2021

Ehsan Rahimi (erahim4@uic.edu), University of Illinois, Chicago
Ali Shamshiripour, University of Illinois, Chicago
Amir Samimi, Sharif University of Technology
Abolfazl Mohammadian, University of Illinois, Chicago

Trucking plays a vital role in economic development in every country, especially countries where it serves as the backbone of the economy. The fast growth of economy in Iran as a developing country has also been accompanied by an alarming situation in terms of fatalities in truck-involved crashes, among the drivers and passengers of the trucks as well as the other vehicles involved. Despite the sizable efforts to investigate the truck-involved crashes, very little is known about the safety of truck movements in developing countries, and about the single-truck crashes worldwide. Thus, this study aims to uncover significant factors associated with injury severities sustained by truck drivers in single-vehicle truck crashes in Iran. The explanatory factors tested in the models include the characteristics of drivers, vehicles, and roadways. A random threshold random parameters hierarchical ordered probit model is utilized to consider heterogeneity across observations. Several variables turned out to be significant in the model, including driver’s education, advanced braking system deployment, presence of curves on roadways, and high speed-limit.

Nabeel Al-Bdairi, University of Wasit
Jason Anderson, Portland State University
Salvador Hernandez, Oregon State University

Although there has been a growing interest in comprehending large-truck crash severity in recent years, what is still not completely understood is the relationship between crash-related factors, injury severity, and unsignalized intersections. Therefore, this research seeks to discuss these relationships and fill a critical gap in the large-truck crash injury severity literature. In this study, a mixed logit model is used to capture the effects of contributing factors on injury severity of large-truck drivers at unsignalized intersections while accounting for unobserved factors (i.e., unobserved heterogeneity). The data used in this study are large-truck-involved crashes that occurred at unsignalized intersections between 2007 and 2013 in the state of Washington. Injury severity sustained by truck drivers are categorized into three categories: severe injury (fatal and incapacitating), minor injury (non-incapacitating and possible injury), and no injury. The results reveal that three parameter estimates, namely, wet roadway surfaces, left turning movements, and drivers who were sober at the time of the crash, are randomly and normally distributed. Moreover, 18 estimated parameters are found to be fixed across observations (i.e., fixed value of the estimated parameter in the mixed logit model). Among these parameters, overcast (i.e., cloudy) weather, driver sobriety, and no restraining systems in vehicles are associated with severe injury, and crashes that occurred in daylight conditions increase the probability of no injury. Results of this study provide a flexible framework to overcome the inherent shortcomings in crash data that lead to biased model estimates and erroneous corresponding inferences by accounting for unobserved heterogeneity.

Xiaobing Li (xli158@ua.edu), University of Alabama
Jun Liu, University of Alabama
Qifan Nie, University of Alabama
Xing Fu, University of Alabama
Shashi Nambisan, University of Nevada, Las Vegas
Asad J. Khattak, University of Tennessee

Truck drivers face many dangers on roadways. Truck-involved traffic crashes may not only threaten the lives of truck drivers but also greatly affect the trucking industry which currently faces a massive truck driver shortage. To improve truck driver safety, studies have been focused on understanding the roles of truck driver behaviors prior to traffic crashes, called pre-crash trucker behaviors. These studies attempted to uncover relationships between pre-crash trucker behaviors and crash severities through modeling traffic crash data. Traffic crashes exhibit complex spatial and temporal patterns interacting with diversified socio-economic, cultural, and geographic contexts, which have not been fully captured in previous studies. The objective of this study is to revisit the roles of pre-crash truck driver behaviors with a focus of examining the spatiotemporal variations in relationships between the behaviors and injury severities. This study employed an integrated spatial-temporal modeling approach, namely the Geographically-Temporally Weighted Ordered Logistic Regression (GTWOLR), to model a Pennsylvania statewide crash database with over 56,000 truck-involved crashes. The results indicate some risk behaviors such as speeding, driving under influence, non-restraint, and cellphone usage were associated with an increased truck driver injury severity. In the spatiotemporal variations, some freight routes or time periods are highlighted for particular behaviors due to the extra high estimation magnitudes, such as I-76 near Pittsburg and the years after 2013 for speeding. The findings are useful for practitioners in developing localized safety improvement programs. More implications are discussed in the paper.

Rajesh Gupta, University of Lucknow
Hamidreza Asgari, Florida International University
Ghazaleh Azimi, Florida International University
Alireza Rahimi, Florida International University
Xia Jin, Florida International University

This paper presents the results of an analysis focusing on recognition of large truck-involved work zone crash patterns. Recognizing the limitations of logistic regression models that were commonly used in crash severity prediction, this study applied machine learning techniques including data resampling and decision tree/random forest models. Using a seven-year large truck involved work zone crash data in the state of Florida, random oversampling and systematic oversampling techniques were explored. Decision trees and random forest models were consequently built for the raw and resampled datasets. From a methodological perspective, results showed that a combination of oversampling with ensemble random forests technique can significantly improve model performance in predicting fatality crashes. Primary contributors included pedestrian involvement, lighting conditions, safety equipment, driver condition, driver age, and work zone locations. In view of fatality patterns, results showed that a combination of different factors can significantly increase the probability of a fatal outcome. Regarding pedestrian crashes, factors such as dark not lighted conditions, distracted truck drivers, airbag deployment, and driver’s age (young drivers outside city limits, senior drivers inside city limits) were highly fatal. For non-pedestrian crashes, the combination of front airbag deployment with any restraint system other than shoulder and belt was quite fatal. Also, abnormal driver condition increased the risk of a fatal outcome. Additionally, the presence of female drivers (in view of multiple vehicle crashes) highly decreased crash severity, probably due to their more careful driving manner compared to males. Interestingly, driver actions and maneuvers as well as roadway design and other physical environment features (i.e., number of lanes, median type, grade and alignment) did not show significant contribution to the model.

Amin Mohammadnazar (amoham17@vols.utk.edu), University of Tennessee, Knoxville
Iman Mahdinia, University of Tennessee, Knoxville
Ramin Arvin, University of Tennessee, Knoxville
Asad J. Khattak, University of Tennessee

Large trucks play an important part in the US' economic growth by transporting a considerable portion of goods and cargoes across the country. However, according to the U.S Department of Labor, in 2017, trucking industry experienced the highest fatalities among all occupations in 2017. Over 75 percent of these fatalities were related to roadway accidents. Police-reported crashes show that angle crashes are the most frequent type of accidents among large truck fatal crashes, which raises concerns over the safety of occupants involved in these types of crashes. Drivers’ physical conditions and driving errors have also been recognized as major factors in truck-involved crashes. Therefore, this study focuses on the effect of at-fault drivers’ physical conditions and driving errors on the severity of large truck angle crashes while controlling for crash characteristics and roadway features. To address unobserved heterogeneity in these crashes, a random parameter ordered probit model was estimated using North Carolina crash and inventory data between 2013-2017. Regarding the physical condition of at-fault drivers, the results show that having a medical condition and impaired driving have strong and positive associations with large truck angle crash severity. It was also found that driving error factors such as aggressive driving, disregarding signs and signals, failure to yield the right of way, and speeding increase the probability of experiencing fatal and injury crashes. This can help technology developers in the trucking industry as well as planners and engineers in the transportation field make informed decisions about safety countermeasures.

Mouyid Islam (mouyid@cutr.usf.edu), University of South Florida
Parisa Hosseini, Rowan University
Mohammad Jalayer, Rowan University

Large truck-involved crashes are one of the core emphasis areas in Strategic Highway Safety Plans (SHSP). Large truck crashes, particularly on rural curved roadways, lead to a disproportionately higher number of fatalities and serious injuries relative to other passenger vehicles over time. The intent of this study is to identify and quantify the factors affecting injury severity outcomes for single-truck crashes on rural curved segments in North Carolina. The crash data are extracted from the Highway Safety Information System (HSIS) from 2010 to 2017. This study applied a mixed logit with heterogeneity in means and variances approach to model driver injury severity. The approach accounts for possible unobserved heterogeneity in the data resulting from driver, roadway, vehicle, traffic, and/or environmental conditions. The model results indicate that there is a complex interaction of driver characteristics such as demographics (male drivers, age below 30 years), physical condition (sleepy while driving), actions (unsafe speed, overcorrection and careless driving), restraint usage (lap-shoulder belt usage and unbelted), roadway and traffic characteristics (undivided road, medium right shoulder width, graded surface, low and medium speed limit, low traffic volume), environmental conditions (rainy condition), vehicle characteristics (tractor-trailer), and crashes characteristics (fixed object crashes and rollover crashes). In addition, this study compared the contributing factor leading to driver injury severity for curved and straight rural segments. The results clearly indicate the importance of driving behavior and roadway design concerning curved segments that need to be prioritized in the trucking agency as well as the roadway design and maintenance agency.

William Sieber, National Institute for Occupational Safety and Health
Marie Sweeney, National Institute for Occupational Safety and Health
Edward Hitchcock, National Institute for Occupational Safety and Health
Cynthia Robinson, National Institute for Occupational Safety and Health
Guang-Xiang Chen, National Institute for Occupational Safety and Health
Jennifer Lincoln, National Institute for Occupational Safety and Health
Imelda Wong, National Institute for Occupational Safety and Health

Driver fatigue and sleepiness may impact driver alertness and performance, which may lead to increased risk of road crashes and poor health.   In 2010, NIOSH conducted a nationally representative survey of long-haul truck drivers to help characterize health and safety risk factors in this population.  Self-reported health conditions, hours of driving or on-duty, and hours of sleep were collected from 900 truck drivers by interview and through a retrospective Sleep and Activity diary.  A multivariable logistic model for high self-reported ratings of fatigue including working/driving character, demographics, and health conditions was developed.  Fatigue was defined as a driver being so tired that the driver needed to take a break or sleep, and was determined by ratings on a Likert Scale. Forty-four percent of drivers indicated self-reported fatigue ratings of 6 or above on the day during the previous week when they felt most fatigued.  Model-based standardized marginal risk ratios showed that high self-reported fatigue ratings increased with number of hours driving or on duty, poor quality of sleep, driving alone, and body mass index (BMI).  High self-reported fatigue ratings were greater for driver health conditions such as heart disease, back pain, and emphysema.  Use of Continuous Positive Air Pressure (CPAP) had a protective effect, as did years worked as a driver in a job requiring a mandatory rest period away from home during each run.   This study suggests that development of trucking policy and programs to minimize long-haul truck driver fatigue should include consideration of working, sleeping, and health conditions.

Vincent Ampadu, University of Wyoming
Anas Alrijjal, University of Wyoming
Khaled Ksaibati (khaled@uwyo.edu), University of Wyoming

As a result of its complexity and inclination for severe crashes during failure, truck braking technology has experienced significant advancement over the past decades. The reduced stopping distance requirement mandated by the Federal Motor Vehicle Safety Standards (FMVSS) rule 121 suggested that the trucking industry should be encouraged to adopt and install disc brakes, especially for fleets that frequently travel over mountain passes. In response to this stopping requirement, some fleets have fitted their steer axles with 16.5 × 5-inch drum brakes, while others have installed disc brakes. This study uses TruckSim TM to model disc brakes and drum brakes on a fully-loaded truck semi-trailer (80,000 lb.) in order to study the performance of each type of brake as downgrades and speeds are varied. The brake performance is measured based on braking distance. A simplified economic comparison based on life cycle cost analysis to determine which road and vehicle conditions give rise to the cost-effectiveness of disc brakes is also performed. The studies demonstrated that disc brakes shorten braking distances by 10- 20%. They also increase the percentage reduction in braking distance by (12-17%) as speed increases and the downgrade gets progressively steeper. Evidence is provided that trucking companies operating their vehicles in steep terrain and at high speeds with disc brakes could benefit from 112%-180% in cost savings in the long term. Finally, at the societal level, by preventing crashes arising from rear-end collisions and runaway truck incidents, disc brakes may be able to save over $649 million annually.

Sarvani Duvvuri, University of North Carolina, Charlotte
Sonu Mathew, University of North Carolina, Charlotte
Srinivas Pulugurtha, University of North Carolina, Charlotte

The presence of trucks in the traffic stream has a psychological influence on drivers in their vicinity due to their size, dimension, and operating characteristics. Truck crashes account for 11% of the traffic fatalities but are only 4.3% of the vehicles in the traffic stream. Trucking activity in a region is governed by its land use development, population, and area type. This research aims to examine the influence of off-network and on-network characteristics on truck crash injury severity to identify potential countermeasures. Crash data for Mecklenburg County for the years 2013 – 2017 was considered for analysis. Buffer analysis was performed to capture the land use and demographic data within the vicinity of each crash. The risk factors and their likelihood associated with the truck crash injury severity were identified using the backward elimination method in the partial proportionality odds model. The results indicate that dark lighting conditions, driver fatigue/impairment, the presence of a van, flatbed, or other semi-trailer, and driver inattention increase the likelihood of a severe or moderate injury truck crash. Further, the likelihood of a severe or moderate injury truck crash is high in commercial, industrial, recreational, and resource land uses. Contrarily, the likelihood of a severe or moderate injury truck crash is less in single-family residential and office land uses. Potential countermeasures to reduce risk and enhance safety include truck signal priority, variable/dynamic speed limit signs for speed harmonization, truck traffic management strategies, incorporation of advanced driver warning/crash avoidance systems, education, and enforcement.

Muhammad Tahmidul Haq, University of Wyoming
Milan Zlatkovic, University of Wyoming
Khaled Ksaibati, University of Wyoming

Disaggregate modeling approach is a new trend in the literature to analyze the injury severity of truck-involved crashes. The assessment of truck driver injury severity based on driver action is still missing in the literature. This paper presents an extensive exploratory analysis that highlights significant variability in the truck driver injury severity based on various action types (i.e. aggressive driving, failure to keep proper lane, driving too fast, and no improper driving). Binary logistic regression with the Bayesian random intercept approach was developed to examine the contributing factors to fatal or any injuries of the truck drivers using ten years (2007-2016) of historical crash data in Wyoming. The log-likelihood ratio tests were performed to justify that separate models by various driving action types are warranted. The results demonstrated the effects of various vehicle, driver, crash, and roadway characteristics, combined with truck driver-specific action, on the corresponding driver injury severity. The gross vehicle weight (GVW), age and gender of the driver, time of day, lighting condition, and the presence of junctions were found to have significantly different impacts on the truck driver injury severity in various driving action related crashes. With the incorporation of the random intercept in the modeling procedure, the analysis found a strong presence (27% – 33%) of intra-crash correlation in driver injury severity within the same crash. Finally, based on the findings of this study, several recommendations are suggested.

James Smith, Western Kentucky University
Mehdi Hosseinpour, Western Kentucky University
Ryan Mains, Western Kentucky University
Nathanael Hummel, Western Kentucky University
Kirolos Haleem, Western Kentucky University

This study takes the initiative and examines various features affecting the severity associated with commercial motor vehicle “CMV” (i.e., large truck and bus) head-on collisions on Kentucky highways. Recent five-year (2015-2019) crash data and variables rarely-explored before (e.g., presence of centerline rumble strips, type of passing zone, and terrain type) were collected and prepared using Google Maps. A total of 378 CMV-related head-on collisions were analyzed. The generalized ordered probit (GOP) model was employed to identify the significant factors affecting the severity level resulting from CMV head-on collisions. The model allows the coefficients to vary across the injury severity categories for reliable parameter estimations. From the preliminary investigation, rolling terrains had the highest share of severe CMV head-on crashes (62% and 71% for multilane and two-lane roadways, respectively). The presence of centerline rumble strips could reduce severe crash outcomes along multilane and two-lane facilities. The GOP model identified various significant predictors of minor and severe injuries from CMV head-on crashes. Occupants wearing seatbelt were 39.3% less likely to sustain severe head-on crash injuries. From the roadway characteristics, presence of median cable and concrete barriers could significantly reduce the probability of severe head-on crash injuries, with median cables being more effective. Regarding the driver characteristics, drug impairment and speeding increased the risk of sustaining fatal/serious injuries by 39.5% and 26.4%, respectively. Necessary safety recommendations are proposed to reduce the severity of CMV head-on-related collisions. One example is installing median cable barriers along roadway stretches with high history of head-on CMV-related crashes.

Nick Stamatiadis, Kentucky Transportation Center
Shraddha Sagar, Gresham Smith and Partners
amantha Wright, University of Kentucky
Eric Green, Kentucky Transportation Cabinet
Reginald Souleyrette, Kentucky Transportation Cabinet

In the United States (U.S.) the annual number of commercial motor vehicle (CMV) crashes has been on an upward trajectory since 2009. In 2016, CMV crashes accounted for 11.8 percent of all fatal crashes in the U.S. And in Kentucky, between 2009 to 2016, the number of CMV crashes rose 27 percent. Of particular concern to state DOTs has been crossover crashes involving CMVs. These occur when a vehicle leaves its intended path and veers into the path of oncoming traffic, typically resulting in head on or sideswipe opposite direction collisions. While some researchers have found that installing cable median barriers can mitigate crossover crashes involving CMVs, no definitive conclusions have been reached. To move toward a resolution of this question, this study leveraged analysis by a panel of experts and the development of safety performance functions (SPFs) and crash modification factors (CMFs) to gauge how cable median barriers can influence the number and severity of crossover CMV crashes on Kentucky interstate routes. Expert panelists contended that cable median barriers will improve safety, a conclusion substantiated by statistical modeling. Despite the study’s limited scope, it appears that installing cable median barriers can prevent or mitigate CMV crashes.

Behrang Assemi (behrang.assemi@qut.edu.au), Queensland University of Technology
Mark Hickman, University of Queensland
Alexander Paz, Queensland University of Technology

Heavy vehicle crashes incur significant economic and social costs. Although most crashes are considered to be related to driver error, the effects of vehicle defects are major in many crashes. Therefore, various vehicle inspections including Queensland’s Certificate of Inspection (COI) scheme have been implemented to improve the safety of heavy vehicles. This study analyses the trends and potential effects of the COI scheme on heavy vehicle crashes. Longitudinal data provided by Queensland’s Department of Transport and Main Roads for the period June 2009 through December 2013 were used to perform the analyses. The data include 474,640 programmed inspections and 2,274 crashes in which heavy vehicles were involved. The results show significant effects of monthly average inspection failure rate as well as monthly average failure severity level on the total number of heavy vehicle crashes. The results also reveal significant effects of monthly average inspection failure rate, average vehicle age, and monthly average mean maximum temperature on the number of defect-related crashes. The implications of these results are discussed with respect to heavy vehicle safety policies.

Mehdi Hosseinpour, Western Kentucky University
Kirolos Haleem, Western Kentucky University

Road departure (RD) crashes are among the most severe crashes that could result in fatal or serious injuries. However, little research has been conducted to investigate RD crash injury severity, especially involving large trucks. Besides, most previous studies neglected to incorporate both roadside and median hazards into RD crash severity analysis. The objective of this study was to identify the significant factors affecting driver injury severity in single-vehicle RD crashes involving large trucks. A random-parameters ordered probit (RPOP) model was developed using extensive crash data collected on roadways in the state of Kentucky between 2015 and 2019. The RPOP model results showed that the effect of local roadways, natural logarithm of annual average daily traffic (AADT), presence of median concrete barriers, cable barrier-involved collisions, and dry surfaces were found to be random across the crash observations. The results also showed that older drivers, ejected drivers, and drivers trapped in their truck were more likely to sustain severe single-vehicle RD crashes. Other variables increasing the probability of driver injury severity have included rural areas, dry road surfaces, higher speed limits, single-unit truck types, principle arterials, overturning-consequences, fired trucks, segments with median concrete barriers, and roadside fixed object strikes. On the other hand, wearing seatbelt, local roads and minor collectors, higher AADT, and hitting median cable barriers were associated with lower injury severities. Potential safety countermeasures from the study findings include installing median cable barriers and flattening steep roadside embankments along those roadway stretches with high history of RD large truck-related crashes.

Zhenyu Wang (zwang9@cutr.usf.edu), University of South Florida
Abhijit Vasili, University of South Florida
Runan Yang, University of South Florida

This study investigated the hierarchical connection among injury severity, non-truck improper actions, and contributing factors in large-truck-involved crashes. Data for four years (2011-2014) of crashes that involved a large truck (≥10,000 pounds) and a non-truck vehicle were collected from suburban roads in Florida. A recursive bivariate probit (RBP) model was fitted with collected data to identify the cause-effect chain, including contributing factors influenced by improper actions, the effects of improper actions on injury severity, and contributing factors indirectly impacting injury severity in large truck-related crashes. Study results indicate that non-truck vehicle improper actions such as excessive speed, careless driving, failure to yield right-of-way, etc., significantly increase the likelihood of fatal and severe injury in large-truck crashes, and factors such as crash month, darkness, intersection-related, surface and shoulder width, truck parking, truck driver age, non-truck driver age, and non-truck alcohol/drug impaired indirectly influence injury severity through their impacts on non-truck improper actions. Two factors—truck right-turn and non-truck driver physical defects—impact injury severity and non-truck improper actions simultaneously. Other factors, including crash year, Annual Average Daily Traffic (AADT), speed limit, crash type, truck type, truck speed, truck alcohol/drug-impaired, and motorcycle involvement, directly contribute to injury severity in large-truck crashes and have no influence on non-truck improper actions.

Sherif Gaweesh, University of Wyoming
Arash Khoda Bakhshi, University of Wyoming
Mohamed Ahmed, University of Wyoming

Traffic crashes can be divided into primary and secondary crashes. Secondary Crashes occur as a consequence of primary crashes within their spatiotemporal location. Secondary crashes comprise nearly 20% of all crashes and 18% of fatal crashes, in which they can possibly have a higher crash severity than the primary crash. Interstate-80 in Wyoming is considered a major freight rural corridor. The Federal Highway Administration selected Wyoming to deploy a Connected Vehicle (CV) program with a focus on commercial truck safety. Distress and rerouting applications were among the CV pilot applications. There is an existing gap in literature about studies that investigated the safety performance of CVs in mitigating the risk of secondary crashes on heavy trucks, more specifically under adverse weather conditions. This study filled this gap by conducting a driving simulator experiment to assess the effectiveness of CV distress and rerouting applications in mitigating the effect of secondary crashes. A total of 23 truck drivers were recruited in this study. Analysis were conducted on the vehicle kinematics obtained from the driving simulator. A CV and a non-CV scenario were designed to compare the participants driving behavior under adverse weather conditions. The results showed that the tested CV applications succeeded to enhance the driving behaviors by reducing the operating speed as well as the speed variation, and all participants avoided a secondary crash under the CV environment. Additionally, distress notification coupled with road closure warning reduced the average operating speed by 26% compared to the static posted speed limit.