Accident; analysis and prevention最新文献

{"title":"How to assess situation awareness while driving with automation? Impact of visual attention during AD mode on measures of situation awareness","authors":"Barbara Metz ,&nbsp;Johanna Wörle ,&nbsp;Myriam Metzulat ,&nbsp;Alexandra Neukum","doi":"10.1016/j.aap.2025.108142","DOIUrl":"10.1016/j.aap.2025.108142","url":null,"abstract":"<div><div>A variety of objective and subjective methods is used to assess situation awareness (SA) in automated driving (AD). In fields like aviation, established methods like SAGAT are used as state of the art for assessing SA, whereas in AD, there is less consensus on the most valid measures. In a driving simulator study with N = 41 participants, four different levels of SA were experimentally created by manipulating visual attention during AD. Different methods for assessing SA were logged during AD-mode and during takeover situations including subjective ratings, gaze behaviour, performance and probe measures. The impact of the manipulation of visual attention during AD mode on the different measures of SA as well as their relation to each other is analysed, reported and discussed. Results show pronounced differences between levels of attention during AD-mode in subjective SA, gaze behaviour and performance measures while preventing visual processing of the driving scenery completely while during AD mode caused surprisingly little impact. A relation between measures of SA and performance can be shown for specifically designed takeover scenarios with increased demands on SA. On the one hand, this implies that there can be a critical impact of SA on performance in takeover situations but it also highlights the robustness and efficiency of drivers’ visual processing that enables safe takeover responses even in situations with only little visual processing before a takeover. The implications of the results for assessing SA as well as the processes behind SA in AD are discussed.</div></div>","PeriodicalId":6926,"journal":{"name":"Accident; analysis and prevention","volume":"220 ","pages":"Article 108142"},"PeriodicalIF":5.7,"publicationDate":"2025-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144365354","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluation of the causal impact of recreational marijuana legalisation on traffic safety in the US","authors":"Anupriya ,&nbsp;Emma McCoy ,&nbsp;Daniel J. Graham","doi":"10.1016/j.aap.2025.108106","DOIUrl":"10.1016/j.aap.2025.108106","url":null,"abstract":"<div><div>Since the legalisation of recreational marijuana in certain US states, traffic fatalities involving drivers testing positive for marijuana have markedly increased, thereby prompting the need to understand how this policy change affects road safety. While marijuana is well-known to impair driving, determining if its recreational use directly causes more traffic fatalities remains contentious due to challenges in roadside impairment testing. Additional challenges arise because (i) Simulations may not accurately replicate driver impairment and road conditions, (ii) Estimation based on observational data must adjust for (unobserved) confounding factors, requiring an innovative model to generate causal inference, and (iii) The dynamic, evolving nature of the process requires capturing temporal relationships. This paper contributes by employing a rigorous study design based on an augmented synthetic control method to assess the causal impact of recreational marijuana legalisation on traffic fatalities. It identifies a consistent but lagged pattern of increased fatality rates in several states post-legalisation, with the effect primarily linked to the drug’s retail availability. These findings disprove any prevailing conjectures that dismiss the link between recreational marijuana use and fatal traffic crashes, highlighting the need for informed policy responses.</div></div>","PeriodicalId":6926,"journal":{"name":"Accident; analysis and prevention","volume":"220 ","pages":"Article 108106"},"PeriodicalIF":5.7,"publicationDate":"2025-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144331364","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effects of early notifications on driver responses to lateral collision warnings","authors":"Wenjing Zhao ,&nbsp;Siyuan Gong ,&nbsp;Dezong Zhao ,&nbsp;Fenglin Liu ,&nbsp;N.N. Sze ,&nbsp;Song Wang","doi":"10.1016/j.aap.2025.108144","DOIUrl":"10.1016/j.aap.2025.108144","url":null,"abstract":"<div><div>Despite the demonstrated effectiveness of advanced driver assistance systems, lateral collisions still occur because of incorrect and deferred responses of drivers to potential hazards attributed to approaching vehicles from the sides. In such a situation, this study introduces early notifications prior to lateral collision warnings, displayed in either a visual-only format or a visual-auditory format. These notifications inform drivers about the presence of approaching vehicles from behind in adjacent lanes, aiming to improve their performance after receiving lateral collision warnings. The improvement is measured by analyzing the differences in drivers’ reaction times and exhibition of appropriate reactions when early notifications are present and absent. Additionally, the effects of confounding factors including display formats (visual-only versus visual-auditory), drivers’ socio-demographics, and driving experience on drivers’ reaction times and exhibition of appropriate reactions are considered in the binary logit and Tobit models. Furthermore, the effects of individual heterogeneity are accounted using the random parameters approach. Results indicate that there is significant heterogeneity in drivers’ reaction time when early notifications is present, prior to lateral collision warnings. Even that drivers’ reaction time may increase in some circumstances, their likelihood of exhibiting appropriate reactions increases. However, it is interesting to find that drivers’ reaction time increases and their likelihood of exhibiting appropriate reactions decreases when early notifications are presented in a visual-auditory format, compared to a visual-only format.</div></div>","PeriodicalId":6926,"journal":{"name":"Accident; analysis and prevention","volume":"220 ","pages":"Article 108144"},"PeriodicalIF":5.7,"publicationDate":"2025-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144313394","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Investigating the impact of in-vehicle warning information complexity on drivers: The role of working memory capacity and cognitive load","authors":"Kunchen Li ,&nbsp;Wei Yuan ,&nbsp;George Yannis ,&nbsp;Fuwei Wu ,&nbsp;Chang Wang","doi":"10.1016/j.aap.2025.108138","DOIUrl":"10.1016/j.aap.2025.108138","url":null,"abstract":"<div><div>In-vehicle warning systems significantly reduce collisions. However, poorly designed warnings, such as those with excessive or insufficient information, intensify the resource consumption of the drivers. The working memory is a crucial component of the cognitive function, which is closely related to the processing of short-term information. Therefore, this paper investigates the impact of the complexity of the warning messages on the behavior and physiological states of the driver, taking into account individual differences in working memory capacity and cognitive load levels. A total of 37 participants are recruited to conduct a 4 (warning information complexity) × 2 (working memory capacity) × 2 (cognitive load) mixed design driving simulation experiment, with working memory capacity treated as a between-subjects factor. An eye-tracker and a physiometer are employed to record participant’s visual motion and heart rate. A correlation analysis is then conducted to identify key dependent variables, and a Generalized Linear Mixed-effects Model (GLMM), which considers random effects, is used to analyze the impact of each experimental factor on the drivers. The obtained results demonstrate that visually rich warnings lead to increased braking reaction times, especially between drivers having low working memory capacity and under high cognitive load. Although detailed warnings are easier to understand, they tend to reduce Root Mean Square of Successive Differences (RMSSD) of the driver under higher cognitive loads, indicating increased tension and annoyance. In addition, the combination of visually simple and auditorily rich warnings has significant advantages, allowing almost all types of participants to perceive risks more quickly, which significantly reduces the collision risks. These findings offer theoretical insights to assist manufacturers in designing human-centered, personalized, and adaptive in-vehicle warning systems.</div></div>","PeriodicalId":6926,"journal":{"name":"Accident; analysis and prevention","volume":"220 ","pages":"Article 108138"},"PeriodicalIF":5.7,"publicationDate":"2025-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144313484","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting occupant response curves in vehicle crashes via Attention-enhanced multimodal temporal Network","authors":"Wenjie Wang ,&nbsp;Xiaoyi Tai ,&nbsp;Chang Zhou ,&nbsp;Zhao Liu ,&nbsp;Ping Zhu","doi":"10.1016/j.aap.2025.108140","DOIUrl":"10.1016/j.aap.2025.108140","url":null,"abstract":"<div><div>Accurately predicting safety responses, especially occupant crash response curves across multiple body regions, plays a crucial role in advancing vehicle crash safety by enabling design optimization and reducing the reliance on costly physical testing and simulations. Machine learning methods have demonstrated good performance in this field, but existing approaches often face challenges in integrating multimodal data and handling multi-task temporal predictions. To address these issues, this work proposes a novel Attention-enhanced Multimodal Temporal Network (AMTN) for predicting occupant crash response curves across multiple body regions during crashes. AMTN integrates numerical parameters and vehicle body crash pulses through a feature extraction module, organically fuses multimodal features via cross-attention mechanisms, and decodes shared features using a modified Temporal Convolutional Network (TCN) with local sliding self-attention. A dynamic adaptive loss and multiple output layers are utilized to evaluate the importance of each task, iteratively update the learning priorities, and finally achieve multi-task prediction. Experiments on the engineering-obtained data demonstrate that the crash response curves predicted by AMTN achieves an average ISO (International Organization of Standards) rating of 0.835 across 11 crash response curves, with critical regions exceeding 0.9. In engineering applications, an ISO rating greater than 0.8 indicates that the predicted curve closely matches the reference curve. Therefore, the experimental results demonstrate that the proposed method has effectively learned the characteristics of the training data and is capable of producing accurate predictions. In summary, this work advances multimodal deep learning for crash safety by enabling efficient, accurate, and interpretable multi-task curve predictions. Consequently, it can be applied to data-driven vehicle safety development, offering significant engineering value by enhancing both development efficiency and quality.</div></div>","PeriodicalId":6926,"journal":{"name":"Accident; analysis and prevention","volume":"220 ","pages":"Article 108140"},"PeriodicalIF":5.7,"publicationDate":"2025-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144297021","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Balancing safety and efficiency for autonomous vehicles at urban uncontrolled crosswalk: challenges and countermeasures","authors":"Xu Chen ,&nbsp;Zihao Xi ,&nbsp;Yan Xu ,&nbsp;Zhuozhi Xiong ,&nbsp;Xuelin Ding ,&nbsp;Hao Wang","doi":"10.1016/j.aap.2025.108111","DOIUrl":"10.1016/j.aap.2025.108111","url":null,"abstract":"<div><div>At uncontrolled crosswalks, defensive autonomous vehicles (DAVs) prioritize yielding to pedestrians but may exacerbate congestion. Interactive autonomous vehicles (IAVs) attempt to prioritize pedestrian passage under safe conditions; however, they demand more advanced technology, and their practical application remains uncertain. Platoon control presents a potential solution to mitigate the limitations of both driving styles. This study develops an agent-based pedestrian-vehicle interaction framework and introduces a platoon formation module that accounts for passenger comfort. The study examines the impact of combining driving styles (DAV and IAV) and the platoon control method on the safety and efficiency of pedestrian-AV interactions. Results indicate that compared to DAV, IAV increases interaction events by 32.2%, yet achieves a yielding compliance rate of only 40% while reducing total delays by 60.7%. Platoon control effectively enhances the safety and efficiency of both driving styles, with its benefits becoming more pronounced as pedestrian and vehicle traffic volumes increase. This study provides methodologies and strategies to address the challenges of integrating AVs into urban environments.</div></div>","PeriodicalId":6926,"journal":{"name":"Accident; analysis and prevention","volume":"220 ","pages":"Article 108111"},"PeriodicalIF":5.7,"publicationDate":"2025-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144288889","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact of jaywalking on pedestrian interaction behavior: A multiagent Markov Game-based analysis","authors":"Elena Abu Khuzam,&nbsp;Gabriel Lanzaro,&nbsp;Tarek Sayed","doi":"10.1016/j.aap.2025.108141","DOIUrl":"10.1016/j.aap.2025.108141","url":null,"abstract":"<div><div>Jaywalking behavior represents a major safety concern especially in traffic environments with intense pedestrian activity. Despite the influence of this behavior on crash risk given that drivers have unexpected interactions with pedestrians and must take additional evasive actions, limited pedestrian models have accounted for jaywalking behavior. This research uses Multiagent Adversarial Inverse Reinforcement Learning (MAAIRL) within a Markov game framework to model road user behavior in jaywalking scenarios at signalized intersections, offering a detailed representation of the dynamic and complex decision-making strategies of pedestrians and drivers in these situations. This approach enables obtaining reward functions that can be used to make inferences about their behaviors and optimal policies that represent the best sequences of decisions, which can be used in developing microsimulation models. Results show that jaywalking pedestrians exhibited erratic movements, with higher acceleration rates and unpredictable paths. In contrast, non-jaywalking pedestrians showed more predictable behavior with smaller variations in their paths and greater distances from vehicles while crossing. Additionally, jaywalking scenarios led to smaller time-to-collision (TTC) and post-encroachment time (PET) values, reduced minimum distances, and faster pedestrian movements compared to non-jaywalking scenarios, which shows the increased crash risks associated with jaywalking. Finally, the MAAIRL model was able to adequately learn the behaviors associated with both non-jaywalking and jaywalking pedestrians. This shows the potential of this framework to model complex real-world scenarios. These findings underscore the importance of improving pedestrian simulation models to take into account the distinct behavioral patterns associated with jaywalking, and such advancements can facilitate a more comprehensive examination of the safety impacts in busy pedestrian environments.</div></div>","PeriodicalId":6926,"journal":{"name":"Accident; analysis and prevention","volume":"220 ","pages":"Article 108141"},"PeriodicalIF":5.7,"publicationDate":"2025-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144280084","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Transferring crash modification factors to automated vehicle environments using surrogate endpoints: Theoretical considerations","authors":"Gary A. Davis,&nbsp;Jingru Gao","doi":"10.1016/j.aap.2025.108112","DOIUrl":"10.1016/j.aap.2025.108112","url":null,"abstract":"<div><div>Although the <em>Highway Safety Manual</em> was developed primarily from statistical summaries of conditions prevailing on North American roads, engineers in other nations have expressed interest in applying, or “transferring,” its predictive methods to places other than those providing the source data. More recently, an emerging issue concerns the application of crash modification factors (CMF) estimated for recent conditions to possibly different conditions in the future, which could change significantly if and when automated vehicles increase their market share. This leads to the question of how the past investment in safety research might be leveraged with a limited experience of newer conditions in order to support reasonable decision-making. The main claim of this paper is that when background knowledge regarding a type of road crash can be reliably represented by a directed acyclic graph, the graph’s connectivity structure can be used to identify a set of surrogate endpoints that will support transfer of a CMF estimated in one situation to a different situation. We present two analytic results that explicate this claim and then use simulation to illustrate the potential applicability of these results. We end with suggestions for further research to help make this approach practical.</div></div>","PeriodicalId":6926,"journal":{"name":"Accident; analysis and prevention","volume":"220 ","pages":"Article 108112"},"PeriodicalIF":5.7,"publicationDate":"2025-06-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144271748","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Safety evaluation of protected bike Lane treatments at Intersections: A causal framework","authors":"Bingyou Dai ,&nbsp;Xuesong Wang ,&nbsp;Qiming Guo ,&nbsp;Lu Yang ,&nbsp;Yu Bai","doi":"10.1016/j.aap.2025.108132","DOIUrl":"10.1016/j.aap.2025.108132","url":null,"abstract":"<div><div>Intersections are a critical focus in bicycle safety research, as approximately one-thirds of bicycle-related crashes occur at these locations. Although protected bike lanes (PBL) at intersections, such as Lateral Shift and Bend-out treatments have been implemented, there is limited crash-based research on their safety performance. Furthermore, the prevailing use of before-after study designs for safety evaluation makes this approach susceptible to selection bias. To address this issue, this study proposes a causal inference framework that combines the advanced generalized causal random forest (GRF) and multimodal large language model (LLM). The LLM is used to extract contextual features from street view images, improving control over unobserved confounding bias. The GRF model is used for effectiveness evaluation by addressing selection bias through residual-based orthogonalization of treatment and outcome. The framework was applied to evaluate the safety impacts of Bend-out and Lateral Shift treatments at intersections. The results indicate that the proposed method outperforms both the baseline and comparative models across all metrics. The average treatment effect (ATE) of Lateral Shift treatments is 1.35 for total crashes and 1.21 for bicycle crashes, suggesting that these treatments tend to increase crashes. For Bend-out treatments, the ATE is −1.61 for total crashes and −0.55 for bicycle crashes, corresponding to a 32.2% reduction in total crashes and a 22.4% reduction in bicycle crashes. Analysis of road user behavior reveals that for Lateral Shift treatments, the low rate of drivers yielding to cyclists is a major issue, with only 30.7% of drivers yielding. To effectively implement Lateral Shift treatments, strengthening enforcement measures should be considered. Furthermore, riding in the wrong direction is a potential risk for both Lateral Shift and Bend-out treatments.</div></div>","PeriodicalId":6926,"journal":{"name":"Accident; analysis and prevention","volume":"220 ","pages":"Article 108132"},"PeriodicalIF":5.7,"publicationDate":"2025-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144263797","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Modeling decision-making during unprotected left turns using interpretable deep learning and uncertainty quantification","authors":"Yubin Chen ,&nbsp;Yajie Zou ,&nbsp;Jun Liu ,&nbsp;Yuanchang Xie ,&nbsp;Jinjun Tang","doi":"10.1016/j.aap.2025.108136","DOIUrl":"10.1016/j.aap.2025.108136","url":null,"abstract":"<div><div>Unprotected left turns present challenges to drivers, as they must manage potential conflicts at intersections, which requires a decision-making process different from that in other driving scenarios. While many studies have modeled human decision-making in unprotected left-turn situations at a behavioral level, most overlook the variability of key information that influences driving behavior and rarely explore the intrinsic mechanisms of decision-making. This study analyzes the decision-making process of drivers in unprotected left-turn scenarios from the perspective of decision uncertainty and explores the relationship between uncertainty and safety. First, a conflict area calculation method is introduced to identify unprotected left-turn interaction events. Next, a transformer model combined with Shapley Additive Explanations is used to identify the key variables driving left-turn decision-making. Finally, Jensen-Shannon divergence are employed to quantify decision-making uncertainty. We explore two types of unprotected left-turn scenarios: left-turn yielding and left-turn proceeding. The experimental results reveal that: (1) left-turning vehicles prioritize static variables, such as waiting time and vehicle type as key variables, while oncoming vehicles focus more on dynamic variables like time to the stop line and speed difference; (2) increased time pressure leads drivers to emphasize on lateral speed and yaw angles during critical decision phases; and (3) higher uncertainty levels are often accompanied by longer negotiation processes and shorter post-encroachment times, which can increase the likelihood of unsafe maneuvers, such as emergency braking. These insights are instrumental in informing decision-making frameworks for autonomous vehicles navigating unprotected left-turn scenarios.</div></div>","PeriodicalId":6926,"journal":{"name":"Accident; analysis and prevention","volume":"220 ","pages":"Article 108136"},"PeriodicalIF":5.7,"publicationDate":"2025-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144271747","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}