Motorcycle Crash Severity Research Articles

Identification of Factors Associated with Both Frequency and Severity of Deer-Related Motorcycle Crashes.

Motorcyclists are particularly vulnerable to injury when encountering large animals such as deer in the roadway. The frequency of these events, factors associated with their occurrence, and the extent of injuries sustained from such crashes are not yet well described. A review of 10 years (2012-2021) of statewide motorcycle crash reports from a single state with a large deer population was performed to ascertain the frequency of deer-associated motorcycle crashes. Environmental and roadway conditions, as well as motorcyclist age, gender, and helmet usage, were identified to determine if associations exist with the severity of injuries sustained. Statewide, 8.5% of reported motorcycle crashes involved deer. Excluding the two most densely populated counties, this rose to 10.7%, with 24.1% of crashes occurring during non-daylight hours being deer related. Factors associated with increased frequency of more severe injury and death included not wearing a helmet, female gender, and rural county crash location. Death and severe injuries were reported in more than twice as many riders who were unhelmeted compared to those wearing helmets (36.1 vs 16.6%; p < 0.01). In a single midwestern state, deer represent a significant danger to motorcyclists, especially in rural settings during non-daylight hours. A cautious approach to riding during this time, utilizing available safety precautions to minimize risk of severe injury and death, and further exploration of viable interventions to reduce the likelihood of these occurrences should be considered.

Factors Affecting Single and Multivehicle Motorcycle Crashes: Insights from Day and Night Analysis Using XGBoost-SHAP Algorithm

This study aimed to identify and compare the risk factors associated with motorcycle crash severity during both daytime and nighttime, for single and multivehicle incidents in Thailand using 2021–2024 data. The research employed the XGBoost (Extreme Gradient Boosting) method for statistical analysis and extensively examined the temporal instability of risk factors. The results highlight the importance of features impacting the injury severity of roadway collisions across various conditions. For single motorcycle crashes, the key risk factors included speeding, early morning incidents, off-road events, and long holidays. In multivehicle crashes, rear-end collisions, interactions with large vehicles, and collisions involving other motorcycles or passenger cars were linked to increased injury severity. The findings indicate that the important factors associated with motorcyclist injury severity in roadway crashes vary depending on the type of crash and time of day. These insights are valuable for policymakers and relevant authorities in developing targeted interventions to enhance road safety and mitigate the incidence of severe and fatal motorcycle crashes.

Bivariate-Logit-Based Severity Analysis for Motorcycle Crashes in Texas, 2017–2021

Due to the number of severe traffic collisions involving motorcycles, a comprehensive investigation is required to determine their causes. This study analyzed Texas crash data from 2017 to 2021 to determine who was at fault and how various factors affect the frequency and severity of motorcycle collisions. Moreover, the study tried to identify high-risk sites for motorcycle crashes. Utilizing bivariate analysis and logistic regression models, the study investigated the individual and combined effects of several variables. Heat maps and hotspot analyses were used to identify locations with a high incidence of both minor and severe motorcycle crashes. The survey showed that dangerous speed, inattention, lane departure, and failing to surrender the right-of-way at a stop sign or during a left turn were the leading causes of motorcycle crashes. When a motorcyclist was at fault, the likelihood of severe collisions was much higher. The study revealed numerous elements as strong predictors of catastrophic motorcycle crashes, including higher speed limits, poor illumination, darkness during the weekend, dividers or designated lanes as the principal road traffic control, an increased age of the primary crash victim, and the lack of a helmet. The concentration of motorcycle collisions was found to be relatively high in city cores, whereas clusters of severe motorcycle collisions were detected on road segments beyond city limits. This study recommends implementing reduced speed limits on high-risk segments, mandating helmet use, prioritizing resource allocation to high-risk locations, launching educational campaigns to promote safer driving practices and the use of protective gear, and inspecting existing conditions as well as the road geometry of high-risk locations to reduce the incidence and severity of motorcycle crashes.

Investigating the risk factors of motorcycle crash injury severity in Pakistan: Insights and policy recommendations

A large number of fatalities and severe injuries are caused by motorcycle crashes worldwide, particularly in developing countries. More than 50% of crashes in Pakistan involve motorcycles. To analyze motorcycle crash severity, various models, including both statistical and machine learning methods, have been applied. Researchers have widely acknowledged that machine learning methods provide superior prediction performance but have weaker interpretability power. However, no study has investigated the consistency of risk factors identified by the two streams of models. The consistency of the findings between these two kinds of methods is vital to improve the interpretability power of machine learning methods for policymaking in an era with more and more applications in the area of traffic safety. This study aims to narrow this research gap by comparing the consistency of crash severity risk factors identified by statistical models and machine learning methods. The study analyzes motorcycle crashes in Rawalpindi city of Pakistan. Multinomial logit model (MNL) and three machine learning models, i.e., the random forest (RF), naive Bayes, and gradient-boosted trees methods, are used to analyze the prediction performance and identify risk factors. The results show that the RF model, with an overall accuracy of 86.7%, outperformed other models. The SHapley Additive exPlanations (SHAP) method was adopted to explore the interpretability of machine learning methods. It was found that the contributing factors to crash injury severity identified by the RF method, such as distracted driving, collisions involving pedestrians, collisions involving a truck, and female riders, are consistent with those determined by the MNL model. These results have clear implications for developing cost-effective safety countermeasures to improve motorcycle safety in Pakistan.

Modeling the Motorcycle Crash Severity on Nonintersection Urban Roadways in the Australian State of Victoria Using a Random Parameters Logit Model

Due to a lack of physical protection and balance, motorcycle riders are one of the most vulnerable road users and are more likely to suffer severe injuries than motorists. Between 2009 and 2020, about 60% of motorcycle crashes occurred on nonintersection urban roadways in Victoria, Australia. While considerable research on intersections and their influence on the severity of motorcycle crashes has been conducted, there are limited studies on motorcycle crashes on nonintersection roadways. Since gathering all information from every motorcycle crash may not be possible, heterogeneity can arise from unobserved factors and cause problems in developing reliable crash severity models. Therefore, this study aims to investigate the factors contributing to motorcycle crash severity on Victorian nonintersection urban roadways while considering the heterogeneity of factors. A total of 10,897 nonintersection motorcycles crash data from the beginning of 2009 to November 2020 in the State of Victoria, Australia, were analyzed. A random parameters (mixed) logit model (RPL) was used for evaluating motorcycle crashes. The severity of motorcycle crashes was divided into three categories: fatal injury, serious injury, and minor injury. Also, marginal effects were calculated to see how each parameter estimate affects crash severity outcomes. The RPL model results showed that some factors increased the likelihood of fatal injuries. These factors included not wearing a helmet, being in the older rider age group, riding during the early morning or midnight hours, weekend motorcycle use, riding in the early morning or midnight hours (00:00–6:29 A.M), and insufficient lighting (dark and dusk/dawn). Also, the following factors enhanced the probability of serious injuries: having a pillion passenger, having a motorcycle age of more than 7 years, riding at higher speed limits (more than 50 km/h) or during peak hours in the morning (6:30−8:59 A.M), and being in the younger age group (less than 26 years old). The findings from this study are valuable resources for road safety policy managers to develop effective strategies for improving motorcyclists’ safety at nonintersections. This may include improving the light conditions at nonintersection, encouraging the motorcyclist to maintain motorcycles regularly, and educating the motorcyclist to wear a helmet, avoid distractions, and ride responsibly on the weekends.

Bayesian Network for Motorcycle Crash Severity Analysis

Given the lack of protective structural barriers and advanced restraints, motorcyclists are vulnerable road users. In 2020 in the United States, motorcycle-involved fatalities occurred 28 times more frequently per vehicle mile traveled than passenger car occupant fatalities, causing 5,579 motorcycle-related fatalities—the highest number of motorcyclists killed since 1975. By identifying patterns and relationships between key contributing factors, strategies for reducing motorcycle crashes can be developed. In addition to current efforts, additional research must be conducted using innovative avenues, with increased funding. Bayesian networks can better discover the relationships between potential speed compliance variables. This study used six years (2014 to 2019) of motorcycle crash data in Louisiana to determine the conditional probabilities of the influential factors. In addition to the high contribution of alcohol involvement, two-way undivided roadways, 35 to 44 year-old drivers involved in improper driving behaviors, and crash types are the underlying factors associated with a considerable increase in motorcycle crash severity. The findings of this study can also be used for decision making and strategy development for motorcycle safety.

Factors Associated With the Severity of Motor Vehicle Crashes Involving Electric Motorcycles and Electric Bicycles: A Random Parameters Logit Approach With Heterogeneity in Means

Exploring the heterogeneity of factors influencing the severity of electric bicycle crashes and electric motorcycle crashes can help target accident prevention policies to improve traffic safety. Therefore, this paper establishes a mean heterogeneity random parameter logit model using crash data from 2016 to 2020 in Guangxi to explore the different factors influencing the severity of crashes involving electric motorcycles and electric bicycles. The results show that the key influences on crash severity differ in electric motorcycle crashes and electric bicycle crashes. At the same time, some common factors affect the two types of crashes to different degrees. In addition, the complex interaction effects of unobserved heterogeneity were considered to explore the random parameters of the two types of crashes. The effect of unobserved heterogeneity on the distribution characteristics of the random parameters was then determined. For example, in electric motorcycle crashes, street lighting at night has a random parameter characteristic. The likelihood of serious crashes decreased when both street lighting at night and vehicle left turn were involved, and decreased when both street lighting at night and no signal control were involved. In electric bicycle crashes, large trucks have a random parameter characteristic. The likelihood of serious crashes increased when both large trucks and motor vehicle lights not turned on were involved, and increased when both large trucks and visibility less than or equal to 200 meters were involved. The results provide a basis for improving the road safety of electric motorcycles and electric bicycles.

A Comparison of Contributing Factors between Young and Old Riders of Motorcycle Crash Severity on Local Roads

This study aims to identify the factors that influence the severity of motorcycle crashes on local roads, particularly given the high speeds often observed for motorcycles on these roads with low traffic volumes and numerous multi-leg intersections. Previous research has shown that a rider’s age can impact their speed behavior. To explore this issue, data on motorcycle crashes from 2015 to 2020 in Thailand—a middle-income developing country—were analyzed using a random parameter logit model with unobserved heterogeneity in means and variances, comparing young (&lt;30-year-old) and older (&gt;50-year-old) riders. The contributing factors were divided into four groups: driver, crash, environmental, and road factors. The transferability test yielded different results for the young rider and old rider models, indicating that it is appropriate to analyze these models separately. A constant value revealed that old riders were more likely to die in a crash than young riders. In terms of the random parameter, the local address and road surface variables were found to be significant in both models. The results of unobserved heterogeneity in means and variances identified significant variables in both models, including gender, exceeding the speed limit, lit roads, unlit roads, mobile phone use, and road surface. These findings were used to develop policy recommendations for reducing the severity of motorcycle crashes on local roads.

Factors contributing to the severity of motorcycle rear-end crashes in Thailand

Objective Motorcycle (MC) rear-end collisions cause many serious injuries and deaths for MC riders. In Thailand, the MC crash investigation data revealed that 18% of all MC crashes were rear-end collisions, which accounted for 18% of all fatalities as well. The aim of this study was to investigate the causes of injuries and deaths from MC rear-end collisions and factors that contribute to their severity level. Between 2016 and 2020, 141 MC rear-end crashes were thoroughly investigated throughout Thailand. Method The ordinal logistic analysis was conducted to analyze factors contributing to severe injuries. The analysis to rear-end collision models comprised four categories: M1 (n = 141) is all types of rear-end collisions to MC, and M2 (n = 114) is the rear-end collision due to other vehicles (OV) collided by MC, M3 (n = 72) is the rear-end collisions for traveling OV collided by MC, and M4 (n = 42) is the rear-end collision for MC hitting the parked OV. The outcomes are verified by the likelihood and Pseudo R2. Result When a MC collides with the rear of another vehicle, there are more fatalities than when other vehicles collide with the rear of a MC. Furthermore, the probability of death is higher if MCs collide with the rear-end of parking vehicles. As for the primary crash contributing factor, motorcyclists’ perception failure was the most frequent. Experience, license status, driving conditions, speed, the time of the crash, the areas of the crash, and types of other vehicles involved significantly influence the severity of rear-end crashes. Conclusion In severe crashes, riders with perception failure are more likely to be involved. Based on the findings of this study, some policies and countermeasures can be drawn to prevent MC rear-end crashes and reduce their severity.

Reliability analysis of motorcycle crash severity outcomes: Consideration of model selection uncertainty

Objective While a large amount of work has been conducted on different types of crash injury severity models, model selection uncertainty remains a critical issue in traffic safety research. The objective of this study is to handle model selection uncertainty by combining multiple models. Methods Motorcycle crashes in Michigan from 2010 to 2014 are collected for the analysis. A model averaging approach is used to integrate useful information from three commonly used crash injury severity models: multinomial logit model, ordered logit model, and ordered probit model to deal with the situation where the model selection uncertainty exists in crash data analysis. The ratios of model posterior probabilities between models are used to quantify the model selection uncertainty. In addition, the effectiveness of the method is illustrated by comparing it with the single-best model. Results The ratios of model posterior probabilities among models approximate to 1. It means that three models have the same importance in statistical analysis of motorcycle injury severity, resulting in model selection uncertainty. The comparison between the results of model averaging approach and single-best model shows that the single-best model tends to overestimate the effects of risk factors on motorcycle injury severities because of ignoring the model selection uncertainty; parameter errors and confidence intervals of model averaging are greater and wider than those of the single-best model due to between-model uncertainty included in the model averaging; some risk factors are significant in the model averaging approach while not in the single-best model. Results from model averaging approach reveal that drunk or riding under influence, angle/sideswipe/head on crashes, speed limit of 35 mph or higher, and signal control play significant roles in the motorcycle crashes. Conclusions The study contributes to the existing crash injury-severity literature by developing a model averaging approach to explore the relationship between motorcyclist’s injury-severity and its contributing factors. The model averaging approach overcomes the limitations of the current crash injury-severity modeling approaches by (1) revealing the potential model selection uncertainty among injury-severity models with model posterior probabilities; (2) more reliably accounting for the effects of risk factors on motorcyclist’ injury severities through integrating all information from the candidate models; and (3) better presenting the underlying unreliability of the analysis results from each individual model.

A comparative study of factors associated with motorcycle crash severities under different causal scenarios

This study was carried out to examine the factors associated with motorcycle crash severity in Alabama, under different manner of crash and causal scenarios using mixed logit modeling. Three crash mechanisms were considered in this study: single-vehicle motorcycle crash with motorcyclist at fault, multi-vehicle collision between a motorcycle and another vehicle with motorcyclist being at fault, and motorcyclist not at fault in a collision between a motorcycle and another vehicle. The model estimation results showed that crashes that happened in rural areas were more likely to be severe, irrespective of the causal unit or manner of collision. The results also show that fatigue among motorcyclists was associated with severe injury, whereas driver fatigue was linked to no injury outcome. Further, it was found that risky behaviors such as speeding, driving/riding under the influence of alcohol or drugs, driving/riding with invalid license were significantly associated with severe injury outcome. Developing the injury-severity models based on the segmented crash data has helped to reveal some similarities and differences in crash outcomes based on the crash mechanism and the at-fault road user. It is expected that these findings would provide a data-driven evidence to improve motorcycle safety in the state.

An analysis of motorcyclists' injury severities in work-zone crashes with unobserved heterogeneity

Work-zone related crashes in Florida are increasing steadily in recent years. With such growing concern, work-zone is recognized in Florida Strategic Highway Safety Plan. The severity and complexity of motorcycle crashes in work-zones is critically important and worth investigating. However, the resulting effect of work-zone on motorcyclists' injuries in work-zone crashes is not fully understood. The purpose of study is to identify the contributing factors of motorcyclists' injury severity sustained in the work-zone crashes in Florida. Recognizing the relatively higher risk of motorcyclists in work-zones with respective to non-work-zones, this study further uncovers the contributing factors for single- and multi-vehicle motorcycle crashes in Florida work-zones. This study investigated motorcyclists' injury severity applying random parameter multinomial logit with possible heterogeneity in means and variances of the random parameters for single-motorcycle and multi-vehicle motorcycle crashes. This study utilizes the Crash Analysis Reporting (CAR) system in Florida over a period of five years from 2012 to 2016 (inclusive). The model result indicates a complex relationship between dark condition, old-aged motorcyclist (50–65), requirement and absence of endorsement, partial ejection, straight roadway segment, shoulder width (up to 1.22 m (4 ft), and 2.74–3.66 m (9–12 ft), urban interstate, activity area, and lane closure and work on shoulder-median work-zone types. The effect of work-zone on single-motorcycle crashes tends to have much more in resulting injury severities relative to multi-vehicle motorcycle crashes. It is more important to investigate the injury severity by single- and multi-vehicle crashes involving motorcycles in work-zones. These risk factors identified in the study are expected to provide more insights for the countermeasures specific to engineering (roadway design) and policy (motorcycle training), which can be considered to improve motorcycle safety in Florida.

Characteristics of crashes and injuries on delivery motorcycles: A retrospective cohort study

Objectives: The food delivery market is growing rapidly. As most delivery riders use motorcycles, motorcycle crashes will increase along with the growing delivery market size. This study aimed at examining the proportions of motorcycle crashes and characteristics of injuries incurred while using motorcycles for occupational purposes. Methods: This retrospective analysis included motorcycle crash patients aged 16 years or older, who were treated in 23 emergency rooms in Korea, between 2014 and 2018. Patients were divided into two groups: delivery riders (delivery group) and others (nondelivery group). Crash and injury characteristics were compared between the two groups. In addition, trends of patients in the delivery group were compared from 2014 to 2018. Results: This study examined 26,982 motorcycle crash patients, including 3894 (14.43%) patients in the delivery group and 23,088 (85.57%) in the nondelivery group. The number of patients in the delivery group increased drastically from 583 in 2014 to 1029 in 2018, whereas the number of patients in the nondelivery group did not considerably increase (4411 in 2014 and 4462 in 2018). The delivery group had a higher proportion of crashes caused by collisions with cars or other motorcycles (p < 0.001); however, injury severity was lower. The delivery group had a lower proportion of head and face injuries but a higher proportion of extremity injuries. Furthermore, 39.9% of all crashes in this group occurred between 17:00 and 21:00. Over time, there were neither any changes in the injury severities, nor any changes in the characteristics of the delivery group, with the exception of increases in both the proportion of motorist insurance and the proportion of wearing a helmet. Conclusions: The results indicated differences in characteristics between delivery motorcycle crashes and other motorcycle crashes. Although delivery motorcycle crash severity was low compared to other motorcycle crashes, the number of patients increased significantly. Therefore, to prevent crashes, it is necessary to improve the working environment. In addition, to prevent the extremity injuries of delivery riders, the policy of wearing extremity protective gears should be considered.

Understanding the Factors That Are Associated with Motorcycle Crash Severity in Rural and Urban Areas of Ghana

Like many countries in sub-Saharan Africa, Ghana has witnessed an increase in the use of motorcycles for both commercial transport and private transport of people and goods. The rapid rise in commercial motorcycle activities has been attributed to the problem of urban traffic congestion and the general lack of reliable and affordable public transport in rural areas. This study investigates and compares factors that are associated with motorcycle crash injury outcomes in rural and urban areas of Ghana. This comparison is particularly important because the commercial use of motorcycles and their rapid growth in urban areas are a new phenomenon, in contrast to rural areas where people have long relied on motorcycles for their transportation needs. Preliminary analysis of the crash data revealed that more of the rural area crashes occurred under dark and unlit roadway conditions, while urban areas recorded more intersection-related crashes. Additionally, it was found that more pedestrian collisions happened in urban areas, while head-on collisions happened more in rural areas. The model estimation results show that collisions with a pedestrian, run-off-road, and collisions that occur under dark and unlit roadway conditions were more likely to result in fatal injury. Findings from this study are expected to help in crafting and targeting appropriate countermeasures to effectively reduce the occurrence and severity of motorcycle crashes throughout the country and, indeed, sub-Saharan Africa.

Modeling the injury severity of small-displacement motorcycle crashes in Hanoi City, Vietnam

Hanoi City is a typical motorcycle dominant city in Vietnam with a modal share of motorcycles of more than 80% and motorcycle riders involved in more than 70% of road fatalities. It is important to understand the major causes of motorcycle crash severity to suggest effective countermeasures. There have been studies elsewhere examining factors related to small-displacement motorcycle crash severity, but work in Vietnam is scarce. This study uses ordinal logistic regression to investigate contributory factors to the odds of severe crashes involving motorcycle riders. A distinction is made between crashes where motorcyclists are involved as the second party (i.e., as victims of the crash) on the one hand and as the first party (i.e., as the person responsible for causing the crash) on the other hand. Results show that age and gender of the first party person, size of the vehicle collided with, and presence of a pillion passenger significantly contribute to the severity of crashes in which motorcyclists are involved as the second party. Significant factors to the severity of crashes in which motorcyclists are involved as the first party are age, time of the crash, road class, size of the second party vehicle, pillion passenger presence, and riding speed. The findings contribute to a deeper understanding of fatal crashes with small-displacement motorcycles in a developing city. This study helps to identify countermeasures to improve road safety of small-displacement motorcycle riders.

Identifying Factors Contributing to the Motorcycle Crash Severity in Pakistan

Motorcycle is a popular mode of transportation in many developing countries, including Pakistan. Since the last decade, the registered number of motorcycles in Pakistan has increased by six times, constituting 74% of the total registered vehicles. However, limited research efforts have been made to investigate motorcycle-related safety issues in Pakistan. Thus, the relationship between potential risk factors and injury outcomes of motorcycle crashes is still unclear in the country. This study, therefore, established a random parameter logit model to examine the factors associated with the motorcycle injury severity. The analysis is based on two years (2014–2015) of data collected through the road traffic injuries surveillance system from Karachi city, Pakistan. The results indicate that the summer season, weekends, nighttime, elderly riders, heavy vehicle, and single-vehicle collisions are positively associated with fatalities, while the presence of pillion passengers and motorcycle-to-motorcycle crashes are negatively associated with fatalities. More importantly, in the specific context of Pakistan, morning hours, young riders, and female pillion passengers whose clothes stuck in the wheel significantly increase the fatal injury outcomes. Based on the findings, potential countermeasures to improve motorcycle safety are discussed, such as strict enforcement to control motorcyclists' risky behavior and speeding, provision of exclusive motorcycles lanes, and education of female pillion passengers. The findings from this study would increase awareness of motorcycle safety and can be used by the policymakers to enhance road safety in Pakistan, as well as in other developing countries with similar situations.

Identification of factors influencing severity of motorcycle crashes in Dhaka, Bangladesh using binary logistic regression model

Dhaka, the capital and megacity of the developing country Bangladesh, has experienced a sharp rise in motorcycle users in the last decade, especially after the introduction of ridesharing services. Therefore, the morbidity and mortality rates of motorcycle crash injuries have also increased and become one of the major safety concerns. However, there is scant empirical evidence on motorcycle crash severity in the context of developing countries. Hence, this study was conducted to identify the factors that influenced the severity of motorcycle crashes in Dhaka. A binary logistic regression model was developed using motorcycle crash data of Dhaka over the period of 2006–2015 to identify the contributing factors of motorcycle crash severity. The model output showed that eleven factors significantly increased the probability of fatal motorcycle crashes. These factors were crashes occurring on weekends, during the rainy season, during dawn and night period, at non-intersections, on straight and flat roads, on highways, hit pedestrian type crashes, crashes involving motorcycles with no defect, crashes with heavier vehicles, crashes involving motorcyclists not wearing helmets, and drivers with alcohol suspicion. These findings would help to formulate prevention strategies to reduce the injury severity of motorcycle crashes in the developing countries.

Using machine leaning techniques for evaluation of motorcycle injury severity

There is a growing interest in the application of the machine learning techniques in predicting the motorcycle crash severity. This is partly due to a progress in autonomous vehicles technology, and machine learning technique, which as a main component of autonomous vehicle could be implemented for traffic safety enhancement. Wyoming's motorcycle crash fatalities constitute a concern since the count of riders being killed in motorcycle crashes in 2014 was 11% of the total road fatalities in the state. The first step of crash reduction could be achieved through identification of contributory factors to crashes. This could be accomplished by using a right model with high accuracy in predicting crashes. Thus, this study adopted random forest, support vector machine, multivariate adaptive regression splines and binary logistic regression techniques to predict the injury severity outcomes of motorcycle crashes. Even though researchers applied all the aforementioned techniques to model motorcycle injury severities, a comparative analysis to assess the predictive power of such modeling frameworks is limited. Hence, this study contributes to the road safety literature by comparing the performance of the discussed techniques. In this study, Wyoming's motorcycle crash injury severities are modeled as functions of the characteristics that give rise to crashes. Before conducting any analyses, feature reduction was used to identify a best number of predictors to be included in the model. Also to have an unbiased estimation of the performance of different machine learning techniques, 5-fold cross-validation was used for model performance evaluation. Two measure, Area under the curve (AUC), and confusion matrix were used to compare different models' performance. The machine learning results indicate that random forest model outperformed the other models with the least misclassification and higher AUC. It was also revealed that a dichotomous response variable, with fatality and incapacitation injury in one category, along with all other categories in another group would result in a lower misclassification rate than a polychotomous response variable. This might result from the nature of motorcycle crashes, lacking a protection compared with passenger cars, preventing machine learning technique to get trained properly. Moreover, the most important variables identified by the random forest model are those related to the operating speed, resentful other party, traffic volume, truck traffic volume, riding under the influence, horizontal curvature, wide roadway with more than two lanes and rider's age.

Modeling severities of motorcycle crashes using random parameters

One of the critical areas of road safety is motorcycle safety. Motorcyclists are more vulnerable to injuries than occupants of other motor vehicles when involved in crashes. Researchers have studied the relationships between motorcycle crash severity and crash contributing factors. They are crash characteristics, roadway geometric design features, traffic characteristics, socio-demographics and environmental conditions. However, few researchers considered unobserved heterogeneity effects when modeling motorcycle crash injury severities, let alone interaction effects. In this research, motorcycle crashes in Wyoming that occurred from 2008 to 2017 were analyzed. Specifically, the injury severities of single motorcycle crashes and multiple vehicle crashes involving motorcycles were modeled. The response was whether the motorcycle crash incurred an incapacitating injury or fatality or not. The binary logistic regression and mixed binary logistic regression modeling structures were implemented. The mixed models revealed effects that otherwise would have been undisclosed in the binary logistic regression models’ results. According to the results of single motorcycle crashes, the majority of motorcycle-animal crashes and of motorcycle-barrier crashes were likely to be severe relative to other single motorcycle crashes. It was also found that horizontal curves increased the risk of severe injuries. Young riders were found to be less at risk of being gravely injured in single motorcycle crashes than older riders as well. Furthermore, riding under the influence and high posted speed limits increased the odds of severe crashes regardless of whether the crashes were single motorcycle crashes or multiple vehicle crashes involving motorcycles. Additionally, the mixed models uncovered interaction effects and unobserved effects pertaining to speed limits. • Severities of motorcycle crashes in Wyoming were modeled using logistic regression. • Random parameters were incorporated in the modeling. • Latent effects were inferred regarding crash types and roadway factors. • Recommendations were made to reduce motorcycle crash severity.

Application of deep learning techniques in predicting motorcycle crash severity

AbstractMachine learning (ML) techniques play a crucial role in today's modern world. Over the last years, road traffic safety is one of the applications where ML‐methods have been successfully employed to prevent road users from being killed or seriously injured. A reliable data‐driven predictive model is essential for this purpose. This could be achieved by successfully applying an intelligent transportation system to identify a driver at a higher risk of crashes. This study investigates the capabilities of different deep learning techniques to predict motorcycle crash severity. This study is based on 2,430 motorcycle crashes in a mountainous area in the United States over a 10‐year period. Different deep networks (DNNs), including deep belief network, standard recurrent neural network (RNN), multilayer neural network, and single‐layer neural network, were considered and compared in terms of prediction accuracy of motorcycle crash severity. Before conducting any analysis, feature reduction was performed to identify the optimal number of variables to include in the models by minimizing the error rate. Different metrics including the area under the curve and confusion matrix were used to compare the different models. Although the analyses were conducted on a relatively small dataset, the results indicate that almost all the DNN models better perform in predicting the severity of motorcycle crashes, compared with the single layer neural network. Finally, the RNN outperforms the other three neural network models. A comprehensive discussion has been made about the methodological approach implemented in this study.