Crash Severity Research Articles

Optimizing Equivalent Property-Damage-Only (EPDO) Prediction Models with Genetic Algorithms: A Case Study on Roundabout Geometric Characteristics

Roundabouts generally offer better traffic safety than other intersections, yet severe crashes still occur. They serve as a viable option to enhance intersection safety and reduce crash severity. Improving crash prediction models enhances the precision of prioritization and safety evaluation, ultimately lowering crash-related costs. This study examines the impact of geometric factors on crash frequency and severity in roundabouts. The equivalent property-damage-only (EPDO) index, which considers both severity and frequency, was included as an independent parameter. Increasing traffic volume significantly affects crash numbers, often overshadowing other contributing factors. This study investigates the effects of central island radius (R), average weaving section width (AWWS), and average entry width (AEW) on crashes. To achieve this, data from four roundabouts were analyzed using Gene Expression Programming (GEP) to develop a predictive model. The model achieved a 99% correlation coefficient, effectively capturing data dispersion. The results showed that R accounted for over 75% of the variance, making it the most influential geometric parameter. The proposed procedure can significantly assist traffic safety engineers in enhancing roundabout safety predictions, particularly in small-scale models where traditional methods may be impractical.

Exploring determinants of motorcyclist non-violation crash injury severities on suburban roads of China: a random parameter logit model with heterogeneity in means and variances

ABSTRACT Motorcyclist non-violation (MN-V) crashes were confirmed to be more closely associated with roadway and environmental factors. However, limited studies have explored how these determinants affect injury severity in MN-V crashes. Using motorcycle crash data from Xi’an, China, this study developed multiple random parameters logit models, incorporating heterogeneity in means and variances for investigating the impact of risk determinants on the injury severities sustained by motorcyclists involved in MN-V crashes. An overall temporal instability in the injury severity determinants was revealed via likelihood ratio tests. Several factors were observed to increase the likelihood of severe injuries, including old (age >55), collision with heavy vehicles, head-on collision, curves, dry road surfaces, and holidays. Additionally, unfavorable riding conditions were demonstrated to decrease the probability of severe injuries, which was attributed to the riders’ risk compensation mechanism in adverse conditions. This study provides new insights into the mechanism of motorcyclist injury severity in MN-V crashes.

Evaluating crash severity at highway-rail grade crossings using an analytic hierarchy process-based hazard index model.

Evaluating crash severity at highway-rail grade crossings using an analytic hierarchy process-based hazard index model.

How does distraction affect cyclists' severe crashes? A hybrid CatBoost-SHAP and random parameters binary logit approach.

How does distraction affect cyclists' severe crashes? A hybrid CatBoost-SHAP and random parameters binary logit approach.

Development of safety performance measures for different crashes severity at urban roundabouts

Development of safety performance measures for different crashes severity at urban roundabouts

Factors Affecting the Injury Severity of Head-On Crashes on Undivided Rural Roads under Different Weather Conditions

Factors Affecting the Injury Severity of Head-On Crashes on Undivided Rural Roads under Different Weather Conditions

A note on data segmentation, sample size, and model specification for crash injury severity modeling

A note on data segmentation, sample size, and model specification for crash injury severity modeling

Partially temporally constrained modeling of speeding crash-injury severities on freeways and non-freeways before, during, and after the stay-at-home order.

Partially temporally constrained modeling of speeding crash-injury severities on freeways and non-freeways before, during, and after the stay-at-home order.

A Hybrid Framework for Predicting Crash Severity in Construction Work Zones Using Knowledge Distillation and Conditional GANs

A Hybrid Framework for Predicting Crash Severity in Construction Work Zones Using Knowledge Distillation and Conditional GANs

Innovative prediction and causal analysis of accident vehicle towing probability using advanced gradient boosting techniques on extensive road traffic scene data.

Innovative prediction and causal analysis of accident vehicle towing probability using advanced gradient boosting techniques on extensive road traffic scene data.

Convolutional neural network-based sensitivity analysis of parent and child features of different vehicle types in single-vehicle crashes

This study proposes a new approach to better understand the risk factors of single-vehicle (SV) crashes across different vehicle models. Using four years (2019–2022) of SV crash data from Shandong Province, involving motorcycles, passenger cars, and trucks, we developed a sensitivity analysis model based on convolutional neural networks (CNNs) to analyze various factors influencing driver injury severity. The model successfully quantified significant parent and child features for different vehicle models, revealing key risk factors for SV crash severity. The results showed that nighttime (20:00 – next 6:59) significantly increased the risk of serious or fatal (FS) injuries for all vehicle types, with drink-driving behaviors at night further exacerbating the frequency of crashes. Additionally, specific groups—drivers aged 51 and over, self-employed individuals, and staff—face higher FS injury risks while riding motorcycles. For passenger cars, rural roads, intersections, without signal control also raise the risk of FS injuries. In trucks, the risk of FS injuries increases under conditions such as driving with only a central divider, in daylight, when overloaded, and on non-intersection roads. This method provides valuable support for the targeted prevention of serious SV crashes across different vehicle types.

Enhancing motorcycle safety: Quantifying the effects of Autonomous Emergency Braking and Adaptive Cruise Control in crashes reduction

Objective Powered 2-wheeler (PTW) safety remains a critical concern, prompting the investigation into the efficacy of advanced rider assistance systems (ARAS) in reducing crash rates. The objective of this study is to provide a comprehensive quantification of the effects of 2 such systems, Adaptive Cruise Control (ACC) and Autonomous Emergency Braking (AEB), on the avoidance or mitigation of motorcycle crashes. Methods Using real-world motorcycle crash data, this research employed kinematic reconstruction techniques and numerical simulations to evaluate the impact of ACC and AEB on crash reduction. A parametric study was conducted, exploring various parameters including trigger distances, deceleration rates, and field of view, to assess their influence on the effectiveness of this systems in crashes involving PTWs. Results The analysis revealed significant findings regarding the effectiveness of ACC and AEB in reducing crash rates involving motorcycles. Notably, ACC, especially in dynamic mode with a trigger distance of 40 m, demonstrated the ability to prevent 53% of crashes, resulting in impact speed reductions ranging from 4 to 25 km/h. Similarly, AEB exhibited remarkable effectiveness, reducing impact speeds by 2.5 to 38.9 km/h, with avoidance rates ranging from 7% to 63%. Conclusions Optimization of key parameters such as triggering strategy, deceleration, sensor range, and field of view is crucial for maximizing the effectiveness of ARAS in motorcycle safety applications. These findings provide valuable insights for manufacturers and policymakers, highlighting the need for tailored approaches to system design and implementation. By refining these systems based on empirical data and analysis, significant improvements can be made in motorcycle safety, ultimately saving lives and reducing the severity of crashes on roadways.

Comparison of Machine Learning Models to Predict Nighttime Crash Severity: A Case Study in Tyler, Texas, USA

Driving at night is riskier in terms of crash involvement than it is during the day. Fortunately, it is clearly established that illumination on roadways can reduce the number and severity of nighttime crashes. However, state and municipal departments of transportation (DOTs) lack the available illumination data. Therefore, the objective of this research is threefold, as follows: (i) to develop machine learning models that use readily available roadway characteristic data to predict the severity of nighttime crashes; (ii) determine the effect that illumination has on crash severity; and (iii) develop a tool to assist DOT decision makers in collecting illumination data. To accomplish this objective, we have extracted data from the Texas Department of Transportation (TxDOT) Crash Record Information System (CRIS) database, which was then further split into a training and a test dataset. Then, seven machine learning techniques, namely binary logistic regression, k-nearest neighbors, naïve Bayes, random forest, artificial neural network, Extreme Gradient Boosting (XGBoost), and a Long Short-Term Memory (LSTM) model, were all applied to the unseen test data. The random forest model produced the most promising results by predicting severe crashes with 97.6% accuracy. In addition, we conducted a pilot study to test the collection of illumination data using a light meter. In the future, we aim to complete the development of a smartphone application, which can be used in conjunction with the random forest model presented in this paper, to collect crowdsourced illumination data and predict nighttime crash hotspots. This may assist DOT decision makers to prioritize funding for illumination at the hot spots.

Exploring the Factors Affecting the Injury Severity of Crashes on Mountainous Freeways with High Proportions of Heavy Vehicles Considering Their Heterogeneous and Interactive Effects

Freeway transportation safety issues have attracted public concern in China for decades. This study aims to identify the factors influencing the injury severity of freeway crashes and to quantify their effects on the likelihood of various crash severity levels, with consideration of heterogeneity and interactions. The empirical analysis is based on three years of crash data from two mountainous freeways in Guangdong, China, covering the years of 2021 to 2023. A random parameters logit model with interaction terms is developed for the analysis. Goodness-of-fit indicators reveal that accommodating the interactive effects can significantly improve model fit performance. The estimation results of the parameters and marginal effects indicate that the factors related to vehicle type, time of day, crash season and cause, 3D curvature, and traffic volume have significant effects on crash severity. Notably, interactive effects are revealed between spring and evening, autumn and fixed objects, and non-local vehicles and improper driving. According to the findings, some countermeasures on safety education, traffic management, and freeway design are provided for preventing freeway crash injury, which is helpful for the development of sustainable transportation systems.

A scoping review on the methods used to assess health-related quality of life and disability burden in evaluations of road safety interventions.

Road traffic crashes globally cause 1.3 million deaths yearly and the rate of nonfatal crashes is increasing. Nonfatal injuries impact long-term quality of life, which is often overlooked in evaluations. The preferred method for using health-related quality of life and disability for evaluating road safety interventions have not been established. A scoping review of peer-reviewed and grey literature was undertaken to understand health-related quality of life and disability measures currently used to evaluate road safety interventions. We included English language studies that used any health-related quality of life or disability measure to evaluate any real-world intervention aimed at reducing the number or severity of road traffic crashes. Nine different health-related quality of life measures were used in the 18 included studies. The most commonly used measure was a quality-adjusted life year, which was used by seven studies, followed by the Glasgow Outcome Scale used by five studies. Two studies used two different health-related quality of life or disability measures. Five studies used primary data (collected directly for the purpose of the study) and 13 studies used existing data sources not explicitly collected for the reported evaluation. Of these 13 studies, 5 used an injury registry as the data source. Six different methods of deriving utility weights for calculating quality-adjusted life years were used. This review found that evaluations of road safety interventions using health-related quality of life or disability measures were rare. There was a lack of consistency in the measures used which prevented comparisons across evaluations. Further, inconsistent methods were used to derive utility weights for quality-adjusted life years. Future evaluations of roads safety interventions need to consider longer-term outcomes. Consistent methods for measuring health-related quality of life and disability burden are needed, as are empirically derived utility weights for quality-adjusted life years.

Temporal transferability assessment of injury severity models for single-vehicle and multi-vehicle crashes at highway ramp areas.

Temporal transferability assessment of injury severity models for single-vehicle and multi-vehicle crashes at highway ramp areas.

Analyzing Feature Importance for Older Pedestrian Crash Severity: A Comparative Study of DNN Models, Emphasizing Road and Vehicle Types with SHAP Interpretation

Analyzing Feature Importance for Older Pedestrian Crash Severity: A Comparative Study of DNN Models, Emphasizing Road and Vehicle Types with SHAP Interpretation

Comparability of driving automation crash databases.

Comparability of driving automation crash databases.

Examining Fatal Motorcycle Crashes in Malaysia: Rider Age, Road Attributes and Collision Partner Dynamics

Abstract: - High severity crashes involving motorcycles in Malaysia were analyzed to understand the crash characteristics, crash partner, and crash scenarios. This would involve studying how these variables influence the occurrence and severity of fatal crashes. Based on the examination, countermeasures are emphasized from the perspective of technologies, particularly those available on passenger vehicles. The objective of this study is to identify key risk factors and provide insights for improving road safety policies and intervention strategies. The highest age group involved in motorcycle crashes are among 16-20, followed by 21-25. 62.7% occurred in daylight, 90.5% during good weather and 75.4% occurred in midblock. By comparing the engine capacity to the total industry volume, the highest risk is on motorcycles>250 cc, as the production is only 0.29%, but contributes to 4.55% of fatal crashes. The highest crash partners are passenger cars at 30.84%, followed by trucks at 26.64% and single-vehicle accidents at 19.64%. In single crash accidents, road departure represents 56.7%, and control loss, 43.3%. The mandatory ABS in new motorcycles in Malaysia may help in reducing the number of stability losses during emergency braking. Except for rear-end crashes, passenger cars are the highest collision partner in all crash scenarios. Considering the frequency of crash scenarios and crash compatibility, crash avoidance technologies that could help reduce fatalities include autonomous emergency braking, blind spot detection, and lane-keeping assist.

Modeling Rollover Crash Risks: The Influence of Road Infrastructure and Traffic Stream Characteristics

Rollover crashes are among the most prevalent types of accidents in developing countries. Various factors may contribute to the occurrence of rollover crashes. However, limited studies have simultaneously investigated both traffic stream and road-related variables. For instance, the effects of T-intersection density, U-turns, roadside parking lots, the entry and exit ramps of side roads, as well as traffic stream characteristics (e.g., standard deviation of vehicle speeds, speed violations, presence or absence of speed cameras, and road surface deterioration) have not been thoroughly explored in previous research. Additionally, the simultaneous modeling of crash frequency and intensity remains underexplored. This study examines single-vehicle rollover crashes in Yazd Province, located in central Iran, as a case study and simultaneously evaluates all the variables. A dataset comprising three years of crash data (2015–2017) was collected and analyzed. A crash index was developed based on the weight of crash intensity, road type, road length (as dependent variables), and road infrastructure and traffic stream properties (as independent variables). Initially, the dataset was refined to determine the significance of explanatory variables on the crash index. Correlation analysis was conducted to assess the linear independence between variable pairs using the variance inflation factor (VIF). Subsequently, various models were compared based on goodness of fit (GOF) indicators and odds ratio (OR) calculations. The results indicated that among ten crash modeling techniques, namely, Poisson, negative binomial (NB), zero-truncated Poisson (ZTP), zero-truncated negative binomial (ZTNB), zero-inflated Poisson (ZIP), zero-inflated negative binomial (ZINB), fixed-effect Poisson (FEP), fixed-effect negative binomial (FENB), random-effect Poisson (REP), and random-effect negative binomial (RENB), the FENB model outperformed the others. The Akaike information criterion (AIC) and Bayesian information criterion (BIC) values for the FENB model were 1305.7 and 1393.6, respectively, demonstrating its superior performance. The findings revealed a declining trend in the frequency and severity of rollover crashes.