高速公路交通事故严重程度与交通流特征的关系研究

为研究交通事故发生前后交通流特征对事故严重程度的影响,以JT高速公路作为研究对象,长期观测和采集交通流及事故数据。将交通事故发生时段的交通流主要衡量指标与事故信息进行数据匹配,形成交通事故与事故小时交通流匹配数据集,并分析流量、速度、大车比例等交通流表征指标与不同等级事故数的分布规律。通过分析发现:在某些流量、速度或大车比例区段,交通事故数及其严重程度处于较高的水平。在此基础上,利用主成分分析(PCA)技术对衡量交通流特征的初始指标进行降维处理,用交通流主成分指标综合反映交通流特征,并建立事故严重程度与交通流主成分指标的统计分析模型。结果表明:交通流主成分指标趋于零的区段的事故严重程度明显高于其他区段。     

Crash data analysis: collective vs. individual crash level approach

INTRODUCTION: Traffic safety literature has traditionally focused on identification of location profiles where "more crashes are likely to occur" over a period of time. The analysis involves estimation of crash frequency and/or rate (i.e., frequency normalized based on some measure of exposure) with geometric design features (e.g., number of lanes) and traffic characteristics (e.g., Average Annual Daily Traffic [AADT]) of the roadway location. In the recent past, a new category of traffic safety studies has emerged, which attempts to identify locations where a "crash is more likely to occur." The distinction between the two groups of studies is that the latter group of locations would change based on the varying traffic patterns over the course of the day or even within the hour. METHOD: Hence, instead of estimation of crash frequency over a period of time, the objective becomes real-time estimation of crash likelihood. The estimation of real-time crash likelihood has a traffic management component as well. It is a proactive extension to the traditional approach of incident detection, which involves analysis of traffic data recorded immediately after the incident. The units of analysis used in these studies are individual crashes rather than counts of crashes. RESULTS: In this paper, crash data analysis based on the two approaches, collective and at individual crash level, is discussed along with the advantages and shortcomings of the two approaches.     

Investigation of hit-and-run crash occurrence and severity using real-time loop detector data and hierarchical Bayesian binary logit model with random effects

Objective: Most of the extensive research dedicated to identifying the influential factors of hit-and-run (HR) crashes has utilized typical maximum likelihood estimation binary logit models, and none have employed real-time traffic data. To fill this gap, this study focused on investigating factors contributing to HR crashes, as well as the severity levels of HR.

Methods: This study analyzed 4-year crash and real-time loop detector data by employing hierarchical Bayesian models with random effects within a sequential logit structure. In addition to evaluation of the impact of random effects on model fitness and complexity, the prediction capability of the models was examined. Stepwise incremental sensitivity and specificity were calculated and receiver operating characteristic (ROC) curves were utilized to graphically illustrate the predictive performance of the model.

Results: Among the real-time flow variables, the average occupancy and speed from the upstream detector were observed to be positively correlated with HR crash possibility. The average upstream speed and speed difference between upstream and downstream speeds were correlated with the occurrence of severe HR crashes. In addition to real-time factors, other variables found influential for HR and severe HR crashes were length of segment, adverse weather conditions, dark lighting conditions with malfunctioning street lights, driving under the influence of alcohol, width of inner shoulder, and nighttime.

Conclusions: This study suggests the potential traffic conditions of HR and severe HR occurrence, which refer to relatively congested upstream traffic conditions with high upstream speed and significant speed deviations on long segments. The above findings suggest that traffic enforcement should be directed toward mitigating risky driving under the aforementioned traffic conditions. Moreover, enforcement agencies may employ alcohol checkpoints to counter driving under the influence (DUI) at night. With regard to engineering improvements, wider inner shoulders may be constructed to potentially reduce HR cases and street lights should be installed and maintained in working condition to make roads less prone to such crashes.     

Predicting reduced visibility related crashes on freeways using real-time traffic flow data

ObjectivesThe main objective of this paper is to investigate whether real-time traffic flow data, collected from loop detectors and radar sensors on freeways, can be used to predict crashes occurring at reduced visibility conditions. In addition, it examines the difference between significant factors associated with reduced visibility related crashes to those factors correlated with crashes occurring at clear visibility conditions.MethodRandom Forests and matched case-control logistic regression models were estimated.ResultsThe findings indicated that real-time traffic variables can be used to predict visibility related crashes on freeways. The results showed that about 69% of reduced visibility related crashes were correctly identified. The results also indicated that traffic flow variables leading to visibility related crashes are slightly different from those variables leading to clear visibility crashes.Impact on IndustryUsing time slices 5–15 minutes before crashes might provide an opportunity for the appropriate traffic management centers for a proactive intervention to reduce crash risk in real-time.     

Analyses of rear-end crashes based on classification tree models

OBJECTIVE: Signalized intersections are accident-prone areas especially for rear-end crashes due to the fact that the diversity of the braking behaviors of drivers increases during the signal change. The objective of this article is to improve knowledge of the relationship between rear-end crashes occurring at signalized intersections and a series of potential traffic risk factors classified by driver characteristics, environments, and vehicle types. METHODS: Based on the 2001 Florida crash database, the classification tree method and Quasi-induced exposure concept were used to perform the statistical analysis. Two binary classification tree models were developed in this study. One was used for the crash comparison between rear-end and non-rear-end to identify those specific trends of the rear-end crashes. The other was constructed for the comparison between striking vehicles/drivers (at-fault) and struck vehicles/drivers (not-at-fault) to find more complex crash pattern associated with the traffic attributes of driver, vehicle, and environment. RESULTS: The modeling results showed that the rear-end crashes are over-presented in the higher speed limits (45-55 mph); the rear-end crash propensity for daytime is apparently larger than nighttime; and the reduction of braking capacity due to wet and slippery road surface conditions would definitely contribute to rear-end crashes, especially at intersections with higher speed limits. The tree model segmented drivers into four homogeneous age groups: < 21 years, 21-31 years, 32-75 years, and > 75 years. The youngest driver group shows the largest crash propensity; in the 21-31 age group, the male drivers are over-involved in rear-end crashes under adverse weather conditions and the 32-75 years drivers driving large size vehicles have a larger crash propensity compared to those driving passenger vehicles. CONCLUSIONS: Combined with the quasi-induced exposure concept, the classification tree method is a proper statistical tool for traffic-safety analysis to investigate crash propensity. Compared to the logistic regression models, tree models have advantages for handling continuous independent variables and easily explaining the complex interaction effect with more than two independent variables. This research recommended that at signalized intersections with higher speed limits, reducing the speed limit to 40 mph efficiently contribute to a lower accident rate. Drivers involved in alcohol use may increase not only rear-end crash risk but also the driver injury severity. Education and enforcement countermeasures should focus on the driver group younger than 21 years. Further studies are suggested to compare crash risk distributions of the driver age for other main crash types to seek corresponding traffic countermeasures.     

The effects of mobile speed camera introduction on road traffic crashes and casualties in a rural county of England

PROBLEM: This study assesses the impact of crash and casualty numbers in correspondence to the introduction of mobile speed cameras in the rural county of Norfolk, England. METHOD: Road traffic accident casualty and crash data were collected for two years before the introduction of cameras and two years subsequently. The casualties and crashes occurring at 29 camera sites were identified and separated from those occurring in the rest of the county. Trends in crashes and casualties, and their severity, were examined graphically and comparisons were made between before and after periods. The regression to the mean effect at individual sites was estimated. RESULTS: After the introduction of cameras, overall crashes declined by 1% and crashes involving fatalities or serious injuries declined by 9% on the roads without cameras. At the camera sites, crashes decreased by 19% and fatal and serious crashes by 44%. The reduction in total crashes was significantly greater than that expected from the effect of regression to the mean in 12 out of 20 sites tested. SUMMARY: The introduction of cameras appears to have resulted in real and measurable reductions in crash risk in this rural county. IMPACT ON INDUSTRY: Our results suggest the deployment of mobile speed cameras is an effective tool for organizations wishing to reduce road traffic casualties in areas where high crash rates have been associated with excessive vehicle speeds.     

Investigating the different characteristics of weekday and weekend crashes

IntroductionThis study provides a systematic approach to investigate the different characteristics of weekday and weekend crashes.MethodWeekend crashes were defined as crashes occurring between Friday 9 p.m. and Sunday 9 p.m., while the other crashes were labeled as weekday crashes. In order to reveal the various features for weekday and weekend crashes, multi-level traffic safety analyses have been conducted. For the aggregate analysis, crash frequency models have been developed through Bayesian inference technique; correlation effects of weekday and weekend crash frequencies have been accounted. A multivariate Poisson model and correlated random effects Poisson model were estimated; model goodness-of-fits have been compared through DIC values. In addition to the safety performance functions, a disaggregate crash time propensity model was calibrated with Bayesian logistic regression model. Moreover, in order to account for the cross-section unobserved heterogeneity, random effects Bayesian logistic regression model was employed.ResultsIt was concluded that weekday crashes are more probable to happen during congested sections, while the weekend crashes mostly occur under free flow conditions. Finally, for the purpose of confirming the aforementioned conclusions, real-time crash prediction models have been developed. Random effects Bayesian logistic regression models incorporating the microscopic traffic data were developed. Results of the real-time crash prediction models are consistent with the crash time propensity analysis. Furthermore, results from these models would shed some lights on future geometric improvements and traffic management strategies to improve traffic safety.Impact on IndustryUtilizing safety performance to identify potential geometric improvements to reduce crash occurrence and monitoring real-time crash risks to pro-actively improve traffic safety.     

Disaggregated traffic conditions and road crashes in urban signalized intersections

Introduction: Road safety studies in signalized intersections have been performed extensively using annually aggregated traffic variables and crash frequencies. However, this type of aggregation reduces the strength of the results if variables that oscillate over the course of the day are considered (speed, traffic flow, signal cycle length) because average indicators are not able to describe the traffic conditions preceding the crash occurrence. This study aims to explore the relationship between traffic conditions aggregated in 15-min intervals and road crashes in urban signalized intersections. Method: First, an investigation of the reported crash times in the database was conducted to obtain the association between crashes and their precursor conditions. Then, 4.1 M traffic condition intervals were consolidated and grouped using a hierarchical clustering technique. Finally, charts of the frequency of crashes per cluster were explored. Results: The main findings suggest that high vehicular demand conditions are related to an increase in property damage only (PDO) crashes, and an increase in the number of lanes is linked to more PDO and injury crashes. Injury crashes occurred in a wide range of traffic conditions, indicating that a portion of these crashes were due to speeding, while the other fraction was associated with the vulnerability of road users. Traffic conditions with: (a) low vehicular demand and a long cycle length and (b) high vehicular demand and a short cycle length were critical in terms of PDO and injury crashes. Practical Applications: The use of disaggregated data allowed for a stronger evaluation of the relationship between road crashes and variables that oscillate over the course of the day. This approach also permits the development of real-time risk management strategies to mitigate the frequency of critical traffic conditions and reduce the likelihood of crashes.     

The impact of environmental, vehicle and driver characteristics on injury severity in older drivers hospitalized as a result of a traffic crash

INTRODUCTION: Compared to younger age groups, older people are more likely to be seriously injured or to die as a result of a traffic crash. METHOD: The aim of the study is to examine the impact of environmental, vehicle, crash, and driver characteristics on injury severity in older drivers involved in traffic crashes by using recently linked police crash records and hospitalization data from New South Wales, Australia. The severity of injury resulting from traffic crashes was measured using the International Classification of Diseases, 10th revision (ICD-10) Injury Severity Score (ICISS). RESULTS: Multivariate analysis identified rurality, presence of complex intersections, road speed limit, driver error, speeding, and seat belt use as independent predictors of injury severity in older people. The type of intersection configuration explained over half of the observed variations in injury severity. IMPACT ON INDUSTRY: Environmental modification such as intersection treatments might contribute to a decrease in the severity of injury in older people involved in road crashes.     

Crash analysis and development of safety performance functions for Florida roads in the framework of the context classification system

Introduction: Safety performance functions (SPF) are employed to predict crash counts at the different roadway elements. Several SPFs were developed for the various roadway elements based on different classifications such as functional classification and area type. Since a more detailed classification of roadway elements leads to more accurate crash predictions, multiple states have developed new classification systems to classify roads based on a comprehensive classification. In Florida, the new roadway context classification system incorporates geographic, demographic, and road characteristics information. Method: In this study, SPFs were developed in the framework of the FDOT roadway context classification system at three levels of modeling, context classification (CC-SPFs), area type (AT-SPFs), and statewide (SW-SPF) levels. Crash and traffic data from 2015-2019 were obtained. Road characteristics and road environment information have also been gathered along Florida roads for the SPF development. Results: The developed SPFs showed that there are several variables that influence the frequency of crashes, such as annual average daily traffic (AADT), signalized intersections and access point densities, speed limit, and shoulder width. However, there are other variables that did not have an influence in crash occurrence such as concrete surface and the presence of bicycle slots. CC-SPFs had the best performance among others. Moreover, network screening to determine the most problematic road segments has been accomplished. The results of the network screening indicated that the most problematic roads in Florida are the suburban commercial and the urban general roads. Practical Applications: This research provides a solid reference for decision-makers regarding crash prediction and safety improvement along Florida roads.     

Identifying crash type propensity using real-time traffic data on freeways

Introduction: We examine the effects of various traffic parameters on type of road crash. Method: Multivariate probit models are specified on 4-years of data from the A4-A86 highway section in the Ile-de-France region, France. Results: Empirical findings indicate that crash type can almost exclusively be defined by the prevailing traffic conditions shortly before its occurrence. Rear-end crashes involving two vehicles were found to be more probable for relatively low values of both speed and density, rear-end crashes involving more than two vehicles appear to be more probable under congested conditions, while single-vehicle crashes appear to be largely geometry-dependent. Impact on Industry: Results could be integrated in a real-time traffic management application.     

A conceptual framework to understand the role of built environment on traffic safety

IntroductionMany U.S. cities have adopted the Vision Zero strategy with the specific goal of eliminating traffic-related deaths and injuries. To achieve this ambitious goal, safety professionals have increasingly called for the development of a safe systems approach to traffic safety. This approach calls for examining the macrolevel risk factors that may lead road users to engage in errors that result in crashes. This study explores the relationship between built environment variables and crash frequency, paying specific attention to the environmental mediating factors, such as traffic exposure, traffic conflicts, and network-level speed characteristics. Methods: Three years (2011–2013) of crash data from Mecklenburg County, North Carolina, were used to model crash frequency on surface streets as a function of built environment variables at the census block group level. Separate models were developed for total and KAB crashes (i.e., crashes resulting in fatalities (K), incapacitating injuries (A), or non-incapacitating injuries (B)) using the conditional autoregressive modeling approach to account for unobserved heterogeneity and spatial autocorrelation present in data. Results: Built environment variables that are found to have positive associations with both total and KAB crash frequencies include population, vehicle miles traveled, big box stores, intersections, and bus stops. On the other hand, the number of total and KAB crashes tend to be lower in census block groups with a higher proportion of two-lane roads and a higher proportion of roads with posted speed limits of 35 mph or less. Conclusions: This study demonstrates the plausible mechanism of how the built environment influences traffic safety. The variables found to be significant are all policy-relevant variables that can be manipulated to improve traffic safety. Practical Applications: The study findings will shape transportation planning and policy level decisions in designing the built environment for safer travels.     

A classification tree based modeling approach for segment related crashes on multilane highways

Introduction

This study presents a classification tree based alternative to crash frequency analysis for analyzing crashes on mid-block segments of multilane arterials.

Method

The traditional approach of modeling counts of crashes that occur over a period of time works well for intersection crashes where each intersection itself provides a well-defined unit over which to aggregate the crash data. However, in the case of mid-block segments the crash frequency based approach requires segmentation of the arterial corridor into segments of arbitrary lengths. In this study we have used random samples of time, day of week, and location (i.e., milepost) combinations and compared them with the sample of crashes from the same arterial corridor. For crash and non-crash cases, geometric design/roadside and traffic characteristics were derived based on their milepost locations. The variables used in the analysis are non-event specific and therefore more relevant for roadway safety feature improvement programs. First classification tree model is a model comparing all crashes with the non-crash data and then four groups of crashes (rear-end, lane-change related, pedestrian, and single-vehicle/off-road crashes) are separately compared to the non-crash cases. The classification tree models provide a list of significant variables as well as a measure to classify crash from non-crash cases. ADT along with time of day/day of week are significantly related to all crash types with different groups of crashes being more likely to occur at different times.

Conclusions

From the classification performance of different models it was apparent that using non-event specific information may not be suitable for single vehicle/off-road crashes.

Impact on Industry

The study provides the safety analysis community an additional tool to assess safety without having to aggregate the corridor crash data over arbitrary segment lengths.     

Investigation of crashes at pedestrian hybrid beacons: Results of a large-scale study in Arizona

Introduction: The pedestrian hybrid beacon (PHB) is a traffic control device used at pedestrian crossings. A recent Arizona Department of Transportation research effort investigated changes in crashes for different severity levels and crash types (e.g., rear-end crashes) due to the PHB presence, as well as for crashes involving pedestrians and bicycles. Method: Two types of methodologies were used to evaluate the safety of PHBs: (a) an Empirical Bayes (EB) before-after study, and (b) a long-term cross-sectional observational study. For the EB before-after evaluation, the research team considered three reference groups: unsignalized intersections, signalized intersections, and both unsignalized and signalized intersections combined. Results: For the signalized and combined unsignalized and signalized intersection groups, all crash types considered showed statistically significant reductions in crashes (e.g., total crashes, fatal and injury crashes, rear-end crashes, fatal and injury rear-end crashes, angle crashes, fatal and injury angle crashes, pedestrian-related crashes, and fatal and injury pedestrian-related crashes). A cross-sectional study was conducted with a larger number of PHBs (186) to identify relationships between roadway characteristics and crashes at PHBs, especially with respect to the distance to an adjacent traffic control signal. The distance to an adjacent traffic signal was found to be significant only at the α = 0.1 level, and only for rear-end and fatal and injury rear-end crashes. Conclusions: This analysis represents the largest known study to date on the safety impacts of PHBs, along with a focus on how crossing and geometric characteristics affect crash patterns. The study showed the safety benefits of PHBs for both pedestrians and vehicles. Practical Applications: The findings from this study clearly support the installation of PHBs at midblock or intersection crossings, as well as at crossings on higher-speed roads.     

Weather impacts on single-vehicle truck crash injury severity

IntroductionThe focus of this paper is on illustrating the feasibility of aggregating data from disparate sources to investigate the relationship between single-vehicle truck crash injury severity and detailed weather conditions. Specifically, this paper presents: (a) a methodology that combines detailed 15-min weather station data with crash and roadway data, and (b) an empirical investigation of the effects of weather on crash-related injury severities of single-vehicle truck crashes.MethodRandom parameters ordinal and multinomial regression models were used to investigate crash injury severity under different weather conditions, taking into account the individual unobserved heterogeneity. The adopted methodology allowed consideration of environmental, roadway, and climate-related variables in single-vehicle truck crash injury severity.Results and conclusionsResults showed that wind speed, rain, humidity, and air temperature were linked with single-vehicle truck crash injury severity. Greater recorded wind speed added to the severity of injuries in single-vehicle truck crashes in general. Rain and warmer air temperatures were linked to more severe crash injuries in single-vehicle truck crashes while higher levels of humidity were linked to less severe injuries. Random parameters ordered logit and multinomial logit, respectively, revealed some individual heterogeneity in the data and showed that integrating comprehensive weather data with crash data provided useful insights into factors associated with single-vehicle truck crash injury severity.Practical applicationsThe research provided a practical method that combined comprehensive 15-min weather station data with crash and roadway data, thereby providing useful insights into crash injury severity of single-vehicle trucks. Those insights are useful for future truck driver educational programs and for truck safety in different weather conditions.     

Investigating the effects of monthly weather variations on Connecticut freeway crashes from 2011 to 2015

IntroductionThe objective of this research is to investigate the effects of monthly weather conditions on traffic crash experience on freeways, considering the interactions between weather, traffic volumes, and roadway conditions. Methods: Data from the state of Connecticut from 2011to 2015 were used. Random parameters negative binomial models with first-order, autoregressive covariance were estimated for representative types of freeway crashes (front-to-rear, sideswipe-same-direction, and fixed-object), most severe crashes (i.e., fatal and injury crashes), and non-injury crashes (i.e., property-damage-only crashes). Results: Major findings are that variations in monthly traffic volumes, roadway geometry, and weather conditions explain much of the variations in monthly traffic crashes. Time effects exist in the panel monthly data for all types of crashes. Taking into account this effect improves model prediction results. When the raw weather measures are highly correlated, using dimension reduction techniques helps to extract more interpretable weather factors. By considering the interaction effects between roadway condition variables, additional findings were found. In general, lower temperature, more heavy fog days, decreased precipitation, lower wind speed, higher monthly traffic volumes, and narrower inside shoulder were found to be associated with higher monthly crashes. The effects of area type and outside shoulder width change dramatically as the number of through lanes changes. Practical applications: The findings of this research could help researchers and general readers gain a better understanding of the effects of monthly weather conditions and other roadway factors on freeway crashes and give engineers practical guidelines on improving freeway safety.     

A clustering approach to integrate traffic safety in road maintenance prioritization

Objectives: We combine data on roads and crash characteristics to identify patterns in road traffic crashes with regard to road characteristics. We illustrate how combined analysis of data regarding road maintenance, maintenance costs, road characteristics, crash characteristics, and geographical location can enrich road maintenance prioritization from a traffic safety perspective.

Methods: The study is based on traffic crash data merged with road maintenance data and annual average daily traffic (AADT) collected in Denmark. We analyzed 3,964 crashes that occurred from 2010 to 2015. A latent class clustering (LCC) technique was used to identify crash clusters with different road and crash characteristics. The distribution of crash severity and estimated road maintenance costs for each cluster was found and cluster differences were compared using the chi-square test. Finally, a map matching procedure was used to identify the geographical distribution of the crashes in each cluster.

Results: Results showed that based on road maintenance levels there was no difference in the distribution of crash severity. The LCC technique revealed 11 crash clusters. Five clusters were characterized by crashes on roads with a poor maintenance level (levels 4 and 3). Only a few of these crashes included a vulnerable road user (VRU) but many occurred on roads without barriers. Four clusters included a large share of crashes on acceptably maintained roads (level 2). For these clusters only small variations in road characteristics were found, whereas the differences in crash characteristics were more dominant. The last 2 clusters included crashes that mainly occurred on new roads with no need for maintenance (level 1). Injury severity, estimated maintenance costs, and geographical location were found to be differently distributed for most of the clusters.

Conclusions: We find that focusing solely on road maintenance and crash severity does not provide clear guidance of how to prioritize between road maintenance efforts from a traffic safety perspective. However, when combined with geographical location and crash characteristics, a more nuanced picture appears that allows consideration of different target groups and perspectives.     

Pedestrians under influence (PUI) crashes: Patterns from correspondence regression analysis

Introduction: Alcohol-related impairment is a key contributing factor in traffic crashes. However, only a few studies have focused on pedestrian impairment as a crash characteristic. In Louisiana, pedestrian fatalities have been increasing. From 2010 to 2016, the number of pedestrian fatalities increased by 62%. A total of 128 pedestrians were killed in traffic crashes in 2016, and 34.4% of those fatalities involved pedestrians under the influence (PUI) of drugs or alcohol. Furthermore, alcohol-PUI fatalities have increased by 120% from 2010 to 2016. There is a vital need to examine the key contributing attributes that are associated with a high number of PUI crashes. Method: In this study, the research team analyzed Louisiana’s traffic crash data from 2010 to 2016 by applying correspondence regression analysis to identify the key contributing attributes and association patterns based on PUI involved injury levels. Results: The findings identified five risk clusters: intersection crashes at business/industrial locations, mid-block crashes on undivided roadways at residential and business/residential locations, segment related crashes associated with a pedestrian standing in the road, open country crashes with no lighting at night, and pedestrian violation related crashes on divided roadways. The association maps identified several critical attributes that are more associated with fatal and severe PUI crashes. These attributes are dark to no lighting, open country roadways, and non-intersection locations. Practical Applications: The findings of this study may be used to help design effective mitigation strategies to reduce PUI crashes.     

Investigation of injury severities in single-vehicle crashes in North Carolina using mixed logit models

IntroductionRoadway departure (RwD) crashes, comprising run-off-road (ROR) and cross-median/centerline head-on collisions, are one of the most lethal crash types. According to the FHWA, between 2015 and 2017, an average of 52 percent of motor vehicle traffic fatalities occurred each year due to roadway departure crashes. An avoidance maneuver, inattention or fatigue, or traveling too fast with respect to weather or geometric road conditions are among the most common reasons a driver leaves the travel lane. Roadway and roadside geometric design features such as clear zones play a significant role in whether human error results in a crash. Method: In this paper, we used mixed-logit models to investigate the contributing factors on injury severity of single-vehicle ROR crashes. To that end, we obtained five years' (2010–2014) of crash data related to roadway departures (i.e., overturn and fixed-object crashes) from the Federal Highway Administration's Highway Safety Information System Database. Results: The results indicate that factors such as driver conditions (e.g., age), environmental conditions (e.g., weather conditions), roadway geometric design features (e.g., shoulder width), and vehicle conditions significantly contributed to the severity of ROR crashes. Conclusions: Our results provide valuable information for traffic design and management agencies to improve roadside design policies and implementing appropriately forgiving roadsides for errant vehicles. Practical applications: Our results show that increasing shoulder width and keeping fences at the road can reduce ROR crash severity significantly. Also, increasing road friction by innovative materials and raising awareness campaigns for careful driving at daylight can decrease the ROR crash severity.     

Exploring pedestrian injury severities at pedestrian-vehicle crash hotspots with an annual upward trend: A spatiotemporal analysis with latent class random parameter approach

Introduction: With the increasing trend of pedestrian deaths among all traffic fatalities in the past decade, there is an urgent need for identifying and investigating hotspots of pedestrian-vehicle crashes with an upward trend. Method: To identify pedestrian-vehicle crash locations with aggregated spatial pattern and upward temporal pattern (i.e., hotspots with an upward trend), this paper first uses the average nearest neighbor and the spatial autocorrelation tests to determine the grid distance and the neighborhood distance for hotspots, respectively. Then, the spatiotemporal analyses with the Getis-Ord Gi* index and the Mann-Kendall trend test are utilized to identify the pedestrian-vehicle crash hotspots with an annual upward trend in North Carolina from 2007 to 2018. Considering the unobserved heterogeneity of the crash data, a latent class model with random parameters within class is proposed to identify specific contributing factors for each class and explore the heterogeneity within classes. Significant factors of the pedestrian, vehicle, crash type, locality, roadway, environment, time, and traffic control characteristics are detected and analyzed based on the marginal effects. Results: The heterogeneous results between classes and the random parameter variables detected within classes further indicate the superiority of latent class random parameter model. Practical Applications: This paper provides a framework for researchers and engineers to identify crash hotspots considering spatiotemporal patterns and contribution factors to crashes considering unobserved heterogeneity. Also, the result provides specific guidance to developing countermeasures for mitigating pedestrian-injury at pedestrian-vehicle crash hotspots with an upward trend.