nighttime crashes: a partially constrained temporal modeling and out-of-sample prediction approach.
Agheli A, Aghabayk K, Shiwakoti N … +1 more , Tay R
Accid Anal Prev · 2026 Jul · PMID 41966717 · Publisher ↗
Pedestrian crashes during nighttime hours pose a significant safety challenge, with reduced visibility being a key contributing factor in injury severity. Despite its importance, the role of pedestrian attire visibility,...
Pedestrian crashes during nighttime hours pose a significant safety challenge, with reduced visibility being a key contributing factor in injury severity. Despite its importance, the role of pedestrian attire visibility, specifically contrasting versus non-contrasting clothing, has received limited empirical attention in real-world crash settings. To address this gap, we apply a partially constrained temporal random parameters logit model with heterogeneity in means, the first application of this framework to assess pedestrian attire visibility and nighttime injury outcomes. Using six years of police-reported data from Chicago (2018-2023), stratified into pre-, during-, and post-COVID periods, likelihood ratio tests confirmed significant temporal instability, supporting the proposed framework. Out-of-sample prediction experiments were conducted to estimate how outcomes might differ if pedestrians in non-contrasting attire had instead worn contrasting clothing. Findings show that pedestrians in non-contrasting attire face higher probabilities of severe injury, especially after COVID, with older pedestrians disproportionately affected. Contrasting attire mitigates much of this risk in dark, unlit conditions, underscoring the importance of conspicuity. Additional results indicate elevated risks associated with pedestrians disregarding traffic control devices, in-roadway walking, and driver failure to reduce speed, whereas protective effects are observed in marked crosswalks, under lower speed limits, and in crashes involving female drivers. Out-of-sample predictions further suggest that reassigning contrasting attire would substantially lower the likelihood of severe injuries. While attire visibility matters, the results underscore the need for systemic interventions, improved roadway lighting, traffic calming, and strengthened driver responsibility, with visibility promotion positioned as a complementary measure rather than a substitute for infrastructure and enforcement.
Causal inference-based analysis of critical design features for right-turn crashes at signalized intersections.
Huang Z, Wang X, Qin X … +5 more , Wu M, Guo Q, Luo G, Wang B, Jiang Y
Accid Anal Prev · 2026 Jul · PMID 41962450 · Publisher ↗
In China, there are usually no exclusive right-turn phases or right-turn-on-red at signalized intersections. Right-turn crashes account for over 30% of intersection-related crashes, highlighting the critical role of inte...
In China, there are usually no exclusive right-turn phases or right-turn-on-red at signalized intersections. Right-turn crashes account for over 30% of intersection-related crashes, highlighting the critical role of intersection design in traffic safety. However, the relationship between specific design features and crash risk remains unclear. This study identifies key design variables and quantifies their causal effects and heterogeneity on right-turn crashes at signalized intersections. A total of 271 signalized intersections in Suzhou, China, were analyzed from 2022 to 2024. The SHapley Additive exPlanations method was applied to rank variable importance, followed by Generalized Random Forest modeling to estimate both Average and Heterogeneous Intervention Effects (HIEs) while controlling for confounding bias. The minimum right-turn radius and intersection skewness were identified as the most influential factors for total and fatal crashes, respectively. A turning radius of approximately 15 m was associated with the lowest total crash risk, and right-angle intersections were linked to reduced fatal crash rates. A bicycle lane barrier width on the minor road between 1.01 and 3 m also contributed to crash reduction. HIEs showed that smaller turning radii improved safety at high-volume, multi-lane intersections, and skewed intersections with low traffic volumes required special attention. Facilities such as bicycle lanes, physical barriers, and channelization islands enhanced safety performance at skewed intersections. As turning radii increase, expanding the bicycle lane barrier width may be beneficial, although wider barriers should be applied cautiously under low traffic conditions. This study provides a causal inference framework for evaluating right-turn design and offers evidence-based guidance for improving intersection safety in urban environments.
A GAN-based framework for predicting pedestrian road safety based on land use and Points of Interest (POIs).
Qiao W, Zheng H
Accid Anal Prev · 2026 Jul · PMID 41955836 · Publisher ↗
Pedestrian road safety represents a critical challenge in rapid urbanization processes, with land use and point of interest (POIs) configurations playing a central role in shaping crash risks. Existing studies mainly rel...
Pedestrian road safety represents a critical challenge in rapid urbanization processes, with land use and point of interest (POIs) configurations playing a central role in shaping crash risks. Existing studies mainly rely on traditional quantitative models, which often struggle to capture nonlinear spatial relationships and lack real-time, visual feedback for planning. This study develops a Generative Adversarial Network (GAN)-based framework to predict pedestrian crashes, integrating land use, POIs, and commuting distance to capture both direct and indirect associations. Using 504 spatial analysis units in Auckland, New Zealand, and 2870 pedestrian crash records (2016-2024), the model performs image-to-image translation to simulate crash risks under different urban configurations. Findings show that pedestrian crash risks vary nonlinearly across space and are linked to the combined effects of land use, POIs, and commuting distance. In the CBD, reducing commuting attraction areas emerges as a key factor associated with improved safety outcomes, with public open space expansion reducing pedestrian crashes by up to 14.7%. Sub-centre areas demonstrate sharp risk amplification (up to 80.4%) when educational, commercial, and transportation POIs simultaneously reach high densities. Residential areas exhibit significant threshold effects, where high-density combinations are associated with extreme risk scenarios (up to 35.5%). Moreover, direct associations significantly outweigh indirect pathways. This study offers a scenario-based visual tool for exploring the safety implications of objection and POI planning, providing empirical evidence for data-driven urban design.
Comparative safety evaluation of ADAS-Equipped electric and gasoline vehicles using real-world crash data.
Ding S, Abdel-Aty M, Chun U
Accid Anal Prev · 2026 Jul · PMID 41955835 · Publisher ↗
The rapid advancement of Battery Electric Vehicle (BEV) in automotive technology and market adoption present safety challenges due to their distinct design and operational characteristics compared to Gasoline Vehicles (G...
The rapid advancement of Battery Electric Vehicle (BEV) in automotive technology and market adoption present safety challenges due to their distinct design and operational characteristics compared to Gasoline Vehicles (GVs). A proactive approach to evaluating the safety performance of Advanced Driver Assistance Systems (ADAS) among BEVs and GVs is necessary to address these emerging risks. A total of 8,118 Florida crash reports with ADAS engaged were analyzed in this study, consisting of 6,052 from GV and 2,066 from BEV. Comparability between BEVs and GVs regarding ADAS safety evaluation across diverse traffic conditions was established using Propensity Score Matching (PSM), examining injury severity, environmental conditions, driver maneuvers, and vehicle types. And then a partially constrained Random Parameter Logit (RPL) model was employed to identify both the unique and shared factors influencing injury severity of BEVs and GVs. PSM showed that ADAS-engaged BEV crashes had lower injury severity (75.13% no-injury vs. 69.82% for GVs) but higher involvement in Vulnerable Road User (VRU) crashes (3.2% vs. 1.8%). The RPL model revealed ADAS-engaged BEV crash risk in work zones (+0.0057 severe injury probability), GV crash risk at processing straight and backing (+0.0042 minor injury probability), and shared risk at speeds above 61 mph (+0.0084 severe injury probability). This study offers an exploratory analysis of ADAS safety performance between BEVs and GVs using real-world crash data. These findings provide valuable insights for manufacturers and other stakeholders to inform decisions regarding the deployment and application of these technologies.
Joint analysis of two nationwide crash databases to model transit bus stop-related pedestrian crash severity in the United States.
Rewalt A, Brakewood C, Khattak AJ
Accid Anal Prev · 2026 Jul · PMID 41950765 · Publisher ↗
The Fatality Analysis Reporting System (FARS) and Crash Report Sampling System (CRSS) report a unique "transit bus stop-related" pedestrian crash type. This study merged fatal and non-fatal transit bus stop-related pedes...
The Fatality Analysis Reporting System (FARS) and Crash Report Sampling System (CRSS) report a unique "transit bus stop-related" pedestrian crash type. This study merged fatal and non-fatal transit bus stop-related pedestrian crashes from FARS and CRSS (2016-2023) to examine injury severity and identify risk factors for pedestrians at bus stops. To address imbalance in severity (KA = 183, BCO = 86), synthetic minority oversampling for nominal and continuous features (SMOTE-NC) was applied. Then, a severity analysis using elastic net penalized logistic regression was conducted on the original and SMOTE-NC sample with 17 predictors, as regularization handles high-dimensional data and multicollinearity better than traditional parametric approaches. The SMOTE-NC model showed severe transit bus stop-related pedestrian crashes were associated with poor lighting, midblock locations, and higher-speed, wider roadways. Results showed amplified effects for predictors in the SMOTE-NC models, including several that were zero in the original model but were selected after oversampling, suggesting oversampling potentially mitigated bias due to class imbalance. The results provided a foundation for identifying tangible solutions to address dangerous bus stops. Improving street lighting or installing bus stop lighting may be effective, as dark, unlighted conditions were associated with higher odds ratios for KA injuries (OR = 3.18) than dark, lighted conditions (OR = 1.69), relative to daylight. Evaluating pedestrian behavior near midblock bus stops, like unmarked crossings, can further mitigate risks exacerbated by darkness and high-speed multilane roads. These findings support data-driven policy and design decisions by transit agencies and transportation planners to address higher-risk bus stops.
Impact of tunnel entrance/exit sidewalls and Medians on driver behavior: Operational safety risk diagnosis based on avoidance dynamics and trajectory evolution.
Chen Y, Dai Z, Du Z … +1 more , Wang S
Accid Anal Prev · 2026 Jul · PMID 41946186 · Publisher ↗
Tunnel entrance/exit represent critical transition zones within underground transport infrastructure with elevated accident risks. This study investigates driver avoidance behavior toward sidewalls and median strips unde...
Tunnel entrance/exit represent critical transition zones within underground transport infrastructure with elevated accident risks. This study investigates driver avoidance behavior toward sidewalls and median strips under the alternating light-dark conditions and spatial constraints of these areas, aiming to mitigate human-space interaction risks during underground-to-surface transitions. Using UAV to collect vehicle driving data for Rail-cum-Road Yangtze River Cross Tunnel in Wuhan, China. Employing an "avoidance degree grading-risk pattern clustering" methodology, vehicle trajectory dynamics were quantified. Revealed drivers' spatial perception and avoidance behavior mechanisms, establishing a safety performance diagnosis and evaluation model. The main conclusions are as follows: 1) Drivers' avoidance behavior outside the tunnel is dominated by psychological expectations and post-adaptation effects. At the entrance, effective avoidance occurs in a "pre-initiated avoidance behavior" pattern. However, at the exit, difficulties in spatial re-evaluation arise due to "adaptive aftereffect" and the absence of external guidance facilities, leading to a high incidence of high-risk "complex conflict avoidance" (65.71% and 52.22%), highlighting the safety challenges posed by environmental transitions. 2) A seven-category risk spectrum of avoidance behaviors was established. Sidewalls were identified as the primary constraint for trajectory deviation. In the exit where sidewall constraints are released, the influence of the median strip is amplified (left deviation at entrance, right deviation at exit). The synergistic design of sidewalls with median is crucial for trajectory envelope and clearance safety. 3) Enhancing continuous visual guidance to align with drivers' psychological expectations and adaptation processes is crucial for improving operational safety in tunnel transition zones. This approach represents a fundamental shift in design philosophy, from passive spatial confinement to active behavioral guidance.
Exploring pedestrian gap acceptance behaviour using immersive CAVE experiments: A multilevel logit regression model.
Zhu M, Wang Z, Sze NN … +1 more , Gu R
Accid Anal Prev · 2026 Jul · PMID 41946185 · Publisher ↗
Unsafe crossing behaviour is a key contributory factor to pedestrian injuries. Understanding the influences of potential factors on pedestrian crossing behaviour is essential. Previous studies have examined the relations...
Unsafe crossing behaviour is a key contributory factor to pedestrian injuries. Understanding the influences of potential factors on pedestrian crossing behaviour is essential. Previous studies have examined the relationship between pedestrian behaviour, road environments, and traffic characteristics. However, the influences of psychological factors, such as safety perception, on pedestrian decision-making are rarely considered. This study investigates the influences of environmental factors, vehicle attributes, personal demographics, and safety perceptions on the gap acceptance behaviour of pedestrians at mid-block crossings using a hybrid experiment and attitudinal survey approach. For instance, a Cave Automatic Virtual Environment (CAVE) method is employed to enhance the immersive experiences of 3-dimensional road environments and dynamic traffic characteristics for pedestrians. A multilevel logit regression method is then employed, controlling for the interdependency between multilevel factors: (1) Participant level: demographics, safety attitude; (2) Trial level: road environment, traffic control, vehicle mix; and (3) Observation level: vehicle class, gap size, and waiting time, in the association measure. Results indicate that the likelihood of gap acceptance increases with pedestrian age, risk-taking attitude, speed limit, gap size, and waiting time. In contrast, the likelihood of gap acceptance decreases with the increased perceived control, presence of on-street parking, and presence of heavy vehicles. These findings shed light on effective remedial measures, such as targeted road safety education and local area traffic management, that can mitigate pedestrian crash risk at high-risk locations with frequent pedestrian-vehicle interactions. Therefore, overall pedestrian safety can be improved, and walkability can be enhanced in the long run.
The influence of cannabis on real time higher-order driving skills: a scoping review.
van der Kint ST, Davidse RJ, van Strijp M
Accid Anal Prev · 2026 Jul · PMID 41936255 · Publisher ↗
Cannabis use is frequently detected in motor vehicle crashes. While research has found an increased crash risk of driving under the influence of cannabis, research on the effects of cannabis, or THC, on driving behaviour...
Cannabis use is frequently detected in motor vehicle crashes. While research has found an increased crash risk of driving under the influence of cannabis, research on the effects of cannabis, or THC, on driving behaviour seems to be somewhat limited to vehicle control. Our objectives were to map available research into the effects of THC on higher-order driving skills during actual driving. A scoping review was conducted by systematically searching Scopus, PubMed and TRID between 2003 and 2024. After applying inclusion criteria to 2326 studies, 40 studies underwent full-text review, after which three studies met the inclusion criteria. These studies examined higher-order driving skills using driving simulators or real-world driving and measured executive function, attention, time perception, decision making, visual search and vigilance. The results of this study revealed a scarcity of research investigating the effects of THC on higher-order driving skills. While the results of the included studies show no or only small effects of THC on higher-order driving skills, there is so little research available that no conclusions can be made. There seems to be limited evidence addressing the impact of THC on critical higher-order cognitive abilities which are essential for safe driving. This highlights a significant gap in the body of research underlining the need for future research to explore these skills in a driving environment. Research using simulators combined with eye-tracking might provide these important insights.
Adapting hazard perception testing to the Czech driving context: a comparison of hazard perception paradigms in differentiating driver experience and collision history.
Ventsislavova P, Sucha M, Harrison L … +2 more , Novotny J, Crundall D
Accid Anal Prev · 2026 Jul · PMID 41936254 · Publisher ↗
Hazard perception (HP) testing is widely regarded as an effective approach for identifying unsafe drivers and mitigating collision risk. The European Transport Safety Council has therefore recommended its adoption within...
Hazard perception (HP) testing is widely regarded as an effective approach for identifying unsafe drivers and mitigating collision risk. The European Transport Safety Council has therefore recommended its adoption within national licensing systems. Evidence indicates that HP assessments must be adapted to local environments to yield comparable safety benefits. Moreover, traditional HP tests depend largely on reaction times, which may introduce post-perceptual biases because responses typically reflect the moment a hazard is judged as threatening. To address this, the present study developed two versions of the HP test tailored to the Czech driving context: a conventional reaction-time-based test and a prediction-based test in which participants were asked to anticipate how a hazardous situation would unfold. Hazardous traffic scenarios were filmed across the Czech Republic and edited into hazard perception and prediction tests. A total of 225 participants were recruited and randomly allocated to either the hazard perception or hazard prediction test. Within each test, participants were grouped according to pre-existing driving experience (novice vs. experienced) and collision history (collision-free vs. collision-involved). The results indicated that experienced drivers outperformed novices on both test versions, showing faster and more accurate hazard detection and superior predictive performance. These findings support the feasibility of incorporating a hybrid HP assessment into the Czech national licensing system. However, significant collision group differences emerged only in the perception task, with collision-involved drivers demonstrating slower hazard response times. This suggests that the key differences may arise from post-perceptual decision-making processes rather than perceptual limitations.
Pedestrian and cyclist perceptions of interactions with fully automated buses.
Downey L, Fonzone A, Fountas G
Accid Anal Prev · 2026 Jul · PMID 41932068 · Publisher ↗
For Automated Buses (ABs) to operate on the public road network they must interact with other road users in ways that are both clear and predictable. This requirement is especially vital in interactions with Vulnerable R...
For Automated Buses (ABs) to operate on the public road network they must interact with other road users in ways that are both clear and predictable. This requirement is especially vital in interactions with Vulnerable Road Users (VRUs) such as pedestrians and cyclists. This study investigates the determinants of VRU perceived safety and trust when sharing the roads with ABs, using data obtained from an online survey distributed to UK residents (n = 2,013). Findings indicated that only 34.6% of respondents would feel comfortable sharing the road with ABs while walking and even fewer (23.0%) would feel comfortable cycling alongside them. The survey included a best-worst scaling experiment to identify which AB driving behaviors respondents considered the most important. An exploratory factor analysis was conducted on items from the BFI-2-S personality inventory to extract factor scores representing the Big Five personality dimensions. A bivariate probit model with heterogeneity in the means of the random parameters was estimated for walking and cycling comfort. The best-fitting model included sociodemographic characteristics, travel behavior patterns, personality, attitudes and experiences with automated technologies and the perceived importance of AB driving behaviors such as yielding to pedestrians at crossing facilities. Extraversion was positively associated with comfort for both pedestrians and cyclists, while negative emotionality and conscientiousness were positively associated with cycling comfort. Taking part in an automated public transport trial was a positive predictor of comfort for both groups, providing support for expanding trials that allow VRUs to interact with ABs in real-world settings.
Training for the unexpected: Enhancing driver preparedness through hazard awareness. A 15-year cohort study.
Kuipers J, Mulder M, Voskes M
Accid Anal Prev · 2026 Jul · PMID 41921244 · Publisher ↗
This cohort study investigates the long-term effects of simulator-based hazard awareness training (HAT) on learner and novice drivers in the Netherlands, using a dataset of 2,372 participants over a 15-year period. Most...
This cohort study investigates the long-term effects of simulator-based hazard awareness training (HAT) on learner and novice drivers in the Netherlands, using a dataset of 2,372 participants over a 15-year period. Most prior studies on HAT have measured only immediate post-training outcomes; no longitudinal cohort study with a control group has previously examined both supervised and unsupervised driving outcomes over a multi-year horizon. Although the HAT and control groups showed small but statistically significant differences in gender composition, education level, and fear of driving at the start of training, the effect sizes were negligible (d ≤ 0.09), and these characteristics are addressed as covariates in the analyses. HAT improved performance during simulator training and supervised driving: HAT students' viewing skills were better during the intersection test, required fewer on-road training hours, passed the driving exam in fewer attempts, and achieved a higher first-attempt pass rate than the control group. These benefits did not persist into unsupervised post-licensing driving. Violations, errors, and accident involvement were comparable between HAT and control group drivers in the first and last year after licensing. Personal characteristics - including gender, licensing age, self-assessed driving competence, and subjective driving difficulty - were stronger and more lasting predictors of post-licensing behaviour than training type. These findings suggest that hazard awareness is a trainable skill, but that training effects on risk-taking behaviour are moderated by individual characteristics that emerge most clearly once drivers operate independently, aligning with findings of a previous study on the same dataset. Teaching higher safety margins during supervised driving may offer a more durable route to reducing accident risk for novice drivers than higher-order skill training alone.
Learning from geometry-aware near misses to real-time COR: A corridor-wide grouped random parameters GEV framework.
Anis M, Zhou Y, Lord D
Accid Anal Prev · 2026 Jul · PMID 41921243 · Publisher ↗
Real-time corridor-wide crash-occurrence risk (COR) prediction is challenging, since existing near-miss EVT models oversimplify collision geometry, neglect vehicle-infrastructure (V-I) interactions, and fail to adequatel...
Real-time corridor-wide crash-occurrence risk (COR) prediction is challenging, since existing near-miss EVT models oversimplify collision geometry, neglect vehicle-infrastructure (V-I) interactions, and fail to adequately account for spatial heterogeneity in traffic and roadway conditions. To do so, this study develops a geometry-aware 2D-TTC near-miss extraction and integrates it with a hierarchical Bayesian structure grouped random parameters (HBSGRP-UGEV) to estimate short-term COR in urban corridors. Building on prior grouped EVT formulations while explicitly accommodating both V-V and V-I near-miss processes within a unified corridor-wide modeling framework. High-resolution trajectories from the Argoverse-2 dataset were analyzed across 28 sites on Miami's Biscayne Boulevard to extract extreme near-miss events. The model incorporates vehicle dynamics and roadway features as covariates, with partial pooling across segments and intersections to capture corridor-wide heterogeneity. Results show that the HBSGRP-UGEV framework outperforms fixed-parameter HBSFP-UGEV models, reducing DIC by up to 7.5% (V-V) and 3.1% (V-I). Predictive validation using ROC-AUC confirms strong accuracy (0.89 for V-V segments, 0.82 for intersections, 0.79 for V-I segments, and 0.75 for intersections). Grouped random-parameters (HBSGRP) framework indicate that relative (speed, distance, and deceleration) dominate V-V near-miss risk on segments, whereas V-I segment risk is primarily associated with relative distance; at intersections, V-V risk is driven by relative (speed and distance), while V-I dynamics exhibit no statistically significant effects. These findings demonstrate the value of a geometry-aware, spatially adaptive framework for proactive corridor safety management, supporting both real-time interventions and long-term Vision Zero goals.
Heterogeneous treatment effects of emotional arousal on aggressive driving among ride-hailing drivers under distraction and passenger presence.
Gao L, Ma Y, Chen S … +2 more , Cheng C, Xing Y
Accid Anal Prev · 2026 Jul · PMID 41916253 · Publisher ↗
This study investigates the heterogeneous treatment effects of emotional arousal on aggressive driving behaviors among ride-hailing drivers using naturalistic driving data. Three types of aggressive driving-abnormal acce...
This study investigates the heterogeneous treatment effects of emotional arousal on aggressive driving behaviors among ride-hailing drivers using naturalistic driving data. Three types of aggressive driving-abnormal acceleration, moderate speeding, and severe speeding-were identified through vehicle kinematics and video recordings. Emotional arousal was classified as positive or negative based on emotional valence. A double machine learning framework (DML) with generalized random forests (GRF) was employed to estimate the average treatment effect (ATE) and conditional average treatment effect (CATE). Results show that both high positive and high negative arousal increase the likelihood of aggressive driving, with negative arousal exerting a stronger effect across all three aggressive behaviors. Abnormal acceleration emerged as the primary means for emotional expression. Drivers who were distracted, without passenger were more susceptible to emotional influence-especially under positive arousal. This study underscores the heterogeneous treatment effects of emotional arousal on aggressive driving. The findings emphasizes the need for more refined and differentiated emotional management and intervention strategies by ride-hailing platforms.
Driving behavior recognition and scoring: A Bayesian approach to driving simulator data analytics.
Pourghasemi M, Rezaei E, Ansariyar A … +3 more , Ardeshiri A, Jeihani M, Tootkaboni M
Accid Anal Prev · 2026 Jul · PMID 41916252 · Publisher ↗
Safety is a key factor in operating commercial motor vehicles (CMVs). Advanced driving assistance systems are designed to assist CMV drivers enhance driving safety and reduce the risk of accidents. This study proposes a...
Safety is a key factor in operating commercial motor vehicles (CMVs). Advanced driving assistance systems are designed to assist CMV drivers enhance driving safety and reduce the risk of accidents. This study proposes a framework for driving style recognition and introduces statistical features that enable modeling the behavioral risk of drivers in real-time. These features include main aspects such as speed, throttle, and steering wheel as well as the those surrounding traffic conditions. A full-scale, high-fidelity driving simulator is used to generate the driving data. The driving records are clustered into different groups and a Bayesian classifier is trained the training portion of data to compute the probability of belonging to aggressive or safe classes. The framework is then extended, under normality assumption and using the distance from the mean, to one that provides driving score based on proximity to the boundary between safe and risky driving. Finally, the accuracy and effectiveness of the framework are evaluated by calculating the confusion matrices and AUCs. The experimental results demonstrate that the proposed framework for driving style recognition achieves accuracy rates of 88% to 91% in the test portion of the driving simulator data set.
About
- Frequency
- Sun
- Papers found
- 200
- RSS feed
- Subscribe