Fatal Injury Crashes Research Articles

Traffic Congestion and Safety: Mixed Effects on Total and Fatal Crashes

This paper examines the effects of traffic congestion on total crashes, fatal or serious injury (FSI) crashes, and fatal-only crashes in peak periods using a zone-level safety analysis in Greater Melbourne. Bayesian mixed-effect negative binomial models are employed to investigate the relationship between a congestion index and the frequency of total and FSI crashes. In addition, Bayesian mixed-effect binary logistic models are adopted to explore the association between the congestion index and the likelihood of having fatal crashes in Statistical Area Level 2 (SA2) zones. Modelling results indicate that traffic congestion tends to increase total crashes in both the AM and PM peak periods and FSI crashes in the AM peak period. In contrast, traffic congestion tends to decrease the likelihood of having fatal crashes at both the AM and PM peaks. These findings suggest that many policies to reduce traffic congestion may also enhance road safety by lowering the overall number of crashes. However, it is crucial to incorporate careful speed management within these policies to reduce the risk of fatal crashes effectively.

Identification of Risk Factors for Severe Younger and Older Driver Crashes in Massachusetts

Systemic safety analysis is a comprehensive and proactive approach to identifying and addressing high-risk roadway features and crash types that contribute to severe crashes. The Massachusetts Department of Transportation identified 14 emphasis areas for systemic improvement, including younger and older driver crashes. Research was undertaken to conduct separate comprehensive analyses of fatal and serious injury crashes involving younger and older drivers, using a systemic safety approach to develop risk factors and identify critical locations within the state. The study utilized crash, roadway, driver’s license, citation, healthy aging, and environmental justice data from various sources. Negative binomial regression models were developed to relate crash frequency with town-specific traffic, behavior, and site characteristics, considering exposure and spatial variance. The findings highlighted the significance of multiple risk factors contributing to severe crash frequency. For younger drivers, the presence of young drivers in the town, risk-taking behaviors such as unbelted and impaired driving, and infrastructure characteristics, particularly high-speed facilities and historically underinvested infrastructure, were key contributors to crash frequency. Older drivers’ crash frequency was influenced by the presence of older drivers, healthy aging indicators, risk-taking behaviors, and high-speed facilities. Based on the model results, risk scores were developed for each variable in the model and applied to each town. Towns were ranked using percentile ranking to classify them into primary and secondary risk categories. The towns identified in the primary risk category (18 towns in the top 5 percentiles) captured approximately 13.5% of severe younger driver crashes and 14.5% of severe older driver crashes.

Identifying Corridor-Level Safety Improvements for Urban and Suburban Arterials in Florida Within a Safe System Framework

Many agencies have adopted a Safe System approach to improving roadway safety. The Highway Safety Manual (HSM) provides methods for assessing safety, but these models are site-specific and require extensive data, making them difficult to use at large scales. This paper develops a corridor-level methodology for holistically looking at corridors made of consecutive intersections and roadway segments to identify safety improvements which align with the Safe System approach while requiring less data than HSM methods. Using a standardized definition, a total of 549 corridors on urban and suburban arterials across Florida were identified which experienced over 10,000 fatal and serious injury (FSI) crashes from 2017 through 2021. A negative binomial regression model was developed to predict mean FSI (MFSI) crashes at the corridor level (using corridor length as exposure), with the predicted values adjusted using the Empirical Bayes method to provide more accurate results. The significant factors in the model were traffic volume, intersection densities and sizing, area type, bus stop presence, citation rate, and corridor lighting presence. Increasing citation rates (citations/year/mile) for unsafe driving behaviors by one unit was predicted to reduce MFSI crash frequency in corridors by 2%, and corridors without lighting were predicted to experience 2.79 times more MFSI crashes compared with corridors with lighting. Two sister corridors in South Florida with similar roadway characteristics but different crash frequencies were also analyzed. Improvements to lighting and access control in the identified high-risk corridor could help reduce FSI crashes. Overall, this corridor approach can help agencies proactively improve roadway safety.

Comparison of Cluster-Based Sampling Approaches for Imbalanced Data of Crashes Involving Large Trucks

Severe and fatal crashes involving large trucks result in significant social and economic losses for human society. Unfortunately, the notably low proportion of severe and fatal injury crashes involving large trucks creates an imbalance in crash data. Models trained on imbalanced crash data are likely to produce erroneous results. Therefore, there is a need to explore novel sampling approaches for imbalanced crash data, and it is crucial to determine the appropriate combination of a machine learning model, sampling approach, and ratio. This study introduces a novel cluster-based under-sampling technique, utilizing the k-prototypes clustering algorithm. After initial cluster-based under-sampling, the consolidated cluster-based under-sampled data set was further resampled using three different sampling approaches (i.e., adaptive synthetic sampling (ADASYN), NearMiss-2, and the synthetic minority oversampling technique + Tomek links (SMOTETomek)). Later, four machine learning models (logistic regression (LR), random forest (RF), gradient-boosted decision trees (GBDT), and the multi-layer perceptron (MLP) neural network) were trained and evaluated using the geometric mean (G-Mean) and area under the receiver operating characteristic curve (AUC) scores. The findings suggest that cluster-based under-sampling coupled with the investigated sampling approaches improve the performance of the machine learning models developed on crash data significantly. In addition, the GBDT model combined with ADASYN or SMOTETomek is likely to yield better predictions than any model combined with NearMiss-2. Regarding changes in sampling ratios, increasing the sampling ratio with ADASYN and SMOTETomek is likely to improve the performance of models up to a certain level, whereas with NearMiss-2, performance is likely to drop significantly beyond a specific point. These findings provide valuable insights for selecting optimal strategies for treating the class imbalance issue in crash data.

Impact of real-time weather conditions on crash injury severity in Kentucky using the correlated random parameters logit model with heterogeneity in means

The present study investigated the impact of real-time weather (air temperature, relative humidity, precipitation, wind speed, and solar radiation) on crash injury severity. Recent crash data (January 2016 to April 2021) on Interstate-75 in the state of Kentucky were merged with real-time weather information (retrieved from Kentucky Mesonet stations) at the 1-hour level. The severity index “SI” (i.e., the ratio of percent severe crashes to percent exposure of a specific weather state during the crash period) was introduced to evaluate the impact of different real-time weather states on fatal and severe injury crashes. Furthermore, the standard mixed logit (MXL), correlated mixed logit (CMXL), and correlated mixed logit with heterogeneity in means (CMXLHM) models were fitted and compared to identify the risk factors contributing to crash injury severity while accounting for unobserved heterogeneity. The results showed that the CMXLHM model was statistically superior to the CMXL and MXL models based on various goodness-of-fit measures (e.g., Akaike information criterion “AIC” and McFadden pseudo R-squared). Results from the SI analysis and CMXLHM model showed that real-time weather-related factors (e.g., air temperature ≥ 70 0F and relative humidity ≥ 90 %) were significantly associated with higher severe injury likelihood. Further, driving under the influence (DUI), young drivers, and vehicle travel speed were associated with greater injury severities. On the other hand, presence of horizontal curve, passenger cars, and hourly traffic volume were associated with lower injury severity likelihood. The study outcomes can help in incident management by suggesting specific real-time weather-related states to feed to dynamic message signs (DMS) to enhance travelers’ safety along the interstates.

Analysis of factors that influence injury severity of single and multivehicle crashes involving at-fault older drivers: A random parameters logit with heterogeneity in means and variances approach

The number of older drivers on our roads have been increasing given the upward shift of older people in the population. It is well-known that older drivers are more susceptible to injury given a crash. They are also likely to be at-fault in certain types of crashes due to health conditions associated with old age. This paper identifies factors that influence the severity of crashes involving at-fault older drivers at a disaggregated level for single and multivehicle crashes. Two random parameters multinomial logit with heterogeneity in means and variances models of injury severity were developed to assess the impact of various crash factors on crash outcomes based on the crash type. The study used five years of crash data from Alabama. The model estimations show that rainy weather and dark roadways were associated with an increased likelihood of serious injuries for single-vehicle crashes. Older male drivers were also found to have a higher probability of being at-fault in fatal single-vehicle crashes. For multivehicle older driver crashes, the model estimation results revealed that fatal injury was more likely to occur when the airbags were deployed. In addition, drivers between 65 and 75 years are less likely to be involved in fatal injury crashes for both single and multivehicle crashes. It was further found that older driver at-fault crashes that occurred in the open country were more likely to be fatal regardless of the manner of collision. The findings of this study are expected to help policymakers and transportation stakeholders to implement countermeasures that will make the road safer for older drivers and all other road users.

Single-vehicle roadway departure crashes at rural two-lane highway curved segments: A diagnosis using pattern recognition

Curved segments account for a disproportionately high proportion of fatal and serious injury crashes, with most of these crashes occurring on rural two-lane (R2L) highways. During the 10-year period from 2008 to 2017, a total of 1 234 fatal single-vehicle roadway departure (SV-RwD) crashes occurred on R2L roads in Louisiana, out of which 635 (51.5 %) crashes occurred on curved segments. Therefore, it is critical to investigate the causes of SV-RwD crashes, specifically those that occur on curved segments. This study aimed to investigate the ‘association knowledge’ of the factors contributing to SV-RwD crashes on R2L curved segments in Louisiana using fatal and injury crash data collected from 2008 to 2017. The study utilized Cluster Correspondence Analysis (CCA), a robust joint dimension reduction and clustering method for handling high-dimensionality and multicollinearity of crash data, to achieve this objective. Based on the cluster validation measures, the study identified five clusters with specific traits, including alcohol-impaired male drivers with no seatbelt usage, young (15–24 years old) female drivers’ crash involvement in cloudy weather conditions, animal-involved crashes in rainy weather conditions, crashes occurring on hillcrest locations under cloudy weather conditions, and crashes in the dark with the presence of streetlights and higher traffic volume. Furthermore, young (15–24 years) female drivers were identified in most clusters, implying that this specific age group of female drivers requires special consideration when dealing with SV-RwD collisions on R2L curved segments. To improve safety on R2L curved segments, policymakers can use the findings of this study to develop targeted countermeasures.

Corridor-level network screening and modeling of fatal and serious injury crashes on urban and suburban arterial corridors in Florida

Network screening is a key aspect of the safety management process. The current standard methods in the Highway Safety Manual assume that intersections and roadway segments are independent, even though interactions between sites likely exist. One way to account for these interactions is by studying roadway corridors made of consecutive intersections and segments. This paper developed a roadway corridor definition to identify urban and suburban arterial corridors in Florida with context classification and lane count as defining features. Applying this definition to 10 Florida counties resulted in 1105 corridors with 14,097 fatal or serious injury crashes from 2017 to 2021. Historic crashes and a developed regression model were used to estimate the expected crash frequency for fatal and serious injury crashes with the empirical Bayes method. After performing k-means clustering on expected crash frequency per mile, 25 high-risk corridors were identified. The difference between the expected and predicted crash frequencies (potential for safety improvement) for each high-risk corridor provided a list of priority corridors for safety improvements. Countermeasures focused on speed reduction in suburban corridors would likely lead to the greatest reductions in fatal and serious injury crashes. This corridor approach allows agencies to quickly identify high-risk areas for priority treatment.

Effects of emotional intelligence on dangerous driving: a comparison between commercial and non-commercial drivers

ABSTRACT Drivers with a Commercial Driver’s License (CDL) experience higher rates of fatal and serious injury crashes compared to non-commercial drivers. Emotional Intelligence (EI) has shown to be a significant predictor of dangerous driving among non-commercial drivers. However, there is limited research investigating EI within commercial drivers. This paper investigates the relationship between drivers’ EI and dangerous driving behaviors, and how this differs between commercial (N = 595) and non-commercial (N = 615) drivers. A survey was conducted, where EI was measured using Trait Emotional Intelligence Questionnaire, and risky driving using Dula Dangerous Driving Questionnaire. Analyses confirmed (1) commercial drivers have lower EI than non-commercial (ANCOVA, p < .01); (2) commercial drivers are more likely to report more dangerous driving behaviors (MANCOVA, p < .01); and (3) lower EI is associated with increased dangerous driving, and these effects are strongest amongst commercial drivers (linear regression, p < .001). Promoting and improving EI can help prevent dangerous driving and reduce crash risk.

Short-Duration Crash Modeling to Understand the Impact of Operating Speed on Freeway Crashes During COVID-19

Gaining an understanding of speed–crash relationships is a critical issue in highway safety research. Because of the ongoing pandemic (COVID-19) there has been a reduction in traffic volume, and some early studies explain that speeding in an environment with less traffic is associated with a high number of crashes, especially fatal and serious injury crashes. This study aims to quantify the impact of operating speed on traffic crash occurrences. The study conflated several databases (speed data, roadway inventory data, and crash data) that contain data from Dallas, Texas, spanning from 2018 to 2020, to examine the speed–crash association. Using the negative binomial Lindley regression model, this study showed that the trends of crash prediction models vary over the years (2018, 2019, and 2020) by different injury severity levels (i.e., fatal crashes, fatal and incapacitating injury crashes). The 2020 models show that operating speed measures (i.e., average operating speed) have a significant impact on crash frequencies. The magnitudes of the speed measures show variations across the models at different injury severity levels.

Leveraging probe data to model speeding on urban limited access highway segments: Examining the impact of operational performance, roadway characteristics, and COVID-19 pandemic

Stay-at-home orders - imposed to prevent the spread of COVID-19 - drastically changed the way highways operate. Despite lower traffic volumes during these times, the rate of fatal and serious injury crashes increased significantly across the United States due to increased speeding on roads with less traffic congestion and lower levels of speed enforcement. This paper uses a mixed effect binomial regression model to investigate the impact of stay-at-home orders on odds of speeding on urban limited access highway segments in Maine and Connecticut. This paper also establishes a link between traffic density and the odds of speeding. For this purpose, hourly speed and volume probe data were collected on limited access highway segments for the U.S. states of Maine and Connecticut to estimate the traffic density. The traffic density was then combined with the roadway geometric characteristics, speed limit, as well as dummy variables denoting the time of the week, time of the day, COVID-19 phases (before, during and after stay-at-home order), and the interactions between them. Density, represented in the model as Level of Service, was found to be associated with the odds of speeding, with better levels of service such as A, or B (low density) resulting in the higher odds that drivers would speed. We also found that narrower shoulder width could result in lower odds of speeding. Furthermore, we found that during the stay-at-home order, the odds of speeding by more than 10, 15, and 20 mph increased respectively by 54%, 71% and 85% in Connecticut, and by 15%, 36%, and 65% in Maine during evening peak hours. Additionally, one year after the onset of the pandemic, during evening peak hours, the odds of speeding greater than 10, 15, and 20 mph were still 35%, 29%, and 19% greater in Connecticut and 35% 35% and 20% greater in Maine compared to before pandemic.

Exploring association of contributing factors to pedestrian fatal and severe injury crashes under dark-no-streetlight condition

The pedestrian crash pattern in the dark-no-streetlight condition is a noteworthy ongoing traffic safety concern. The persistently high percentage of pedestrian fatalities at night in the U.S., as well as in the state of Louisiana, necessitates new research to supplement existing studies. This study utilized 10 years (2010–2019) of pedestrian fatal and severe injury crashes in Louisiana that occurred in the dark without streetlights to identify the associated crash patterns. The methodology is based on Multiple Correspondence Analysis (MCA), an exploratory approach used to discover the association of multiple categorical variables from a crash dataset. The findings suggest that driver characteristics (age, gender, and physical condition), pedestrian action, pedestrian alcohol impairment, and physical settings (posted speed limit, location, and roadway type) have a substantial impact on pedestrian collisions at night without streetlights. Moreover, the obtained combination clouds of MCA reveal associations such as elderly pedestrian (>64 years) alcohol impairment resulting in fatalities, crashes in an open country location with a high posted speed limit, crashes involving pedestrians in dark clothing on high-speed (50–55 mph) roadways, alcohol-impaired driver involvement in crashes on two-way roads without physical separation, severe injury crashes at intersections, male pedestrian crashes on midblock locations during weekends, and young (15–24 years) female driver's involvement in crashes while pedestrians were walking against the traffic. Based on the findings, this research also suggests safety recommendations that can assist highway safety practitioners in determining appropriate countermeasures to reduce pedestrian crashes in the dark-no-streetlight condition.

Пропозиції щодо удосконалення правового регулювання використання тахографів на автомобільному транспорті

The probability of death or serious injury in an accident with a road vehicle involved is almost 50% higher compared to other accidents due to the large size and mass of the vehicles involved, as well as due to the fact the driver who fell asleep cannot undertake braking or motion to a side , to avoid a collision or abate an impact. Although it is impossible to calculate the exact number of fatigue-related crashes, studies show that driver fatigue may be a contributing factor in about 20% of all traffic crashes and approximately 25% of all fatal and serious injury crashes. In order to reduce the number of road accidents by means preventing the driving by tired drivers of com-mercial vehicles with a gross weight of more than 3.5 ton, as well as those that carry more than 9 people in the pro-cess of road transportation, tachograph systems have been implemented in many countries. The major documents implementing the tachograph systems (depending on the country) are Regulation (EC) No. 561/2006 of the European Parliament and of the Council or the European agreement concerning the work of crews of vehicles engaged in international road transport (AETR). In order to contribute to the improvement of the level of safety during the internal road transportation of passengers and cargo, Ukraine, in accordance with the Association Agreement between the European Union and the European Atomic Energy Community and their Member States, of the one part, and Ukraine, of the other part, has undertaken to implement the requirements of the Council Regulation (EEC) No. 3821/85 and Regulation (EC) No. 561/2006 of the European Parliament and of the Council in the legislative and regulatory acts of Ukraine. When implementing the requirements of the aforementioned Regulations into the legislative and regulatory acts of Ukraine, it is necessary to take into account the recent changes to the specified documents, as well as the introduc-tion into the EU legislation of associate documents that are directly related to the aforementioned Regulations. Keywords: road safety, tachograph system.

Factors Associated with Work Zone Crashes

Nationally, the annual number of fatal crashes in work zones over the past decade (2010–2019) increased from 521 to 762, which led to rising fatalities from 546 in 2010 to 842 in 2019. Traffic safety in work zones with the aim of reducing fatal and serious injury crashes has been under investigation by researchers and related agencies for decades. The main objective of this study was to investigate the factors associated with work zone crashes to recommend policies targeted at addressing them. Data collected by the Kansas Department of Transportation on crashes that occurred between 2016 and 2020 were used for this research. For the statistical model, binary logistic regression was used in which the dichotomous response variable was defined as either crashes in work zones or crashes outside work zones. According to the model’s results, several significant factors were associated with work zone crashes. Those that had a reasonable expectation of improvement through state policies included the following: alcohol involvement; driver distraction; large commercial vehicle involvement; pedestrian involvement; single-vehicle crashes; speed limits, and speeding. Although these may already be areas targeted in a state’s strategic highway safety plan, special consideration of the overrepresentation of these factors in work zones can provide additional policy options for consideration. Existing countermeasures can be explored, or new countermeasures can be developed based on the needs of the existing conditions.

Understanding the drowsy driving crash patterns from correspondence regression analysis

Abstract: Drowsy driving-related crashes have been a key concern in transportation safety. In Louisiana, 14% (1,758 out of 12,512) of police-reported drowsy driving-related crashes during 2015–2019 resulted in injury (fatal, severe, or moderate). Amid the calls for action against drowsy driving by national agencies, it is of paramount importance to explore the key reportable attributes of drowsy driving behaviors and their potential association with crash severity. Method: This study used 5-years (2015–2019) of crash data and utilized the correspondence regression analysis method to identify the key collective associations of attributes in drowsy driving-related crashes and interpretable patterns based on injury levels. Results: Several drowsy driving-related crash patterns were identified through crash clusters – afternoon fatigue crashes by middle-aged female drivers on urban multilane curves, crossover crashes by young drivers on low-speed roadways, crashes by male drivers during dark rainy conditions, pickup truck crashes in manufacturing/industrial areas, late-night crashes in business and residential districts, and heavy truck crashes on elevated curves. Several attributes – scattered residential areas indicating rural areas, multiple passengers, and older drivers (aged more than 65 years) – showed a strong association with fatal and severe injury crashes. Practical Applications: The findings of this study are expected to help researchers, planners, and policymakers in understanding and developing strategic mitigation measures to prevent drowsy driving.

Understanding the role of faster emergency medical service response in the survival time of pedestrians

Fatalities and severe injuries among vulnerable road users, particularly pedestrians, are rising. In addition to the loss of life, about 6,000 annual pedestrian deaths in the U.S. cost society about $6 billion. Contrary to the assumption that all fatal pedestrian-involved crashes are similar, instantaneous death is substantially more severe than death that occurs several days after the crash. Instead of homogenizing all fatal pedestrian crashes, this study takes into account the severity of fatal injury crashes as a timeline based on the survival time of pedestrians. This study extracts valuable information from fatal crashes by examining pedestrians' survival time ranging from early death to death within 30 days of the crash. The Fatality Analysis Reporting System dataset is utilized from 2015 to 2018. The emergency medical service (EMS) response time is the key post-crash measure, while controlling for pedestrian, driver, roadway, and environmental characteristics. Notably, the response time and survival time can cause endogeneity, i.e., the response times may be shorter for more severe crashes. Due to the spatial and temporal nature of traffic crashes, to extract the association of different variables with pedestrians' survival time, a geographically and temporally weighted truncated regression with a two-stage residual inclusion treatment (local model) is estimated. The local model can overcome the endogeneity limitation (between EMS response time and survival time) and uncover the potentially spatially and temporally varying correlates of pedestrians’ survival time with associated factors to account for unobserved heterogeneity. Moreover, to verify the variations are noticeable, a truncated regression with the two-stage residual inclusion treatment is developed (global model). The modeling results indicate that while capturing the unobserved heterogeneity, the local model outperformed the global model. The empirical results show that EMS response time, speeding, and some pedestrian behaviors are the most important factors that affect pedestrians’ survival time in fatal injury crashes. However, the effect of factors on pedestrians’ survival time is noticeably varied spatially and temporally. The results and their implications are discussed in detail in the paper.

Evaluation of an Increase in Roadside Drug Testing in Victoria Based on Models of the Crash Effects of Random and Targeted Roadside Tests

Drug driving continues to be overrepresented in both fatal and serious injury crashes in Victoria. As an enforcement countermeasure, preliminary oral fluid tests to detect drug driving were introduced in Victoria, Australia in December 2004. Recent research has modelled the relationships between prevalences of THC and methamphetamine in fatally and seriously injured drivers and (a) the annual numbers of random and targeted drug tests during 2010-2016 and (b) the positive detection rates from these tests. The increase in roadside drug tests in Victoria from 42,000 in 2013 (1% of licensed drivers) to 100,000 per year (2.2% of drivers), particularly targeted tests, is estimated to have saved 33 fatal crashes (13.7% reduction) and at least 80 serious injury crashes (1.4% reduction) per year. Based on the findings from this research, further increases in targeted and random roadside drug tests are warranted, up to at least 390,100 total tests per year, which are estimated to save a further 46 fatal crashes and at least a further 134 serious injury crashes per year.

Quantifying the Lost Safety Benefits of ADAS Technologies Due to Inadequate Supporting Road Infrastructure

Advanced driver assistance systems (ADAS) provide warnings to drivers and, if applicable, intervene to mitigate a collision if one is imminent. Autonomous emergency brakes (AEB) and lane keep assistance (LKA) systems are mandated in several new vehicles, given their predicted injury and fatality reduction benefits. These predicted benefits are based on the assumption that roads are always entirely supportive of ADAS technologies. Little research, however, has been conducted regarding the preparedness of the road network to support these technologies in Australia, given its vastly expansive terrain and varying road quality. The objective of this study was to estimate what proportion of crashes that are sensitive to AEB and LKA, would not be mitigated due to unsupportive road infrastructure, and therefore, the lost benefits of the technologies due to inadequate road infrastructure. To do this, previously identified technology effectiveness estimates and a published methodology for identifying ADAS-supportive infrastructure availability was applied to an estimated AEB and LKA-sensitive crash subset (using crash data from Victoria, South Australia and Queensland, 2013–2018 inclusive). Findings demonstrate that while the road networks across the three states appeared largely supportive of AEB technology, the lack of delineation across arterial and sub-arterial (or equivalent) roads is likely to have serious implications on road safety, given 13–23% of all fatal and serious injury (FSI) crashes that occurred on these road classes were LKA-sensitive. Based on historical crash data, over 37 fatalities and 357 serious injuries may not be avoided annually across the three Australian states based on the lack of satisfactory road delineation on arterial and sub-arterial (or equivalent) roads alone. Further, almost 24% of fatalities in Victoria, 24% of fatalities in Queensland and 21% of fatalities in South Australia (that are AEB- or LKA-sensitive) are unlikely to be prevented, given existing road infrastructure. These figures are conservative estimates of the lost benefits of the technologies as they only consider fatal and serious injury crashes and do not include minor injury or property damage crashes, the benefits of pedestrian-sensitive AEB crashes in high-speed zones or AEB fitted to heavy vehicles. It is timely for road investments to be considered, prioritised and allocated, given the anticipated penetration of the new technologies into the fleet, to ensure that the road infrastructure is capable of supporting the upcoming fleet safety improvements.

Profiling fatal and serious injury crashes in South Australia according to the Swedish transport administration model

Objectives This article describes a study examining fatal road crashes in South Australia (SA) according to the safety model developed by the Swedish Road Administration (STA). The STA model is based on the biomechanical limits that human beings can tolerate and specifies a number of elements that must be present for “compliance” with the model: driving at or below the speed limit, driver not intoxicated by alcohol or drugs, vehicle occupants wearing a seatbelt, car with a EuroNCAP rating of 5 stars, car equipped with electronic stability control, and a road with an EuroRAP rating of 4 stars. Method The study used a sample of 105 fatal crashes in SA in 2010 and 2011 and a comparison sample of 136 serious injury crashes from 2014 to 2017. Results It was found that there were high levels of noncompliance with all elements of the STA model, with multiple forms of noncompliance in a clear majority of fatal cases. Conclusions Considerable gains in safety could occur with greater penetration into the SA fleet of 5-star cars fitted with ESC and other vehicle safety technology, especially as the majority of both sets of crashes were of a type involving loss of vehicular control. It is also important to note that risky road user behavior was less common in the serious injury crashes and so a road safety strategy focusing on changing road user behavior will be ineffective for addressing the major factors contributing to non-fatal road trauma.

A deep generative approach for crash frequency model with heterogeneous imbalanced data

Crash frequency model is often subject to excessive zero observation because of the rare nature of crashes. To address the problem of imbalanced crash data, a deep generative approach – augmented variational autoencoder – was proposed to generate synthetic crash data for the association measure between crash and possible explanatory factors. This approach was characterized by a factorized generative model and refined objective function. For instance, the generative model can handle heterogeneous data including real-valued, nominal and ordinal distributions. On the other hand, the refined objective function can control for the random effect by better recognizing both the zero-crash and non-zero crash cases. In this study, comprehensive traffic and crash data of multiple distribution types in Hong Kong in the period between 2014 and 2016 were used. To assess the data generation performance of the proposed augmented variational autoencoder method, a conventional data synthesis technique (synthetic minority oversampling technique-nominal continuous) was also considered. Performances of crash frequency models of total crashes and fatal and severe injury crashes are assessed. For total crashes, the results of parameter estimation, in terms of statistical fit, prediction accuracy, and explanatory factors identified, of the crash frequency model based on synthetic data using the augmented variational autoencoder method adhered closer to that based on original data, compared to that based on synthetic data using the synthetic minority oversampling technique-nominal continuous method. For fatal and severe injury crashes, zero-crash observations were prevalent, with the ratio of zero-crash to non-zero crash cases of 9 to 1. Crash data was first balanced using the proposed augmented variational autoencoder method. Then, fatal and severe injury crash frequency models using correlated random parameter models based on original data and balanced data were estimated respectively. Results indicate that fatal and severe injury crash frequency model based on balanced data outperforms its counterpart, with the lowest root mean square error, lowest mean absolute error, and highest number of crash explanatory factors identified. More importantly, correlation between the random parameters can be revealed. Findings of this study should shed light to both researchers and practitioners for the development of crash frequency models, with which the problem of excessive zero observations is prevalent when highly disaggregated traffic and crash data by time and space are used.