In-depth analysis of drivers’ merging behavior and rear-end crash risks in work zone merging areas

This study investigates the drivers’ merging behavior and the rear-end crash risk in work zone merging areas during the entire merging implementation period from the time of starting a merging maneuver to that of completing the maneuver. With the merging traffic data from a work zone site in Singapore, a mixed probit model is developed to describe the merging behavior, and two surrogate safety measures including the time to collision (TTC) and deceleration rate to avoid the crash (DRAC) are adopted to compute the rear-end crash risk between the merging vehicle and its neighboring vehicles. Results show that the merging vehicle has a bigger probability of completing a merging maneuver quickly under one of the following situations: (i) the merging vehicle moves relatively fast; (ii) the merging lead vehicle is a heavy vehicle; and (iii) there is a sizable gap in the adjacent through lane. Results indicate that the rear-end crash risk does not monotonically increase as the merging vehicle speed increases. The merging vehicle's rear-end crash risk is also affected by the vehicle type. There is a biggest increment of rear-end crash risk if the merging lead vehicle belongs to a heavy vehicle. Although the reduced remaining distance to work zone could urge the merging vehicle to complete a merging maneuver quickly, it might lead to an increased rear-end crash risk. Interestingly, it is found that the rear-end crash risk could be generally increased over the elapsed time after the merging maneuver being triggered.     

Analysis of work zone rear-end crash risk for different vehicle-following patterns

This study evaluates rear-end crash risk associated with work zone operations for four different vehicle-following patterns: car–car, car–truck, truck–car and truck–truck. The deceleration rate to avoid the crash (DRAC) is adopted to measure work zone rear-end crash risk. Results show that the car–truck following pattern has the largest rear-end crash risk, followed by truck–truck, truck–car and car–car patterns. This implies that it is more likely for a car which is following a truck to be involved in a rear-end crash accident. The statistical test results further confirm that rear-end crash risk is statistically different between any two of the four patterns. We therefore develop a rear-end crash risk model for each vehicle-following pattern in order to examine the relationship between rear-end crash risk and its influencing factors, including lane position, the heavy vehicle percentage, lane traffic flow and work intensity which can be characterized by the number of lane reductions, the number of workers and the amount of equipment at the work zone site. The model results show that, for each pattern, there will be a greater rear-end crash risk in the following situations: (i) heavy work intensity; (ii) the lane adjacent to work zone; (iii) a higher proportion of heavy vehicles and (iv) greater traffic flow. However, the effects of these factors on rear-end crash risk are found to vary according to the vehicle-following patterns. Compared with the car–car pattern, lane position has less effect on rear-end crash risk in the car–truck pattern. The effect of work intensity on rear-end crash risk is also reduced in the truck–car pattern.     

Estimation of rear-end crash potential using vehicle trajectory data

Recent advancement in traffic surveillance systems has allowed for obtaining more detailed vehicular movement such as individual vehicle trajectory data. Understanding the characteristics of interactions between leading vehicle and following in the traffic flow stream is a backbone for designing and evaluating more sophisticated traffic and vehicle control strategies. This study proposes a methodology for estimating rear-end crash potential, as a probabilistic measure, in real time based on the analysis of vehicular movements. The methodology presented in this study consists of two components. The first estimates the probability that a vehicle's trajectory belonging to either ‘changing lane’ or ‘going straight’. A binary logistic regression (BLR) is used to model the lane-changing decision of the subject vehicle. The other component derives crash probability by an exponential decay function using time-to-collision (TTC) between the subject vehicle and the front vehicle. Also, an aggregated measure, crash risk index (CRI) is used in the analysis to accumulate rear-end crash potential for each subject vehicle. The result of this study can be used in developing traffic control and information systems, in particular, for crash prevention.     

A probabilistic quantitative risk assessment model for the long-term work zone crashes

Work zones especially long-term work zones increase traffic conflicts and cause safety problems. Proper casualty risk assessment for a work zone is of importance for both traffic safety engineers and travelers. This paper develops a novel probabilistic quantitative risk assessment (QRA) model to evaluate the casualty risk combining frequency and consequence of all accident scenarios triggered by long-term work zone crashes. The casualty risk is measured by the individual risk and societal risk. The individual risk can be interpreted as the frequency of a driver/passenger being killed or injured, and the societal risk describes the relation between frequency and the number of casualties. The proposed probabilistic QRA model consists of the estimation of work zone crash frequency, an event tree and consequence estimation models. There are seven intermediate events – age (A), crash unit (CU), vehicle type (VT), alcohol (AL), light condition (LC), crash type (CT) and severity (S) – in the event tree. Since the estimated value of probability for some intermediate event may have large uncertainty, the uncertainty can thus be characterized by a random variable. The consequence estimation model takes into account the combination effects of speed and emergency medical service response time (ERT) on the consequence of work zone crash. Finally, a numerical example based on the Southeast Michigan work zone crash data is carried out. The numerical results show that there will be a 62% decrease of individual fatality risk and 44% reduction of individual injury risk if the mean travel speed is slowed down by 20%. In addition, there will be a 5% reduction of individual fatality risk and 0.05% reduction of individual injury risk if ERT is reduced by 20%. In other words, slowing down speed is more effective than reducing ERT in the casualty risk mitigation.     

Development of crash-severity-index models for the measurement of work zone risk levels

Highway work zones interrupt regular traffic flows and create safety problems. Improving safety without sacrificing the main function of highways is a challenging task that traffic engineers and researchers have to confront. In this study, the concept of using crash severity index (CSI) for work zone safety evaluation was proposed and a set of CSI models were developed through the modeling of work zone crash severity outcomes. A CSI is a numerical value between zero and one that is estimated from given work zone variables. It is interpreted as the likelihood of having fatality/fatalities when a severe crash occurs in a given work zone. The CSI models were developed using a three-step approach. First, a wide range of crash variables were examined in a comprehensive manner and the significant risk factors that had impact on crash severity were selected. Second, the CSI models were developed using logistic regression technique by incorporating the selected risk factors. Finally, the developed models were validated using the recent crash data and their ability in assessing work zone risk levels were analyzed. Results of this study showed that CSI models can provide straightforward measurements of work zone risk levels.     

Effects of work zone presence on injury and non-injury crashes.

Work zones in the United States have approximately 700 traffic-related fatalities, 24,000 injury crashes, and 52,000 non-injury crashes every year. Due to future highway reconstruction needs, work zones are likely to increase in number, duration, and length. This study focuses on analyzing the effect of work zone duration mainly due to its policy-sensitivity. To do so, we created a unique dataset of California freeway work zones that included crash data (crash frequency and injury severity), road inventory data (average daily traffic (ADT) and urban/rural character), and work zone related data (duration, length, and location). Then, we investigated crash rates and crash frequencies in the pre-work zone and during-work zone periods. For the freeway work zones investigated in this study, the total crash rate in the during-work zone period was 21.5% higher (0.79 crashes per million vehicle kilometer (MVKM)) than the pre-work zone period (0.65 crashes per MVKM). Compared with the pre-work zone period, the increase in non-injury and injury crash rates in the during-work zone period was 23.8% and 17.3%, respectively. Next, crash frequencies were investigated using negative binomial models, which showed that frequencies increased with increasing work zone duration, length, and average daily traffic. The important finding is that after controlling for various factors, longer work zone duration significantly increases both injury and non-injury crash frequencies. The implications of the study findings are discussed in the paper.     

A multinomial logit model-Bayesian network hybrid approach for driver injury severity analyses in rear-end crashes

Rear-end crash is one of the most common types of traffic crashes in the U.S. A good understanding of its characteristics and contributing factors is of practical importance. Previously, both multinomial Logit models and Bayesian network methods have been used in crash modeling and analysis, respectively, although each of them has its own application restrictions and limitations. In this study, a hybrid approach is developed to combine multinomial logit models and Bayesian network methods for comprehensively analyzing driver injury severities in rear-end crashes based on state-wide crash data collected in New Mexico from 2010 to 2011. A multinomial logit model is developed to investigate and identify significant contributing factors for rear-end crash driver injury severities classified into three categories: no injury, injury, and fatality. Then, the identified significant factors are utilized to establish a Bayesian network to explicitly formulate statistical associations between injury severity outcomes and explanatory attributes, including driver behavior, demographic features, vehicle factors, geometric and environmental characteristics, etc. The test results demonstrate that the proposed hybrid approach performs reasonably well. The Bayesian network reference analyses indicate that the factors including truck-involvement, inferior lighting conditions, windy weather conditions, the number of vehicles involved, etc. could significantly increase driver injury severities in rear-end crashes. The developed methodology and estimation results provide insights for developing effective countermeasures to reduce rear-end crash injury severities and improve traffic system safety performance.     

Analysis of driver casualty risk for different work zone types

Using driver casualty data from the Fatality Analysis Report System, this study examines driver casualty risk and investigates the risk contributing factors in the construction, maintenance and utility work zones. The multiple t-tests results show that the driver casualty risk is statistically different depending on the work zone type. Moreover, construction work zones have the largest driver casualty risk, followed by maintenance and utility work zones. Three separate logistic regression models are developed to predict driver casualty risk for the three work zone types because of their unique features. Finally, the effects of risk factors on driver casualty risk for each work zone type are examined and compared. For all three work zone types, five significant risk factors including road alignment, truck involvement, most harmful event, vehicle age and notification time are associated with increased driver casualty risk while traffic control devices and restraint use are associated with reduced driver casualty risk. However, one finding is that three risk factors (light condition, gender and day of week) exhibit opposing effects on the driver casualty risk in different types of work zones. This may largely be due to different work zone features and driver behavior in different types of work zones.     

Modeling work zone crash frequency by quantifying measurement errors in work zone length

Work zones are temporary traffic control zones that can potentially cause safety problems. Maintaining safety, while implementing necessary changes on roadways, is an important challenge traffic engineers and researchers have to confront. In this study, the risk factors in work zone safety evaluation were identified through the estimation of a crash frequency (CF) model. Measurement errors in explanatory variables of a CF model can lead to unreliable estimates of certain parameters. Among these, work zone length raises a major concern in this analysis because it may change as the construction schedule progresses generally without being properly documented. This paper proposes an improved modeling and estimation approach that involves the use of a measurement error (ME) model integrated with the traditional negative binomial (NB) model. The proposed approach was compared with the traditional NB approach. Both models were estimated using a large dataset that consists of 60 work zones in New Jersey. Results showed that the proposed improved approach outperformed the traditional approach in terms of goodness-of-fit statistics. Moreover it is shown that the use of the traditional NB approach in this context can lead to the overestimation of the effect of work zone length on the crash occurrence.     

Characteristics of rear-end accidents at signalized intersections using multiple logistic regression model

Multi-vehicle rear-end accidents constitute a substantial portion of the accidents occurring at signalized intersections. To examine the accident characteristics, this study utilized the 2001 Florida traffic accident data to investigate the accident propensity for different vehicle roles (striking or struck) that are involved in the accidents and identify the significant risk factors related to the traffic environment, the driver characteristics, and the vehicle types. The Quasi-induced exposure concept and the multiple logistic regression technique are used to perform this analysis. The results showed that seven road environment factors (number of lanes, divided/undivided highway, accident time, road surface condition, highway character, urban/rural, and speed limit), five factors related to striking role (vehicle type, driver age, alcohol/drug use, driver residence, and gender), and four factors related to struck role (vehicle type, driver age, driver residence, and gender) are significantly associated with the risk of rear-end accidents. Furthermore, the logistic regression technique confirmed several significant interaction effects between those risk factors.     

Validating a driving simulator using surrogate safety measures

Traffic crash statistics and previous research have shown an increased risk of traffic crashes at signalized intersections. How to diagnose safety problems and develop effective countermeasures to reduce crash rate at intersections is a key task for traffic engineers and researchers. This study aims at investigating whether the driving simulator can be used as a valid tool to assess traffic safety at signalized intersections. In support of the research objective, this simulator validity study was conducted from two perspectives, a traffic parameter (speed) and a safety parameter (crash history). A signalized intersection with as many important features (including roadway geometries, traffic control devices, intersection surroundings, and buildings) was replicated into a high-fidelity driving simulator. A driving simulator experiment with eight scenarios at the intersection were conducted to determine if the subjects' speed behavior and traffic risk patterns in the driving simulator were similar to what were found at the real intersection. The experiment results showed that speed data observed from the field and in the simulator experiment both follow normal distributions and have equal means for each intersection approach, which validated the driving simulator in absolute terms. Furthermore, this study used an innovative approach of using surrogate safety measures from the simulator to contrast with the crash analysis for the field data. The simulator experiment results indicated that compared to the right-turn lane with the low rear-end crash history record (2 crashes), subjects showed a series of more risky behaviors at the right-turn lane with the high rear-end crash history record (16 crashes), including higher deceleration rate (1.80+/-1.20 m/s(2) versus 0.80+/-0.65 m/s(2)), higher non-stop right-turn rate on red (81.67% versus 57.63%), higher right-turn speed as stop line (18.38+/-8.90 km/h versus 14.68+/-6.04 km/h), shorter following distance (30.19+/-13.43 m versus 35.58+/-13.41 m), and higher rear-end probability (9/59=0.153 versus 2/60=0.033). Therefore, the relative validity of driving simulator was well established for the traffic safety studies at signalized intersections.     

Generalized nonlinear models for rear-end crash risk analysis

A generalized nonlinear model (GNM)-based approach for modeling highway rear-end crash risk is formulated using Washington State traffic safety data. Previous studies majorly focused on causal factor identification and crash risk modeling using Generalized linear Models (GLMs), such as Poisson regression, Logistic regression, etc. However, their basic assumption of a generalized linear relationship between the dependent variable (for example, crash rate) and independent variables (for example, contribute factors to crashes) established via a link function can be often violated in reality. Consequently, the GLM-based modeling results could provide biased findings and conclusions. In this research, a GNM-based approach is developed to utilize a nonlinear regression function to better elaborate non-monotonic relationships between the independent and dependent variables using the rear end accident data collected from 10 highway routes from 2002 through 2006. The results show for example that truck percentage and grade have a parabolic impact: they increase crash risks initially, but decrease them after the certain thresholds. Such non-monotonic relationships cannot be captured by regular GLMs which further demonstrate the flexibility of GNM-based approaches in the nonlinear relationship among data and providing more reasonable explanations. The superior GNM-based model interpretations help better understand the parabolic impacts of some specific contributing factors for selecting and evaluating rear-end crash safety improvement plans.     

Driving speed and the risk of road crashes: a review

Driving speed is an important factor in road safety. Speed not only affects the severity of a crash, but is also related to the risk of being involved in a crash. This paper discusses the most important empirical studies into speed and crash rate with an emphasis on the more recent studies. The majority of these studies looked at absolute speed, either at individual vehicle level or at road section level. Respectively, they found evidence for an exponential function and a power function between speed and crash rate. Both types of studies found evidence that crash rate increases faster with an increase in speed on minor roads than on major roads. At a more detailed level, lane width, junction density, and traffic flow were found to interact with the speed-crash rate relation. Other studies looked at speed dispersion and found evidence that this is also an important factor in determining crash rate. Larger differences in speed between vehicles are related to a higher crash rate. Without exception, a vehicle that moved (much) faster than other traffic around it, had a higher crash rate. With regard to the rate of a (much) slower moving vehicle, the evidence is inconclusive.     

Investigation of work zone crash casualty patterns using association rules

Investigation of the casualty crash characteristics and contributory factors is one of the high-priority issues in traffic safety analysis. In this paper, we propose a method based on association rules to analyze the characteristics and contributory factors of work zone crash casualties. A case study is conducted using the Michigan M-94/I-94/I-94BL/I-94BR work zone crash data from 2004 to 2008. The obtained association rules are divided into two parts including rules with high-lift, and rules with high-support for the further analysis. The results show that almost all the high-lift rules contain either environmental or occupant characteristics. The majority of association rules are centered on specific characteristics, such as drinking driving, the highway with more than 4 lanes, speed-limit over 40 mph and not use of traffic control devices. It should be pointed out that some stronger associated rules were found in the high-support part. With the network visualization, the association rule method can provide more understandable results for investigating the patterns of work zone crash casualties.     

Modeling rear-end collisions including the role of driver's visibility and light truck vehicles using a nested logit structure

This paper presents an analysis of the effect of the geometric incompatibility of light truck vehicles (LTV)--light-duty trucks, vans, and sport utility vehicles--on drivers' visibility of other passenger cars involved in rear-end collisions. The geometric incompatibility arises from the fact that most LTVs ride higher and are wider than regular passenger cars. The objective of this paper is to explore the effect of the lead vehicle's size on the rear-end crash configuration. Four rear-end crash configurations are defined based on the type of the two involved vehicles (lead and following vehicles). Nested logit models were calibrated to estimate the probabilities of the four rear-end crash configurations as a function of driver's age, gender, vehicle type, vehicle maneuver, light conditions, driver's visibility and speed. Results showed that driver's visibility and inattention in the following (striker) vehicle have the largest effect on being involved in a rear-end collision of configuration CarTrk (a regular passenger car striking an LTV). Possibly, indicating a sight distance problem. A driver of a smaller car following an LTV, have a problem seeing the roadway beyond the LTV, and therefore would not be able to adjust his/her speed accordingly, increasing the probability of a rear-end collision. Also, the probability of a CarTrk rear-end crash increases in the case that the lead vehicle stops suddenly.     

Driver responses to differing urban work zone configurations

This study reports the results of a simulator-based assessment of driver response to two different urban highway work zone configurations. One configuration represented an existing design which was contrasted with a second configuration that presented a reduced taper length prototype work zone design. Twenty-one drivers navigated the two different work zones in two different conditions, one with and one without a lead vehicle; in this case a bus. Measures of driver speed, braking, travel path, and collision frequency were recorded. Drivers navigated significantly closer to the boundary of the work area in the reduced taper length design. This proximity effect was moderated by the significant interaction between lead vehicle and taper length and such interactive effects were also observed for driver speed at the end of the work zone and the number of collisions observed within the work zone itself. These results suggest that reduced taper length poses an increase in risk to both drivers and work zone personnel, primarily when driver anticipation is reduced by foreshortened viewing distances. Increase in such risk is to a degree offset by the reduction of overall exposure to the work zone that a foreshortened taper creates. The benefits and limitations to a simulation-based approach to the assessment and prediction of driver behavior in different work zone configurations are also discussed.     

Estimation of rear-end vehicle crash frequencies in urban road tunnels

According to The Handbook of Tunnel Fire Safety, over 90% (55 out of 61 cases) of fires in road tunnels are caused by vehicle crashes (especially rear-end crashes). It is thus important to develop a proper methodology that is able to estimate the rear-end vehicle crash frequency in road tunnels. In this paper, we first analyze the time to collision (TTC) data collected from two road tunnels of Singapore and conclude that Inverse Gaussian distribution is the best-fitted distribution to the TTC data. An Inverse Gaussian regression model is hence used to establish the relationship between the TTC and its contributing factors. We then proceed to introduce a new concept of exposure to traffic conflicts as the mean sojourn time in a given time period that vehicles are exposed to dangerous scenarios, namely, the TTC is lower than a predetermined threshold value. We further establish the relationship between the proposed exposure to traffic conflicts and crash count by using negative binomial regression models. Based on the limited data samples used in this study, the negative binomial regression models perform well although a further study using more data is needed.     

Predicting crash likelihood and severity on freeways with real-time loop detector data

Real-time crash risk prediction using traffic data collected from loop detector stations is useful in dynamic safety management systems aimed at improving traffic safety through application of proactive safety countermeasures. The major drawback of most of the existing studies is that they focus on the crash risk without consideration of crash severity. This paper presents an effort to develop a model that predicts the crash likelihood at different levels of severity with a particular focus on severe crashes. The crash data and traffic data used in this study were collected on the I-880 freeway in California, United States. This study considers three levels of crash severity: fatal/incapacitating injury crashes (KA), non-incapacitating/possible injury crashes (BC), and property-damage-only crashes (PDO). The sequential logit model was used to link the likelihood of crash occurrences at different severity levels to various traffic flow characteristics derived from detector data. The elasticity analysis was conducted to evaluate the effect of the traffic flow variables on the likelihood of crash and its severity.The results show that the traffic flow characteristics contributing to crash likelihood were quite different at different levels of severity. The PDO crashes were more likely to occur under congested traffic flow conditions with highly variable speed and frequent lane changes, while the KA and BC crashes were more likely to occur under less congested traffic flow conditions. High speed, coupled with a large speed difference between adjacent lanes under uncongested traffic conditions, was found to increase the likelihood of severe crashes (KA). This study applied the 20-fold cross-validation method to estimate the prediction performance of the developed models. The validation results show that the model's crash prediction performance at each severity level was satisfactory. The findings of this study can be used to predict the probabilities of crash at different severity levels, which is valuable knowledge in the pursuit of reducing the risk of severe crashes through the use of dynamic safety management systems on freeways.     

Determining the drivers' acceptance of EFTCD in highway work zones

Traffic safety is a major concern in the temporary one-lane, two-way highway work zones due to the increasing of construction and maintenance operations. To prevent rear-end crashes and to mitigate the severity of these crashes caused by the inattentive driving, the utilization of the Emergency Flasher Traffic Control Device (EFTCD) was under consideration by government agencies, in addition to existing temporary traffic control devices installed in the one-lane, two-way highway work zones. The EFTCD was a newly proposed traffic warning device implemented through the use of vehicles’ hazard warning flashers. The primary objective of the research project was to investigate the drivers’ acceptance of the proposed EFTCD by measuring the mean speed changes of vehicles with and without EFTCD and by evaluating the drivers’ opinions of the EFTCD using the survey method. Field experimental results revealed that the EFTCD effectively reduced the mean vehicle speeds in the upstream of two work zones. A slow speed is more likely to reduce the severity of a crash in work zones. In addition, survey results indicated that 60% of the drivers thought the EFTCD signified a need for speed reduction and 82% of drivers recommended the implementation of the EFTCD in one-lane, two-way work zones. These results provide the necessary scientific justifications for the government agencies to decide if the EFTCD should be implemented in the one-lane, two-way highway work zones to prevent rear-end crashes and to mitigate the severity of these crashes.     

Impact of road environment on drivers’ behaviors in dilemma zone: Application of agent-based simulation

At a signalized intersection, there exists an area where drivers become indecisive as to either stop their car or proceed through when the traffic signal turns yellow. This point, called a dilemma zone, has remained a safety concern for drivers due to the great possibility of a rear-end or right-angle crash occurring. In order to reduce the risk of car crashes at the dilemma zone, Institute of Transportation Engineers (ITE) recommended a dilemma zone model. The model, however, fails to provide precise calculations on the decision of drivers because it disregards the supplemental roadway information, such as whether a red light camera is present. Hence, the goal of this study was to incorporate such roadway environmental factors into a more realistic driver decision-making model for the dilemma zone. A driving simulator was used to determine the influence of roadway conditions on decision-making of real drivers. Following data collection, each driver's decision outcomes were implemented in an Agent-Based Simulation (ABS) so as to analyze behaviors under realistic road environments. The experimental results revealed that the proposed dilemma zone model was able to accurately predict the decisions of drivers. Specifically, the model confirmed the findings from the driving simulator study that the changes in the roadway environment reduced the number of red light violations at an intersection.