Analysis of the severity of vehicle-bicycle crashes with data mining techniques

Interpretable machine learning for evaluating risk factors of freeway crash severity

Machine learning (ML) models are widely employed for crash severity modelling, yet their interpretability remains underexplored. Interpretation is crucial for comprehending ML results and aiding informed decision-making. This study aims to implement an interpretable ML to visualize the impacts of factors on crash severity using 5 years of freeways data from Iran. Methods including classification and regression trees (CART), K-nearest neighbours (KNNs), random forest (RF), artificial neural network (ANN) and support vector machines (SVM) were applied, with RF demonstrating superior accuracy, recall, F1-score and ROC. The accumulated local effects (ALE) were utilized for interpretation. Findings suggest that light traffic conditions (volume / capacity < 0.5 ) with critical values around 0.05 or 0.38, and higher proportion of large trucks and buses, particularly at 10% and 4%, are associated with severe crashes. Additionally, speeds exceeding 90 km/h, drivers younger than 30 years, rollover crashes, collisions with fixed objects and barriers, nighttime driving and driver fatigue elevate the likelihood of severe crashes.

Analyzing the transition from two-vehicle collisions to chain reaction crashes: A hybrid approach using random parameters logit model, interpretable machine learning, and clustering

Chain reaction crashes (CRC) begin with a two-vehicle collision and rapidly intensify as more vehicles get directly involved. CRCs result in more extensive damage compared to two-vehicle crashes and understanding the progression of a two-vehicle collision into a CRC can unveil preventive strategies that have received less attention. In this study, to align with recent research direction and overcome the limitations of econometric and machine learning (ML) modelling, a hybrid approach is adopted. Moreover, to tackle the existing challenges in crash analysis, addressing unobserved heterogeneity in ML, and exploring random parameter effects and interactions more precisely, a new approach is proposed. To achieve this, a hybrid random parameter logit model and interpretable ML, joint with prior latent class clustering is implemented. Notably, this is the first attempt at using a clustering with hybrid modeling. The significant risk factors, their critical values, distinct effects, and interactions are interpreted using both marginal effects and the SHAP (SHapley Additive exPlanations) method across clusters. This study utilizes crash, traffic, and geometric data from eleven suburban freeways in Iran collected over a 5-year period. The overall results indicate an increased risk of CRC in congested traffic, higher traffic variation, and on horizontal curves combined with longitudinal slopes. Some parameters exhibit distinct or fluctuating effects, which are discussed across different conditions or considering interactions. For instance, during nighttime, heightened congestion on 2-lane freeways, increased traffic variation in less congested conditions, and adverse weather combined with horizontal curves and slopes pose risks. During daytime, increased traffic variation within highly congested sections, higher proportion of heavy vehicle traffic in moderately congested sections, and two lanes in each direction coupled with curves, elevate the levels of risk. The results of this study provide a better understanding of risk factors impact across different conditions, which are usable for policy makers.

Interpretable ensemble imbalance learning strategies for the risk assessment of severe-low-level wind shear based on LiDAR and PIREPs

The occurrence of severe low-level wind shear (S-LLWS) events in the vicinity of airport runways poses a significant threat to flight safety and exacerbates a burgeoning problem in civil aviation. Identifying the risk factors that contribute to occurrences of S-LLWS can facilitate the improvement of aviation safety. Despite the significant influence of S-LLWS on aviation safety, its occurrence is relatively infrequent in comparison to non-SLLWS incidents. In this study, we develop an S-LLWS risk prediction model through the utilization of ensemble imbalance learning (EIL) strategies, namely, BalanceCascade, EasyEnsemble, and RUSBoost. The data for this study were obtained from PIREPs and LiDAR at Hong Kong International Airport. The analysis revealed that the BalanceCascade strategy outperforms EasyEnsemble and RUSBoost in terms of prediction performance. Afterward, the SHapley Additive exPlanations (SHAP) interpretation tool was used in conjunction with the BalanceCascade model for the risk assessment of various factors. The four most influential risk factors, according to the SHAP interpretation tool, were hourly temperature, runway 25LD, runway 25LA, and RWY (encounter location of LLWS). S-LLWS was likely to happen at Runway 25LD and Runway 25LA in temperatures ranging from low to moderate. Similarly, a high proportion of S-LLWS events occurred near the runway threshold, and a relatively small proportion occurred away from it. The EIL strategies in conjunction with the SHAP interpretation tool may accurately predict the S-LLWS without the need for data augmentation in the data pre-processing phase.

Investigating pedestrian-vehicle crashes on interstate highways: Applying random parameter binary logit model with heterogeneity in means

In the U.S., the interstate highway system is categorized as a controlled-access or limited-access route, and it is unlawful for pedestrians to enter or cross this type of highway. However, pedestrian-vehicle crashes on the interstate highway system pose a distinctive safety concern. Most of these crashes involve 'unintended pedestrians', drivers who come out of their disabled vehicles, or due to the involvement in previous crashes on the interstate. Because these are not 'typical pedestrians', a separate investigation is required to better understand the pedestrian crash problem on interstate highways and identify the high-risk scenarios. This study explored 531 KABC (K = Fatal, A = Severe, B = Moderate, C = Complaint) pedestrian injury crashes on Louisiana interstate highways during the 2014-2018 period. Pedestrian injury severity was categorized into two levels: FS (fatal/severe) and IN (moderate/complaint). The random parameter binary logit with heterogeneity in means (RPBL-HM) model was utilized to address the unobserved heterogeneity (i.e., variations in the effect of crash contributing factors across the sample population) in the crash data. Some of the factors were found to increase the likelihood of pedestrian's FS injury in crashes on interstate highways, including pedestrian impairment, pedestrian action, weekend, driver aged 35-44 years, and spring season. The interaction of 'pedestrian impairment' and 'weekend' was found significant, suggesting that alcohol-involved pedestrians were more likely to be involved in FS crashes during weekends on the interstate. The obtained results can help the 'unintended pedestrians' about the crash scenarios on the interstate and reduce these unexpected incidents.

Self-reported cycling behavior and previous history of traffic accidents of cyclists

BACKGROUND

Cyclists are vulnerable traffic users and studying the cycling behavior of professional and elite cyclists, their previous history of traffic accidents combined with the current knowledge on high-risk behaviors of this group can be a useful basis for further studies on ordinary cyclists. This study aimed to determine the relationship between cycling behavior and the previous history of traffic accidents among members of the Cycling Federation of Guilan province in 2022.

METHODS

A descriptive-analytical study was performed in which the Bicycle Rider Behavior Questionnaire (BRBQ) constructed in the Porsline platform was distributed using the WhatsApp social network. All participants were asked to self-report their cycling behavior. The final analysis was performed by using STATA software (version 14).

RESULTS

The study subjects included a total of 109 cyclists with a mean age of 38.62 ± 10.94 years and a mean cycling experience of 13.75 ± 11.08 years. Using the logistic regression model, the relationship between gender (P = 0.039), years of cycling experience (P = 0.000), and education level (P ≤ 0.00), with previous traffic accidents, was found significant. There was also a significant relationship between stunts and distractions (P = 0.005), signaling violation (P = 0.000), and control error (P = 0.011) with previous traffic accidents. A significant association existed between stunts and distractions (P = 0.001) and signaling violation (P = 0.001) with a previous history of traffic injury within the last 3 years.

CONCLUSIONS

The findings of this study can be used to establish cyclist safety and preventative planning in society. In behavior change intervention programs, it is best to target male cyclists with higher-level education. In addition, the behavior of the cyclists whose predominant term of signaling violations must be corrected should be targeted. It is necessary to shape information campaigns and educational programs aimed for cyclists with common high-risk behaviors, especially signaling violations.

Injury severity analysis of single-vehicle and two-vehicle crashes with electric scooters: A random parameters approach with heterogeneity in means and variances

In recent years, the electric scooter has become one of the most popular means of transportation on short trips. Due to the lag in the formulation of transportation policies and regulations, coupled with the increasing number of electric scooter crashes, there has been growing concern about the safety of pedestrians and electric scooter riders. For the first time in the extant literature, this study aims to analyze injury severity of electric scooter crashes by unobserved heterogeneity modeling approaches. A random parameters approach with heterogeneity in means and variances is utilized to examine the factors influencing injury severity, using data collected from the STATS19 road safety database. Electric scooter crashes are classified as single-vehicle crashes and two-vehicle crashes, with injury severity categorized into two groups: fatalities or serious injuries, and slight injuries. The model estimation was conducted by considering several variables including roadway, environment, temporality, vehicle, and rider characteristics, as well as second-party vehicle and driver characteristics and manners of collision specific to two-vehicle crashes. The results of the model estimation reveal that certain factors had relatively stable effects with the varying degree of crash injury severity outcomes in both single-vehicle crashes and two-vehicle crashes. These factors include nighttime incidents, weekdays, male riders, and an increase in rider age, all of which are associated with more severe injury outcomes. Moreover, the random parameters logit model with heterogeneity in means and variances is more flexible in accounting for unobserved heterogeneity and exhibits better goodness of fit. This study improves the understanding of electric scooter safety, and the finding can better inform public policy regarding electric scooter use to improve road safety and reduce injury severity of electric scooter crashes.

Crash injury severity prediction considering data imbalance: A Wasserstein generative adversarial network with gradient penalty approach

For each road crash event, it is necessary to predict its injury severity. However, predicting crash injury severity with the imbalanced data frequently results in ineffective classifier. Due to the rarity of severe injuries in road traffic crashes, the crash data is extremely imbalanced among injury severity classes, making it challenging to the training of prediction models. To achieve interclass balance, it is possible to generate certain minority class samples using data augmentation techniques. Aiming to address the imbalance issue of crash injury severity data, this study applies a novel deep learning method, the Wasserstein generative adversarial network with gradient penalty (WGAN-GP), to investigate a massive amount of crash data, which can generate synthetic injury severity data linked to traffic crashes to rebalance the dataset. To evaluate the effectiveness of the WGAN-GP model, we systematically compare performances of various commonly-used sampling techniques (random under-sampling, random over-sampling, synthetic minority over-sampling technique and adaptive synthetic sampling) with respect to dataset balance and crash injury severity prediction. After rebalancing the dataset, this study categorizes the crash injury severity using logistic regression, multilayer perceptron, random forest, AdaBoost and XGBoost. The AUC, specificity and sensitivity are employed as evaluation indicators to compare the prediction performances. Results demonstrate that sampling techniques can considerably improve the prediction performance of minority classes in an imbalanced dataset, and the combination of XGBoost and WGAN-GP performs best with an AUC of 0.794 and a sensitivity of 0.698. Finally, the interpretability of the model is improved by the explainable machine learning technique SHAP (SHapley Additive exPlanation), allowing for a deeper understanding of the effects of each variable on crash injury severity. Findings of this study shed light on the prediction of crash injury severity with data imbalance using data-driven approaches.

Analyzing the injury severity in single-bicycle crashes: An application of the ordered forest with some practical guidance

This paper investigates the injury severity of cyclists in single-bicycle crashes (SBCs) in the UK. The data for analysis is constructed from the STATS19 road traffic casualty database, covering the period of 2016-2019. A machine learning-based ordered choice model termed Ordered Forest (ORF) is used. In our empirical analysis, ORF is found to produce more accurate class predictions of the SBC injury severity than the traditional random forest algorithm. Moreover, the factors associated with the injury severity are revealed, including the time and location of occurrence, the age of cyclists, roadway conditions, and crash-related factors. Specifically, old cyclists are more likely to be seriously injured in SBCs. Rural areas, higher speed limits, run-off crashes, and hitting objects are also related to an increased probability of serious injuries. While SBCs occurring at junctions, and/or during peak hours (i.e., 6:30-9:30 and 16:00-19:00) are less severe. To achieve the ambition of a step change in cycling and walking put forward by the UK Department for Transport, SBCs deserve more public attention. Lastly, regarding the implementation of ORF in crash injury severity analysis, we provide some practical guidance based on a series of simulation experiments.

Systematic literature review of 10 years of cyclist safety research

Cyclist safety is a research field that is gaining increasing interest and attention, but still offers questions and challenges open to the scientific community. The aim of this study was to provide an exhaustive review of scientific publications in the cyclist safety field. For this purpose, Bibliometrix-R tool was used to analyse 1066 documents retrieved from Web of Science (WoS) between 2012 and 2021. The study examined published sources and productive scholars by exposing their most influential contributions, presented institutions and countries most contributing to cyclist safety and explored countries open towards international collaborations. A keywords analysis provided the most frequent author keywords in cyclist safety shown in a word cloud with E-bike, behaviour, and crash severity representing the primary keywords. Furthermore, a thematic map of cyclist safety field drafted from the author's keywords was identified. The strategic diagram is divided in four quadrants and, according to both density and centrality, the themes can be classified as follows: 1) motor themes, characterized by high value of both centrality and density; 2) niche themes, defined by high density and low centrality; 3) emerging or declining themes, featured by low value of both centrality and density; and 4) basic themes, distinguished by high centrality and low density. The motor themes (i.e., the main topics in cyclist safety field) crash severity and bike network were further explored. The research findings will be useful to develop strategies for making bike a safer and more confident form of transport as well as to guide researchers towards the future scientific knowledge.

Risk-Compensation Trends in Road Safety during COVID-19

The COVID-19 pandemic has had a global impact, disrupting the normal trends of our everyday life. More specifically, the effects of COVID-19 on road safety are still largely unexplored. Hence, this study aims to investigate the change in road safety trends due to COVID-19 using real-time traffic parameters. Results from the extensive analyses of the 2017 to 2020 data of Interstate-4 show that traffic volume decreased by 13.6% in 2020 compared to the average of 2017–2019’s volume, whereas there is a decreasing number of crashes at the higher volume. Average speed increased by 11.3% during the COVID-19 period; however, the increase in average speed during the COVID-19 period has an insignificant relationship with crash severities. Fatal crashes increased, while total crashes decreased, during the COVID-19 period; severe crashes decreased with the total crashes. Alcohol-related crashes decreased by 22% from 2019 to 2020. Thus, the road-safety trend due to the impact of COVID-19 has evidently changed and presents a unique trend. The findings of the study suggest a larger need for a more in-depth study to analyze the impact of COVID-19 on road safety, to minimize fatalities on roads through appropriate policy measures.

An application of in-vehicle recording technologies to analyze injury severity in crashes between taxis and two-wheelers

Conventionally, the crash data used in traffic safety analysis have been collected by the police dispatched to the crash scene. Therefore, crash information inevitably includes errors that influence traffic safety analysis. Such errors can include the crash speed, crash time, crash location, and other crash characteristics. The advances in in-vehicle video recording (IVVR) technologies have recently enabled traffic safety professionals to use more accurate crash information based on crash data reconstruction methods. Although a few studies have been conducted to identify the factors affecting the crash injury severity using such detailed crash data, there was no effort to analyze the factors affecting the injury severity in crashes between taxis and two-wheelers (TWs), including bicycles and motorcycles. Therefore, this study analyzes the injury severity of TW riders in taxi-TW crashes with the accurate crash data collected by taxis equipped with IVVR devices in Incheon, Korea. Two hundred and forty-eight crash data from two years (2010-2011) were used to perform this objective. The factors affecting the injury severity to TW riders were identified based on a partial proportional odds model for these data. Seven variables were found to affect the injury severity significantly: crash speed, second collision, third collision, Delta-V, crashes that occurred with a non-helmeted motorcycle rider, crashes where the collision type was sideswipe, and crashes under rainy or snowy weather conditions. On the other hand, two variables regarding crashes, where the taxi driver behavior helped reduce visible and severe injuries, were changing lanes and the young TW riders (<18 years).

Analysis of Bicycle-Motor Vehicle Crashes in San Antonio, Texas

Bicycling is inexpensive, environmentally friendly, and healthful; however, bicyclist safety is a rising concern. This study investigates bicycle crash-related key variables that might substantially differ in terms of the party at fault and bicycle facility presence. Employing 5 year (2014-2018) data from the Texas Crash Record and Information System database, the effect of these variables on bicyclist injury severity was assessed for San Antonio, Texas, using bivariate analysis and binary logistic regression. Severe injury risk based on the party at fault and bicycle facility presence varied significantly for different crash-related variables. The strongest predictors of severe bicycle injury include bicyclist age and ethnicity, lighting condition, road class, time of occurrence, and period of week. Driver inattention and disregard of stop sign/light were the primary contributing factors to bicycle-vehicle crashes. Crash density heatmap and hotspot analyses were used to identify high-risk locations. The downtown area experienced the highest crash density, while severity hotspots were located at intersections outside of the downtown area. This study recommends the introduction of more dedicated/protected bicycle lanes, separation of bicycle lanes from the roadway, mandatory helmet use ordinance, reduction in speed limit, prioritization of resources at high-risk locations, and implementation of bike-activated signal detection at signalized intersections.