International Conference on Transportation and Development 2021: Transportation Operations, Technologies, and Safety

The Highway Capacity Manual provides concise, well-defined procedures for estimating capacity and level of service for most types of roadways and intersections. While not perfect, these procedures allow for a quick comparison of roadway geometries when making design decisions. When detailed information is necessary, the analyst can use microscopic simulation; however, there is a stiff cost to doing so. However, performing an operational analysis (e.g., VISSIM) is not always an option in projects due to the higher associated costs. This is mostly the case in smaller projects, in the initial phases of larger projects where there is limited budget available for providing preliminary results, or in projects with low traffic and simple/non-complex geometry. In these cases, having an HCM-based method for evaluating all possible freeway segments becomes rather important. Also, there is pressure to limit right-of-way needed for capacity improvement projects. This constraint could result in the use of multiple weaving areas. The current Highway Capacity Manual (HCM 2016) provides procedures for three types of freeway segments: (1) basic segments away from ramps, (2) merge and diverge segments, and (3) weaving segments, which are formed when a merge is followed closely by a diverge, increasing the amount of lane changing as traffic lines up to for its desired destination. If the influence of a single ramp is felt beyond the next ramp to a second ramp, we have a multiple weave. The configuration is either two entrance ramps followed by an exit ramp or a single entrance ramp followed by two exit ramps. We use the former in this work. The HCM recommends segregating the weaving movements into separate merge, diverge, and simple weaving segments, which ignores the interrelated behavior of the merging and diverging traffic. In this work, we are using microscopic simulation (VISSIM) to find relations between macroscopic variables flow, density, and speed. Capacity is evaluated by gradually raising flows for a range of geometries and flow ratios. Models are developed to express capacity in terms of lane configuration, flow ratios, truck percentage, and overall flow rate. Also, the impact of distances between the ramps and the percent trucks will be measured. We plan to provide a procedure to allow analysts to evaluate capacity in multiple weaves without having to resort to simulation.

Roundabouts have become a vital part of the transportation infrastructure in the United States. It is critical that engineers consider all variables and accurately predict the performance of roundabouts prior to installation. Two common methods for assessing the performance of roundabouts are the Highway Capacity Manual (HCM) method in the U.S. and the Kimber equation in Europe. The objectives of this study were (1) to determine if geometric factors, such as inscribed diameter, influence roundabout capacity; and (2) to compare driver critical gap behavior in Southern Illinois to values from previous research. In this study, field observations measured vehicle critical gaps and capacity flow rates at roundabouts in Southern Illinois. The results found (1) correlations between roundabout diameter and entry capacity, which suggest an opportunity for future refinement of the HCM methods; and (2) that a critical gap estimate of 4.2 s is appropriate for Southern Illinois, which differs from default values from the HCM (5 s).

In this research, performance measures of the Utah Department of Transportation (UDOT) Traffic Incident Management (TIM) program were analyzed. Performance measures were determined for responding Incident Management Teams (IMT) and Utah Highway Patrol (UHP) units at 168 crashes. Statistical analyses were conducted to understand how the existing IMT program was performing. Results of statistical analyses showed that the effect of IMT response time (RT) on roadway clearance time (RCT) was statistically significant, while RT was not significant on incident clearance time (ICT) at the 95% confidence level. On average, for each minute delay of IMT response time, 0.8 min is added to the RCT ranging from 0.1 to 1.4 min.

Recent five-year bicycle-vehicle crash pattern (2011–2015) in Alabama shows a 78% increase in bicycle fatal crash rate. This may be due to lack of understanding of cycling regulations. This study develops a comprehensive online survey to examine drivers’ and bicyclists’ understanding of cycling regulations in Alabama. Sixty-six questions were developed. The survey included questions on cycling regulations, helmet use, bicyclist/driver behavior, cycling difficulties, and suggested improvements. A total of 1,070 responses were received. Generally, respondents felt unsafe while cycling on Alabama’s roadways. Approximately 63% of respondents felt that drivers were mostly at-fault in bicycle-vehicle crashes. Around 17.5% did not know that riding a bicycle on sidewalks is illegal in Alabama. About 86% of respondents said they wear their helmet “Always” or “Most of the Time”. About 93% of respondents expressed how they felt drivers did not respect bicyclists’ right to be on the roadway. Some of the difficulties identified by bicyclists were motorists passing closely at excessive speeds, presence of rumble strips on roadways, and existence of some road sections with poor cycling conditions. Some recommendations include better cycling education for drivers and bicyclists, adding more bicycle parking, and promoting cycling through intelligent transportation systems applications, e.g., displaying dynamic message signs with “SHARE THE ROADWAY” information.

Increasing congestion challenges call for better utilization of existing infrastructure by opting for data-driven and cost-effective solutions. Intelligent transportation systems (ITS)-based solution like the variable speed limit (VSL) is a popular advanced data-driven solution to enhance mobility and safety by harmonizing speeds along a road. Prior to allocating resources to a road, a thorough analysis of travel and speed patterns is needed. This paper focuses on investigating road link-level data to identify the threshold of speeds for different peak hours and, thereby, investigating potential parameters influencing the speed of road links. Multivariate clustering analysis was first used to identify potential road links susceptible to speed variation for the implementation of VSL. A supervised machine learning algorithm, forest-based classification and regression, was then used to model and examine the influence of the average annual daily traffic (AADT), the historical speed of the road link, and the average speeds of upstream and downstream road links on the average speed of corresponding road link. Modeling and validation were performed using morning and evening peak hour data for Mecklenburg County, North Carolina, USA, for all the road links as well as for road links with low- and high-speed variation.

Explosive population growth in urban areas increases traffic on already congested highways, particularly during rush hours. To tackle this type of issue, transportation agencies need to evaluate various countermeasures preferably with a tool designed to evaluate the impact of each potential policy. This work describes the design of a dynamic traffic assignment (DTA) simulation-based model to evaluate different polices for alleviating urban congestion. The DTA model was based on data extracted from an already existing static regional traffic demand model. The static model provided travel demand data and the network geometry. The Austin metropolitan area was used to test the traffic impact of two polices: (1) removing the tolls from nearby alternate highways to hopefully attract travelers using the overburdened main urban highway corridor; and (2) adding capacity to the existing main urban highway corridor. Results, including systemwide, corridor, and link levels, show that removing the tolls from the selected highway can reduce the total system travel time, but it is not an ideal solution for the urban traffic because the toll roads are too far away from the urban area to be a justified alternative for the urban travelers, supported by the results from the corridor and link levels. On the other hand, adding extra capacity on existing congested highway does attract more users and ultimately negates the benefit of the additional capacity. In addition, passenger cars and trucks behave differently under different polices.

Growth of road users in urban areas results in consequential higher interactions between pedestrians and vehicles causing delay to the mainstream traffic flow. The delay caused by pedestrians who make random jaywalking along the carriageway is substantial when such behavior exists. Thus, the prime objective of this research was to study the interaction and develop a delay model to estimate the collective delay caused to the mainstream traffic which encounters jaywalking pedestrians. This delay is influenced by the characteristics of the crossing pedestrians and the behavior of on-coming vehicles. The data collection was done in a suburban city near Colombo, Sri Lanka. The data were extracted from video footages taken using a drone camera. Both the movement of the vehicles and crossing pedestrians on the subject lane were tracked using automated software in order to enhance the accuracy of the results. The delay caused to mainstream vehicle was derived using the deceleration and acceleration behavior. The proposed delay model exemplifies that the pedestrian-vehicle gap and the pedestrian speed along with other relevant pedestrian characteristics such as age, pedestrian speed at the start of the vehicle speed drop, and vehicle-related characteristics such as vehicle speed at the start and end, veh-ped gap at the vehicle speed drop, subject lane, and vehicle type are highly significant to the delay of the subject vehicle on the mainstream. The overall R value of 0.63 was observed from the regression analysis of the proposed delay model. The applicability of the proposed model for each pedestrian and vehicle characteristic has been determined and evaluated based on their level of significance.

The development of faster, improved, and less costly sensor technologies has made it easier to obtain large and much more disaggregate (“BIG” data) streams of traffic related data along freeways and arterials. As such, the raw data that are collected have great impact on providing more credible insight with regard to information such as speed and travel time distribution(s), which can subsequently provide reliable travel information. The average speeds on a segment of roadway can be impacted by different factors that can exist for a specific period. However, there are limited studies that have focused specifically on investigating speed distribution fitting under free-flow and recurrent congestion based on day-to-day real traffic observations. Whereas it is common to assume that speeds are normally distributed, it is essential to take into consideration all possible alternative models for the specification of a distribution that can most appropriately characterize daily vehicles’ speed variability. The aim of this study is to investigate the characteristics of speed distribution patterns under free-flow and recurrent congestion by fitting different models based on day-to-day real traffic observations. Conceptually, vehicles’ speed can experience different ranges of uncertainty because of stochastic traffic flow, random delay due to weather events, traffic direction, vehicle classification, and the number of lanes. Another goal of this study is to determine the speed distribution patterns based on different temporal aggregation levels on two different types of roadways. Traffic data, along with vehicle classifications, have been collected from two study sites located on Interstates I-71 and I-75 in the state of Ohio. In this study, different statistical probability distribution models were explored. The probability distribution models are analyzed based on different scenarios, such as temporal aggregations, spatial scales, lane types, and vehicle classifications. A goodness of fit test has been conducted for a comprehensive evaluation of the performance of distribution models. Results showed that the Burr, log-logistic, and t-location distributions would provide a relatively promising fitting to vehicle speed in the urban segment considering all temporal and lanes aggregations. Standard normal distribution, Burr, and t-location models are evaluated as superior fitting distribution to vehicle speed compared to their alternatives in the rural segment. Maximum likelihood has been used to estimate the probability distributions parameters. The results also exhibited that heavy vehicles, particularly class 9 trucks, have high acceptance rate distribution fitting compared to other classes. Also, the distribution of class 9 vehicles traveling on the outer lane shows a significant difference in the probability distribution parameters compared to inner lanes. This approach offered details on speed distribution and provided critical insights into the dynamics of congestion and the variability of travel times. Other applications that can benefit from these data could be enhancing the reliability of travel time prediction and estimation, mode choice, and travel demand models.

The categorization of the traffic state into breakdown and non-breakdown is critical to traffic flow analysis and effective traffic management and operations. Due to the data availability, the identification of the traffic states has been mainly based on the three macroscopic measures (speed, occupancy, and volume). Emerging new technologies will allow the collection of microscopic measures that can be used in combination with the macroscopic measures for better recognition of the traffic state. Since stop-and-go operations result in traffic disturbance, this study developed disturbance metrics based on microscopic measures to examine their capability for better traffic state categorization. The utilized disturbance metrics are the number of oscillations (NO) and a measure of disturbance durations in terms of the time exposed time-to-collision (TET). The study found that adding traffic disturbance metrics in the data clustering when identifying the traffic states will result in better traffic breakdown recognition by capturing stop-and-go in the traffic stream.

Connectivity between vehicles will enable platoons of connected automated vehicles (CAV) to operate on their dedicated lanes on highways in the future. Platoon is an entity comprised of several parameters which generate the throughput of the platoon lane. These parameters are the headways, speed, and length of platoon. The parameters are either discrete or continuous and have a random nature, while being within certain acceptable ranges determined for the safety and functionality of the platoon lane. This article investigates how the change of the platoon parameter’s effect the throughput of the platoon lane. Safety, comfort, and operational robustness are the major factors in determining the formation and operation of platoons. Currently, researchers and engineers are testing the influence of platoon parameters in the operation of these lanes. This article investigates the mathematical structure of the platoon lane parameters and how the change of the platoon’s parameter effect the throughput of the platoon lane. The results provided a sensitivity analysis and the nature of the platoon lane parameters and can be used for research and analysis of platoon lanes.

This study presents an in-service performance evaluation (ISPE) of high-tension median cable barriers and strong-post W-beam guardrails along Interstate 85 (I-85) in the State of Alabama. Crash data from 2012 to 2016 were used. The study compares the safety performance of median cable barriers and median guardrails in terms of crossover/non-crossover percentages and injury crash percentages. The study also conducts an ISPE of roadside guardrails on I-85. Extensive data collection effort has been made through reviewing in detail police reports and capturing features rarely explored in previous studies, e.g., median width, left shoulder width, barrier offset distance or lateral position (i.e., distance from barrier to median edge), blockout type, and number of cables/strands. A total of 495 crashes were used, broken down as 220 median cable barrier crashes (MCBCs), 79 median guardrail crashes (MGCs), and 196 roadside guardrail crashes (RGCs). Roadside guardrails showed the highest success in preventing crossover crashes (98%), followed by median cable barriers (94.5%), and finally median guardrails (86.1%). Median cable barriers were more successful in preventing sport utility vehicle (SUV) crossovers (91.2% success rate) compared to median guardrails (79.2% success rate). Median cable barriers had less injury crash involvement (12.7%) compared to median W-beam guardrails (34.2%). It was also found that composite blockouts were more involved in injury-related median guardrail crashes (32.9%), as opposed to roadside guardrail crashes (23.3%). Some recommendations from the study have included ensuring installing cable barriers on medians that are at least 50-ft wide, as well as ensuring installing W-beam guardrails on medians that are 50–70-ft wide.

This study investigated an idea of using a location-based surrogate measure of risk (credit rating) to improve crash prediction. The work presented here is the initial step in this investigation specifically examining if any relationship exists between location-based credit rating and crash frequencies. Because traffic volume is highly correlated with traffic crashes, the sites examined were grouped into bins by annual average daily traffic (AADT). Over 1,300 intersections in California were examined with 5 years of traffic and crash data. Credit scores were obtained for the zip codes where these intersections were located to determine if there is any relationship between credit score and number of crashes. The analysis showed positive and negative trends between credit score and crash risk for differing levels of credit rating and AADT. Future work is recommended to further examine the trends revealed in this research and determine if the use of driver-related risk surrogate measures to account for the heterogeneity of driver populations across jurisdictions could improve traffic crash prediction models through the development of a crash modification factor (CMF).

In U.S., every year more than 36,000 people die and approximately 4.4 million are injured in traffic crashes. Immediately after a crash occurrence, emergency medical services (EMS) play a critical role in saving lives by providing patients with the earliest urgent medical care and a means of transportation to medical facilities. It is crucial that the EMS response time (EMS-RT), defined as the time required for EMS to reach the site after notification of the crash, be kept as low as possible in order to reduce the crash severity. This EMS-RT may depend on many factors including weather condition, light condition, traffic condition, land use, and location of EMS facilities. In this study, the factors that may have an impact on the EMS-RT were identified using statistical modeling technique. More than 58,000, EMS-RTs were investigated for this study, and results revealed that EMS-RT is significantly higher in adverse weather conditions, on weekdays, within nighttime hours, in urban areas, and on highways as predicted. Also, as the number of crashes in a specific day increases, the EMS-RT is notably on the rise as well.

A driving simulator experiment was conducted to evaluate the effectiveness of an audio-visual warning system (visual display on the windshield) that communicates the presence of a potential non-connected red-light-running vehicle to the driver of a connected vehicle. Twenty participants were recruited and randomly placed in a control group (warning system activated at the stop bar) and three treatment groups (warning system activated at 50, 100, and 150 ft upstream of the stop bar). Kruskal–Wallis test showed a statistically significant difference in the mean of reaction time between the different groups. Mann–Whitney Wilcoxon test was used to compare between two groups. Drivers reduced their speeds for an average of 2.15, 2.24, 2.59, and 3.15 s and came to a complete stop for 29.73%, 17.5%, 29.27%, and 47.7% when the warning system was activated at the stop bar and 50, 100, and 150 ft upstream of the stop bar, respectively.

Traffic incidents often require lane or road closures to facilitate the incident clearance and ensure the safety of first responders. With the integrated corridor management (ICM) strategies, upstream traffic could be diverted to parallel roadways and the incident-induced traffic congestion is managed in an integrated system rather than individual facilities. This study delivers an example of estimating the incident-induced user delay costs with ICM strategies. A dynamic traffic assignment (DTA) tool is used to quantify the travel delays caused by an incident under various ICM strategies. In particular, a case study based on a fatal crash on the Interstate 35 in Austin, Texas, is performed under three types of ICM strategies. The DTA can capture more realistic traffic flow characteristics by considering time-dependent link information, compared to the static traffic assignment (STA) method. DTA parameters are specified to approximate the driver behavior under different ICM strategies. Emerging traffic data were collected to validate the study results. All tested ICM strategies are found to be associated with decreased travel times and user delay costs, but the one that offers both timely detour options and proper adjustments to traffic controls on parallel roads can be more effective in reducing corridor-wide delays. This study contributes by demonstrating the implementation of DTA models to model driver behavior during traffic incidents and the deployment of ICM strategies. DTA models are a valuable tool to evaluate the region-wide effectiveness of traffic management strategies during traffic incidents.

A driving simulator experiment was used to determine if an advanced warning message (in the form of auditory and visual cues) can have an impact on the detection of pedestrians/bicyclists. While twenty-one participants were recruited for the experiment, the analysis was performed for data collected from 19 participants. The participants were asked to detect the presence of pedestrians/bicyclists by pressing a button device on the steering wheel. The warning system was activated at 20, 30, and 40 s ahead from reaching a pedestrian/bicyclist. When the warning system was activated, statistical tests suggest that participants detected the presence of pedestrians/bicyclists 25 ft earlier than when no warning system was activated. The variances in the location where pedestrians/bicyclists were detected were lower when the warning system was activated. When the warning system was activated, for most events (approximately 73%), there was a speed reduction of 4.9 mph on average observed.

A “road diet” is a technique that reconfigures the roadway by reducing the number of lanes in hopes of achieving traffic calming and improving safety. Common road diets involve changing a roadway from a four-lane undivided roadway to a three-lane roadway with a center two-way turn lane. This study aims to evaluate the safety effectiveness of road diets in Virginia using the empirical Bayes method and focusing on the common cross sections of AAAA to BATAB and AAAA to BuATABu (A = automobile/general purpose travel lane, B = standard bicycle lane, Bu = buffered or separated bicycle lane, and T = one-way or two-way left-turn lane). A total of 40 sites were identified in 15 Virginia cities. A crash modification factor (CMF) of 0.65 is recommended for total segment crashes and a CMF of 0.41 is recommended for fatal and injury (FI) crashes. The standard errors for both these CMFs are 0.125 and 0.154, respectively, and both CMFs were found to be statistically significant at a 95% confidence level. Further research and data collection are needed to expand this work for other roadway cross sections as well as development of intersection CMFs. Based on the computed results, it is concluded that road diets can potentially reduce crashes and public agencies should consider the benefits of road diets in both safety and mobility when justifying roadway improvements.

When presenting construction to a freeway, it almost guarantees the flow of traffic will be hindered and the risk of a crash increases. Two prominent different types of maintenance of traffic have been used: rolling slow down and roadblock. In a rolling slow down, a pilot car will enter in front of a platoon of vehicles and moderately slow them down while allowing the cars ahead to continue at speed. This change of pace will cause a gap in traffic. These gaps are utilized by workers to adjust lane configurations or cross the roadway with machinery. Contrary to the idea of “keeping the traffic at a forward momentum”, the idea of a temporary road closure is being considered. If the roadway was closed for a short amount of time, the construction workers would not have any risk due to passing drivers. The purpose of this study was to use comparative analysis to determine the most safe and effective way to maintain traffic during high risk roadway change. For this study, four types of data were collected: crash data, traffic data, cost of implementation, and geometric conditions. From this data, operational analysis, cost analysis, and safety analysis were done. Results show that roadblock performs better than rolling slowdown.

This study estimated the Highway Safety Manual (HSM) calibration factors for freeway ramp segments using Kansas data and further identified the challenges faced during the calibration effort. At the time of this study, no comprehensive study has been conducted under this research focus. The results indicated that the HSM methodology overpredicted all single vehicle crashes and multiple vehicle fatal and injury crashes resulting calibration factors 0.165, 0.368, and 0.957, respectively, and underpredicted multiple vehicle property damage only crashes resulting a calibration factor of 2.737 for entrance ramps in Kansas. Considering exit ramps, the HSM methodology underpredicted all multiple vehicle crashes resulting calibrations factors 5.426 and 7.973 and overpredicted all single vehicle crashes resulting calibration factors 0.179 and 0.550 in Kansas. The major challenge of this entire calibration was the identification of ramp crashes, where the study manually reviewed 9,720 interchange-related crash reports to locate crashes occurred on ramps.

The outcome of novel coronavirus (COVID-19) has a profound impact on all industries including the transportation sector. Transportation industry which is deemed as an essential activity should continue to operate through the lockdowns while making it certain to exercise social distancing. Although some researchers, Departments of Transportation (DOT), Federal Highway Administration (FHWA), and private organizations have provided vital information and guidelines to mitigate the unintended consequences of the pandemic, the impacts of COVID-19 on various types of transportation modes have been rarely analyzed and compared. This paper gives rich information into the system and cognizance of the effect of COVID-19 and the developments that the transportation area has needed to embrace so as to cling to the lockdown rules while being operational and productive. The study contributes towards advising policymakers, stakeholders, and agencies on certain insights gained from the management of the COVID-19 from a transportation area point of view. Moreover, four key regulations are concluded for leaders inside the transportation framework business to reevaluate and act to manage the pandemic emergency. This research analyzes how each sector has addressed the short-term and long-term effects of the pandemic and results of the study concluded that the survival of transportation industry is not solely reliant on the duration of the COVID-19 pandemic but on the resilience adapted by the different sectors to cope up with the situation. The findings also indicate that transportation departments should include a management plan to reduce the aftermaths of the pandemic and use the lockdown as an opportunity to address surging maintenance backlogs, improve technology, and safeguard the workforce.

Transportation agencies seek efficient methods to figure out how to reduce the severity of crashes and prioritize safety improvement investments. Police-reported crash data are the main source of information for transportation engineers to systematically address safety. However, other additional data sets exist that can help explain factors associated with variance in crash outcomes. In this research, crash data collected by Kentucky State Police are linked to associated emergency medical services runs. First, time and location were used to link the data set as the initial assessment. Furthermore, other related fields, such as crash type and the number of injuries, were utilized to investigate the matches that were not unique. Later, the practical implications of using the linked data set are addresses.

Older drivers aged 65 years and older are more vulnerable to fatal crashes due to cognitive impairments and frailty. The objective of this study was to determine factors associated with intersection-related crashes involving older drivers in the Midwestern states. A five-year fatal crash data set from 2014 to 2018, from the Fatality Analysis Reporting System (FARS), was utilized. The logistic stepwise selection procedure was applied to identify statistically significant predictor variables. The model identified eight statistically significant predictor variables out of thirty. The predictors that increase older-driver single-vehicle fatal crashes at the intersections are if trafficway type is two-way undivided, if the land use is urban, if the intersection is with control, if the time is between 8 p.m. and 8 a.m., if the age is 75+ years old, if the posted speed limit is lower than 55 mph, if the pre-crash event is speeding, and if the most harmful event is other than rollover or hitting a tree.

The paper presents an innovative method to disclose and quantify the cause-and-effect mechanism between such a connected vehicle (CV) behavior performance and the traffic safety of a highway facility by using a safety surrogate approach. In this study, the CV-affected driving behavior parameters are integrated into the state-of-the-art traffic simulation model, in which the interactions between the CV functional attributes (e.g., CV function type, message type, message timing, and drivers’ compliance level to CV messages) and drivers’ behavior attributes (e.g., perception-reaction time, desired speed, and desired following distance) can be virtually observed and determined. The developed system is examined via a case study of a freeway site in the Cincinnati area, Ohio. The results indicate that the CV-affected reaction time, desired headway, and speed are responsible for the reduction of the traffic conflicting frequency and intensity, which are insightful to the successful deployment of the CV safety technologies into the existing highway systems.

Aggressive driving and driver distraction have widely been reported as the two main causes of traffic crashes. According to the police crash database for the years 2015–2019 in Kentucky, crashes affected by driver aggressive violations and distraction activities accounted for nearly 20% and 36% of the total crashes involving commercial motor vehicles (CMVs), respectively, while they were responsible for 31% and 41% of the total severe crashes, respectively. This study aims at comparing the injury severity outcomes (severe versus non-severe) concerning aggressive-driving and distracted-driving crashes under various circumstances. Recent five years of CMV-related crashes (2015–2019) were collected from the Kentucky Transportation Cabinet (KYTC). Separate Z-tests of proportions and chi-square tests of independence were, respectively, applied to identify the factors affecting the severity of crashes for each of aggressive-driving and distracted-driving CMV-involved crashes. The overall results of Z-test comparing the proportions of severe versus non-severe injuries showed that both aggressive- and distracted-driving crashes significantly increased injury severity. The test further demonstrated that, for all the significant variables, the proportion of severe injuries in aggressive-driving crashes was higher than that in distraction-related crashes. Interestingly, road segments with lanes equal to or less than 11 ft increased the risk of severe aggressive-related CMV crashes (odds ratio = 1.30), but reduced the risk of severe injuries in distraction-related CMV crashes (odds ratio = 0.77). Head-on crashes, use of alcohol or drugs, angle collisions, going straight ahead, and paved right shoulder were the top five factors increasing the risk of severe CMV-involved crashes caused by aggressive-driving and distraction behaviors. On the other hand, seatbelt use, sideswipe collisions, urban areas, at-fault CMVs, and annual average daily traffic (AADT) per lane greater than 10,000 were the top factors reducing the probability of severe injuries in CMV crashes related to driver aggressiveness and distraction. The study findings suggest that launching driving safety campaigns, removal of road distracting elements (e.g., billboards) at high crash risk spots, and intensifying traffic enforcement can reduce the severity of CMV crashes affected by aggressive-driving and driver-distraction acts.

Large trucks are a vital mode for freight transportation. Increasing demand in freight transportation increases the risk of truck-involved crashes on highways. Truck-involved crashes counted for 4,761 deaths in the United States during the year 2017, which is 12% more than the death toll of the year 2008. We gathered four years (2016–2019) of motor vehicle crashes involving large trucks in New Jersey for further analysis. The data set is classified into three injury severity categories, including severe injury, possible injury, and no injury. To predict the crash severity, we trained and tested the data set with the three most commonly used machine learning models, including support vector machine, random forest, and boosting methods. The performance of the models is evaluated by their precision, accuracy, and recall. Further, a sensitivity analysis was performed to demonstrate the impact of the top contributing factors in truck-related crashes. The findings of the study will help practitioners and policymakers determine the effects of influential parameters on injury severity outcomes and take necessary countermeasures to minimize the frequency and severity of large-truck crashes.

The Highway Safety Improvement Program allows for funds to be allocated to safety-related projects on roadways. Given the high need for safety improvements and limited funding, prioritization of safety projects is crucial. To assist transportation agencies in making safety improvement project and resource allocation decisions, a computerized suite of tools has been developed that parses the crash database to identify crash hotspots, suggest potential treatments, and calculate possible safety benefits. To demonstrate the applicability of the tools, a case study was conducted using real data provided by Texas Department of Transportation, Austin District. The results returned a set of 6,174 candidate project recommendations with corresponding most effective treatment. The tool can assist engineers in making more insightful decisions regarding safety projects and allocate the limited budget more effectively and efficiently.

Effective visibility methods are crucial for the safe operation of snowplows. Despite modern devices and methods, snowplow crashes cause economic and casualty losses. According to the National Climatic Data Center (NCDC) database, the number of heavy snow events was more than 25% of all winter weather events over the past decade in Texas. The objective of this research is to survey the effectiveness of the current snowplow visibility methods in Texas. Fifteen TxDOT districts (covering nearly all of north Texas) responded to the survey. The survey results brought forward an accurate representation of the effectiveness of warning lights, message signs, retroreflective markings, rear airfoils, and smart snowplow systems—the snowplow visibility methods in Texas. This research led to recommending warning lights and retroreflective markings to be used in specific colors and locations to increase the visibility of snowplows. On the other hand, rear airfoils were not recommended as they had not increased the conspicuity of the vehicles effectively. These recommendations guide TxDOT to implement more effective snowplow visibility methods.

Periodically monitoring, maintaining, and managing infrastructure assets are one of the key challenges faced by transportation agencies around the world. As per regulations, inspectors need to perform these tasks often to ensure optimum performance of the infrastructure with low life cycle costs. During the inspections, either the traffic operations are interrupted or the inspector can only get partial access to the infrastructure elements under inspection. These inspections include collecting data in different formats like notes, forms, and photos, and most of the time depends on the experience of the inspector, therefore, are subjective. Stationary and mobile laser-based instruments, which are expensive, are being used to scan the area, and these results upon analysis can provide an objective evaluation of the asset condition. However, there is a need to identify an effective data collection tool that is economical and can access hard-to-reach areas. The advent of unmanned aerial vehicle (UAV) platforms coupled with photogrammetry technologies has provided a solution to create a digital twin of the infrastructure asset and monitor its condition with minimum human interactions. The digital twin is a model-centric environment that offers a holistic view of real-field conditions of the infrastructure in a virtual environment facilitating effective information sharing and analysis. This study presents two case studies where a UAV mounted with a visible camera was used to create digital twins of different phases of infrastructure for better monitoring and management of transportation assets. This study provides an approach that can benefit transportation agencies in conducting preventive rather than reactive maintenance works, which can result in significant cost savings.

In the USA, the infrastructure is aging rapidly due to a combination of factors including loading, materials, weather, and other environmental conditions. This is evident from ASCE’s recent report card on the infrastructure condition. Roads, which form the backbone of the infrastructure and provide connectivity to remote areas, were graded “D” overall. The current situation necessitates the selection and adoption of proper pavement management practices and related data collection tools to support proactive monitoring that can help in preventive maintenance. In recent times, there has been a rise in holistically analyzing and visualizing transportation infrastructure assets using intuitive models. These are being accomplished using unmanned aerial vehicle (UAV) platforms mounted with various compact sensors like visible, infrared, multi-spectral, and hyperspectral cameras; and LiDAR. Over a period, the initial geometric design conditions deteriorate and further lead to distresses on the pavement. In this study, close-range photogrammetry, accomplished using a visible camera mounted on a UAV, is used to study geometrical features and assessments of the pavement condition. The collected aerial images are processed to build three-dimensional (3D) models for assessing the current condition of the design features of pavements including longitudinal and lateral slope information. One of the key benefits of this approach is the facility it provides to analyze a single 3D model for different pavement attributes. This technology is expected to complement the existing traditional monitoring methods and quickly identify problematic sections of pavements for future repairs.

Autonomous vehicles (AVs) are new technologies with the potential of becoming a viable mobility option that mitigates key transportation barriers. This study aims to understand the concerns and preferences of a proposed AV service in Arlington, Texas, through focus group discussions with both university and public participants. Results indicate focus groups were both concerned about the spatial accessibility, capacity, trip cost, and disability friendliness of the AV service. University participants believed that the AV service would be safer than other services because of its speed; however, they were less tolerant of accidents caused by self-driving vehicles than human errors. Accessibility to recreational facilities at late hours was important for both groups. Participants expected the AV service could improve the mobility of seniors in Arlington. Besides, technology improvements were considered as a factor that increases the reliability of AV service. This study provides new insights into implementing efficient AV services.

This study explores how people with disabilities perceive and accept autonomous vehicles (AVs) as a technology to improve their mobility. A focus group discussion was conducted to explore individuals’ preferences towards integrating level 4 AVs into the existing microtransit service in Arlington, Texas. Participants demonstrated a positive perception towards the integration of AVs into the current microtransit infrastructure. The results suggest that accessibility to a well-designed built environment is vital in adopting AVs by people with disabilities. Moreover, AVs’ accessibility to healthcare facilities is one of the main concerns identified by focus groups of persons with disabilities. In particular, participants with visual impairment were hopeful that future AV services could improve their mobility through advanced apps, booking systems, and vehicle equipment. This study offers several implications for designing AV service in line with the needs of persons with disabilities while combining with the current microtransit service.

According to the latest statistics by the National Highway Safety Administration, approximately 1,830 young drivers (15–20 age) are involved in fatal crashes. Driving tasks such as keeping the vehicle within the lane, observing traffic signs, turning, and identifying any sudden danger involve various cognitive processes including attention, memory, vision, spatial orientation, motor control, executive function, and decision making. This study proposes an innovative way to assess driver’s behavior, especially focusing on young drivers, by applying the emerging sensing technologies, functional near infrared spectroscopy (fNIRS), driving simulator, and analytical methods (e.g., machine learning). Study findings show machine learning can achieve an average of 97.5% accuracy when using fNIRS data and 70% accuracy when using driving data. While the team acknowledges the limited data set is used in this paper, this study opens the doors for applying the emerging machine learning methods to driver’s cognitive processes to further understand young individual’s driving behavior.

There have been limited efforts to investigate the potential of using detailed trajectory data obtained from connected vehicles and/or other sensors in deriving measures for use in real-time traffic state estimation. This study utilizes a hybrid machine learning approach that classifies the traffic states as a function of traffic disturbance and safety surrogate metrics estimated based on detailed trajectories combined with macroscopic traffic metrics. The investigated disturbance metrics are the number of oscillations, and a measure of disturbance duration based on the time exposed time to collisions. The study, first, used unsupervised clustering techniques to classify traffic states into “breakdown” and “non-breakdown” in terms of both mobility and safety. Then, the categorized traffic state was used as a binary response to the macroscopic and microscopic metrics, as features, to train supervised machine learning techniques for predicting traffic flow breakdown in the following 5-min interval in real-time operations. The study found that the utilizing disturbance and safety surrogate metrics in the real-time classification of traffic flow state increases the accuracy of prediction.

It is not well understood how the imminent use of autonomous trucks for freight transportation would change the serviceability of our existing freight infrastructure. For instance, it is not clear if the current highways are maintained well enough to support the early adoption of autonomous trucks; it is not clear if new facilities will be needed to support the novel technologies that would be driving the autonomous trucks; it is also not clear how these need will evolve as autonomous trucking technologies. A careful study of existing literature, including research papers, technical reports, and news articles, was performed to categorize the currently available and under-development automated trucking technologies into various levels. This review was then used to establish a clear definition of autonomous trucking technologies. Next, elements of the current freight infrastructure system that are most likely to be impacted by autonomous trucking technologies were evaluated. The impacts were evaluated for three levels of autonomous trucking technologies. Also, potential challenges specific to each infrastructure element for the successful adoption of autonomous trucking technologies were identified. Based on the study, transportation infrastructure elements such as road signs and markings, parking, managed/dedicated lanes, inspection and maintenance infrastructure, transfer hubs, and roadside equipment were identified to be majorly impacted by the use of autonomous trucking technologies. The findings of this study are expected to be highly beneficial for freight infrastructure planners in formulating freight infrastructure development plans to facilitate the adoption of autonomous trucking technologies and maximize their benefits while minimizing their risks.

The focus of this paper is on modeling and assessing the influence of connected and automated vehicle (CAV) penetration levels on a freeway (I 85) and an urban arterial (NC 49) road operational performance. The real-world speed, traffic volume, and signal data obtained for a small network in Charlotte, North Carolina, were used to develop and calibrate the non-CAV environment simulation model. The driving behavior, acceleration/deceleration profiles, time headway, and space headway were modified to calibrate the simulation model under non-CAV environment. Different scenarios with varying penetration levels of CAV and non-CAV environment (0% to 100%, increments of 20) were built and compared for each road facility. The macroscopic traffic flow parameters, viz., speed, flow, and occupancy (traffic concentration, an indicator of density), were computed and evaluated for different scenarios under a steady state condition. The results help understand the benefits of CAV deployment at an aggregate level by road facility type.

The license plate recognition (LPR) data have become widely available at urban intersections in China. The LPR data is featured with measured headways with time stamps at all approaches of the intersections such that traffic flow rate and operational conditions can be further estimated in replacement of traditional loop data for actuated traffic signal control. This paper presents a study of developing innovative algorithms to dynamically determine the signal control parameters against varying traffic conditions measured by the LPR data sets. The performance of the proposed control algorithm was evaluated via simulation by comparing the LPR-based actuated signal control scheme and the existing pre-timed control, with the use of the data obtained at intersections in City of Jinan, Shandong Province, China. The results indicate an obvious improvement of the traffic control performance and initially prove the developed algorithms adaptive to the LPR-based actuated traffic signal control schemes.

COVID-19 and resulting lockdowns have had a tremendous impact on our daily activities, and subsequently, the traffic on the roads. In this project, we analyze the effect that COVID-19 has had on driving behavior in Austin with respect to vehicle speed and traffic crashes. Data on were collected from the City of Austin open data portal. It was found that the crash count was significantly lower in 2020 in comparison to 2019. Even though there was significant drop in number of crashes, the death count increased close to 14% since 2019. Meanwhile, the average speed in Austin had increased during the pandemic and was higher in 2020 than in 2019. The largest increases in speed occurred during the morning from 7 to 9 a.m. and evening from 4 to 6 p.m. These changes in crash count and speed were observed to be the strongest with the onset of the lockdown and wane as the pandemic progressed.

Wildlife vehicle collision, commonly called roadkill, is a nascent threat to both humans and wild animals. The collision results in property damage, injuries, death, and financial losses to society and mankind. An automobile system is integrated with alert notification, image processing, and machine learning models. This study explores a newer dimension for wild animal detection and signals the driver during active nocturnal hours. The intelligent system uses histogram of oriented gradients (HOG), which extracts the essential thermography image features; next, the extracted features are fed to the pre-trained, convolutional neural network (1D-CNN). This intelligent system has been tested on a set of real scenarios and gives approximately 91% and 92% accuracy in the alert notification and detection of the wild animals in the transportation road system in the city of San Antonio, TX, USA. This proposed system will contribute to the reduction of vehicle collisions caused by wild animals.

Intelligent transportation technologies are aiming to improve last-mile urban logistics and reduce traffic congestion. Congested metropolitan cities are implementing last-mile delivery robots to make the delivery cheaper and faster. A key factor for the success of automated delivery robots (ADRs) in the last-mile is its ability to meet the fluctuating demand for robots at each micro-hub. Delivery companies rent robots from micro-hubs scattered around the city, use them for deliveries, and return them at micro-hubs. This paper studies the dynamic assignment of the robots to satisfy their demands between the micro-hubs. A mixed integer linear programming (MILP) model is developed, which minimizes the total transportation cost by determining the optimum required fleet size. The result determines the number of robots required for each planning period to meet all the demands. It provides algorithms to operate and schedule the robot-sharing program efficiently with the same approximation guarantee for all related problems.

Freight transportation is a growing industry in both developed and developing countries. Products ranging from groceries to medical supplies require quick and reliable shipping. In this era of e-commerce, consumers expect reliable shipping service standards. Further, quicker shipping options are considered a requirement rather than an option for the freight companies responsible for first- and last-mile deliveries. The increasing demand for freight transportation and better shipping standards is leading to high truck traffic on roads. Transportation planners, engineers, designers, and freight operators have been making significant efforts to improve the reliability of truck transportation services. Travel time reliability measures are an effective way for planners to gain an understanding of traffic stream performance that will help them make informed decisions. These measures are developed using passenger car or mixed traffic stream data, making their applicability to trucks (exclusively) questionable. This research focuses on gaining an understanding of whether travel time reliability measures of trucks and passenger cars significantly differ from each other. Travel time data for roads in Mecklenburg County, Wake County, and Buncombe County in North Carolina were considered for analysis. Descriptive statistics were computed, and statistical analysis was conducted to examine the distinctive nature of their behavior. Differences in values and significance of performance measures were observed by facility type, time-of-the-day, and day-of-the-week.

Previous studies have found correlations between the implementation of transit signal priority (TSP) and the reduction in number of crashes. To further understand how TSP affects traffic safety, a more in-depth evaluation was carried out using detailed bus automatic vehicle location (AVL) data. The data were from Minneapolis-Saint Paul Metro Transit Bus Route 5, where TSP was implemented at 30 signalized intersections in early 2019. A surrogate safety measure, bus speed volatility (BSV), was used to estimate TSP’s safety effects, with a higher BSV indicating more safety risks. A regression analysis was carried out on 23,123 event-level observations, with event defined as a bus traversal of a TSP-equipped intersection. Results indicate that with a TSP request, BSV was significantly lower than without a TSP request, confirming TSP’s effectiveness in smoothing bus trips through intersections, thus reducing risks of bus collisions and passenger fall-overs.

This research focuses on exploring features in a microscopic traffic simulation software to model and assess the influence of light rail transit (LRT) system before it is built, during the testing phase, and after it is in operation on traffic operational performance measures at intersections. A 2.5-mi urban arterial road in the city of Charlotte, North Carolina, was chosen for modeling and assessment in this research. The selected corridor, along N. Tryon St. from Mallard Creek Church Rd. to University City Blvd., has eight signalized intersections. Vissim traffic simulation software was used to develop simulation models and generate traffic operational performance measures such as vehicle delay, queue length, and level of service (LOS) and examine the influence at selected intersections on the study corridor. The results indicate there is an increase in vehicle delay along minor streets with a decrease in vehicle delay along the major streets parallel to the LRT route after it is in operation.

The growing number of disabled and senior citizens in the United States encourages cities to provide this population with paratransit services. This study explores the travel patterns of a paratransit service while considering the age-based trip characteristics of riders. We utilize a paratransit trip data set specialized for seniors and people with disabilities in the City of Arlington, TX. While a small number of eligible users take most trips, the ridership is significantly higher in adults under 65 years of age. Younger users mostly ride paratransit service for work/school trips, while older adults have a higher tendency for medical and recreational trips. Older adults tend to have fewer and shorter trips with a different temporal pattern than younger users. The results suggest that the availability of alternative modes negatively influences paratransit usage. This study could also be useful for future AV planners seeking to predict individuals’ travel patterns of the proposed AV service.

The COVID-19 outbreak has brought up challenges in numerous aspects of everyday life that require revisiting both the concepts and policies that are supposed to help adjust to the new situation. In this sense, public transit systems have undergone many changes while adapting to the new situation. Stay at home orders or avoid using public transportation was encouraged to help avoid the disease’s spread. It is important to note that pandemic does not treat everyone in the same way, and preventive orders may not apply to the transit-dependent population. In Jefferson County, Kentucky, the public transit system (bus) made bus routes and bus routes after the pandemic. The question is who is impacted by the pandemic and the limitations of using public transportation as it is the convenient mode of transportation for many transit-dependent populations. This study assessed the bus ridership before the pandemic, during the orders, and after the lockdown to help understand bus ridership variation across 580 TAZs in Jefferson County, Kentucky. It will also examine the route coverage in Jefferson County at the TAZ level to examine the areas that will low route coverage after the adjustments and its relation to the income.

Automated mobility technology is beginning to emerge as a viable means to create sustainable and effective public mobility systems within denser urban environments. Automated mobility districts (AMDs) describe major urban districts or activity centers in which deployments of multiple automated vehicle (AV) transit and ride-hailing fleets are supported to meet public mobility needs. The authors put forward a framework to enable AMDs and their governing and management jurisdictional authorities to manage safety of AV operations based on lessons learned from the last century of automated guideway transit and roadway intersection traffic control systems. The essential concept is that of operational management and safety-critical control of multiple AV fleets using a “system-of-systems” approach to system safety analysis. The safety analysis would focus on safe passage of the AV fleet vehicles through complex roadway intersections and junctions, especially in the presence of other non-automated modes such as pedestrians and manually operated vehicles.