Strategic Highway Research Program 2 Research Articles

Lessons learned from naturalistic driving data processing in a secure data enclave: Preliminary discoveries from analyzing dash camera videos

This paper provides preliminary insights on the challenges of processing Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study (NDS) videos and data, particularly those with Personally Identifiable Information (PII). Insights and lessons learned are presented from a study designed to evaluate the effectiveness of High Visibility Crosswalks (HVCs). Over a one-month period, 15,379 videos were processed in the secure data enclave of Virginia Tech Transportation Institute (VTTI). As these videos are not available outside of the secure data enclave due to PII restrictions, researchers visiting the secure data enclave for the first time may face several challenges: navigating the software interface; identifying the video views and frames of interest; and identifying and extracting information of interest from the video views, etc. These challenges, the procedures followed to address them, and the process for identifying and classifying distracted driving behaviors are discussed. Lastly, hypothesis tests are conducted to investigate distracted driving behavior, with the results revealing that HVCs have the potential to make drivers more cautious in their proximity. The information presented in this paper is expected to aid researchers who intend to utilize SHRP2 NDS or similar videos for future research, to preemptively plan for the video processing phase.

Incorporating Driving Behavior Metrics Derived from Naturalistic Driving Data into Macroscopic Safety Modeling

This research leveraged datasets from the Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study (NDS) to explore the potential benefits of incorporating macroscopic measures derived from NDS into traditional safety modeling. Large datasets with traversals from more than 1,700 unique drivers were used to extract driving behavior on freeway segments. New sets of time series totaling over 1,600 h of driving were developed, including vehicle dynamics exclusively during car-following, while also tracking the spacing between the instrumented vehicle and the vehicle being followed. This paper focuses on the statistical modeling of crash frequency incorporating macroscopic metrics derived from the new time-series datasets as regards the mean, median, variance, and 85th percentile of vehicle spacing, vehicle speed, and traffic density. Results of this exploration indicate that an increase in the traffic density variance, an increase in the speed variance, and a decrease in the mean vehicle spacing had significant effects associated with increases in multi-vehicle crash frequencies. These results can be used to estimate the safety effect of countermeasures that may change speed, density, and or spacing, along with changes in annual average daily traffic (AADT).

Crash/Near-Crash Analysis of Naturalistic Driving Data Using Association Rule Mining

This study explores the associations between crash/near-crash (C/NC) events and roadway, driver-related, and environmental factors in naturalistic driving studies (NDS). We used the Naturalistic Engagement in Secondary Tasks (NEST) dataset, which is massive and detailed and contains 50 million miles of naturalistic driving data resulting from the Strategic Highway Research Program 2 (SHRP2). Association rule mining (ARM) is applied to extract the rules for frequently occurring events. The generated association rules are filtered by four metrics (support, confidence, lift, and conviction) and validated by the lift increase criterion. A three-step analysis is performed to obtain a comprehensive understanding of the rules of C/NC events. The 20 most frequent items are first selected to investigate their relationship with the C/NC events. Subsequently, the association rules are used to identify the factors contributing to C/NC events. Finally, correlations between contributing factors and different severities of crashes (I—most severe, II—police-reportable, III—minor crash, and IV—low-risk tire strike) are analyzed by ARM. The results demonstrate that C/NC events occur most frequently on straight and level road segments with no controlled intersections or traffic control devices when drivers are performing secondary tasks. Thus, the reasons for these crashes are carelessness and overconfidence. In addition, a median strip or barrier and a wider road can significantly reduce the frequency and severity of crash events. Moreover, gender, age, average annual mileage, and secondary tasks are highly correlated with the frequency and severity of C/NC events. Drivers with visual-spatial disabilities or crash records are more likely to be involved in the most severe crash events. Near-crash events occur more frequently at higher traffic density and on roads with traffic control devices and controlled intersections. These conditions may keep drivers alert, preventing crashes.

Construction of a Composite Pavement (Asphalt over Concrete) in Virginia

The U.S. Strategic Highway Research Program 2 (SHRP2) identified composite pavement as a “renewal solution” and the Virginia Department of Transportation (VDOT) received implementation funding to demonstrate its feasibility. In 2017, this funding was applied to support major rehabilitation of two westbound lanes of US 60 in Henrico County, Virginia, a project that essentially replaced 1.1 mi of deteriorated concrete pavement with a new composite system consisting of continuously reinforced concrete pavement (CRCP) overlaid with stone matrix asphalt (SMA). This paper documents the design and construction of this new composite pavement. It also includes results from through-the-thickness temperature monitoring and a summary of performance for the first 4 years after opening to traffic. Close oversight by researchers provided a review of challenges and associated lessons learned from the design and construction process. Current VDOT specifications and standards were sufficient for successful construction. The asphalt overlay appeared to reduce the temperature gradient within the CRCP and to moderate daily temperature swings. This insulating effect of asphalt is expected to lower curling and associated stresses in the concrete pavement, which may permit reduced concrete thickness requirements in designs and also enable longer service life. The cracks in CRCP were tight and were not expected to reflect through the SMA. Moreover, the SMA mixture was tested to be crack and rut resistant. Therefore, a crack- and rut-free long service life for this composite pavement is expected. A recent visual survey by researchers found no observable distresses after 4 years under traffic.

How are different sources of distraction associated with at-fault crashes among drivers of different age gender groups?

IntroductionDistracted driving has been well researched, however the comparison between different age-gender groups on the impact of distracted driving has not been explored. Most crash analysis research does not distinguish driver responsibility, so the role that distractions has in at-fault crashes is unknown. Without distinguishing at-fault crashes from all-cause crashes, distracted driving's detrimental effects could be underestimated. ObjectiveThis study aims to systematically assess the risk of at-fault crashes associated with different sources of distraction among six groups by driver age (Teens 16–19, Adults 20–64, Seniors 65+) and gender. MethodsCrashes where a study participant was deemed at fault were identified using human expert annotated variables from the Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study dataset. Generalized linear mixed models were performed to assess the adjusted odds ratios of 10 distraction types associated with the at-fault crashes while controlling for environmental factors. ResultsThe main findings are (1) The highest contributing distraction types in at-fault crashes were In-Cabin Objects, Mobile Device, External Scenes, and In-Vehicle Information Systems (IVIS) as indicated by their influence on multiple age-gender groups and the magnitude of odds ratios; (2) Teens and adults were more distraction-prone than seniors, although seniors had the greatest at-fault crash risks associated with In-Cabin Objects, Mobile Device, and IVIS; (3) Distractions impacted females and males similarly; (4) At-fault crashes were more likely to have the significant distraction types present than all-cause crashes. ConclusionThis study adds to the limited literature on at-fault crashes particularly as it explores the role of driver demographics and distracted driving. Analyzing the risks of distracted driving by age-gender group shows that specific activities can be riskier for a certain population. The effects of distractions may be overlooked without fault determination. Distractions by external scenes, in-vehicle technologies, and in-cabin objects should not be overlooked, in addition to mobile device use.

Evaluating Relationships between Perception-Reaction Times, Emergency Deceleration Rates, and Crash Outcomes using Naturalistic Driving Data

Perception-reaction time (PRT) and deceleration rate are two key components in geometric design of highways and streets. Combined with a design speed, they determine the minimum required stopping sight distance (SSD). Current American Association of Highway Transportation Officials (AASHTO) SSD guidance is based on 90th percentile PRT and 10th percentile deceleration rate values from experiments completed in the mid-1990s. These experiments lacked real-world distractions, and so forth. Thus, the values from these experiments may not be applicable in real-world scenarios. This research evaluated (1) differences in PRTs and deceleration rates between crash and near-crash events and (2) developed predictive models for PRT and deceleration rate that could be used for roadway design. This was accomplished using (1) genetic matching (with Rosenbaum’s sensitivity analysis) and (2) quantile regression. These methods were applied to the Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study (NDS) data. The analysis results indicated that there were differences in PRT and deceleration rates for crash and near-crash events. The specific estimates were that, on average, drivers involved in crash events took 0.487 s longer to react and decelerated at 0.018 g’s (0.58 ft/s2) slower than drivers in equivalent near-crashes. Prediction models were developed for use in roadway design. These models were used to develop tables comparing existing SSD design criteria with SSD criteria based on the results of the predictive models. These predicted values indicated that minimum design SSD values would increase by 10.5–129.2 ft, dependent on the design speed and SSD model used.

Robust Data-Driven Framework for Driver Behavior Profiling Using Supervised Machine Learning

Driver behavior profiling has been gaining increased attention due to its relevance in many applications. For instance, car insurance telematics and fleet management entities have been recently using smartphones’ embedded sensors, On-Board Diagnostics II (OBDII) units and other on-board IoT devices to collect data on vehicles’ behavior and evaluate the risk profile of drivers. In this context, this paper presents a robust data-driven framework for calculating drivers’ risk profile measured in terms of the additive inverse of the predicted risk probability. The Strategic Highway Research Program 2 (SHRP2) naturalistic driving study (NDS) dataset, which is the largest dataset of its kind to date, is utilized to build the risk prediction models. Crash and near-crash events are used to quantify riskiness whereas balanced baseline driving events (i.e., events captured during normal day to day driving episodes) are used to reflect total exposure or driving time per driver. Thirteen mutually exclusive behavioral risk predictors are identified, and the feature matrix is formulated. A sensitivity analysis is then performed to find the best number of balanced baseline events below which drivers are filtered out. Different machine learning models are selected, customized, and compared to achieve best risk prediction performance. Finally, the utilization of the proposed prediction model within an envisioned driver profiling cloud-based framework is briefly discussed.

Evasive behavior-based method for threat assessment in different scenarios: A novel framework for intelligent vehicle

The threat assessment process is a crucial part of intelligent vehicles (IVs) for evaluating the levels of criticality and taking possible measures to avoid the collision, especially for the collision avoidance systems (CAS). In this study, a novel threat assessment framework based on the driver’s evasive behavior, namely the CPIC, is proposed, which integrates the crash probability (CP) and inevitable crash (IC) state to be widely used by different CAS in different scenarios. In the first step of the CPIC framework, the detailed evasive driver behavior models (E-DBMs) in the form of probability density functions (PDFs) were introduced to generate more realistic collision-avoidance trajectories. Two techniques for sampling these trajectories, namely the Markov Chain Monte Carlo (MCMC) and adaptive Gaussian mixture framework (GMM) methods, were utilized to ensure the samples were from the area of high probability density in the E-DBMs. The CP value could be derived by considering multiple collision-avoidance trajectories. To confirm the IC state in step 2, the CPIC framework employed the driving limit-based approach for IC checking, which combined the CP value to double-check the unavoidable collision. A total of 82 critical events from the real-world naturalistic driving study, the Strategic Highway Research Program 2 (SHRP2), were extracted to verify the performance of the CPIC framework in different scenarios. Results show that the proposed method clearly revealed the risk levels when two vehicles were approaching, and 80 events were successfully identified as near crashes/crashes. Moreover, the real-time performance of the CPIC framework was also demonstrated. The findings indicate this CPIC framework could be used in practical applications of IVs in different scenarios.

Nonparametric Multivariate Adaptive Regression Splines Models for Investigating Lane-Changing Gap Acceptance Behavior Utilizing Strategic Highway Research Program 2 Naturalistic Driving Data

Gap acceptance is one of the crucial components of lane-changing analysis and an important parameter in microsimulation modeling. Drivers’ poor gap judgment, and failure to accept a necessary safety gap, make it one of the major causes of lane-changing crashes on roadways. Several studies have been conducted to investigate lane-changing gap acceptance behavior; however, very few studies examined the behavior in complex real-world situations, such as in naturalistic settings. This study examined lane-changing gap acceptance behavior from the big Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study (NDS) datasets using a nonparametric multivariate adaptive regression splines (MARS) approach to better understand the complex effects of different factors in gap acceptance behavior. The study developed a unique methodology to identify lane-changing events of the non-NDS-vehicles using the front-mounted radar data from NDS vehicles and extract necessary parameters for analyzing gap acceptance behavior. In addition, surrogate measures of safety, that is, time-to-collision (TTC), was utilized to understand the impact of lane-changing on the NDS following vehicle safety. Moreover, different distributions of gap acceptance were fitted to identify the trend of gap acceptance behavior. The results from the MARS model revealed that different factors including relative speed between lane-changing vehicle (LCV) and lead vehicle (LV)/following vehicle (FV), traffic conditions, acceleration of LCV and FV, and roadway geometric characteristics have significant effects on gap acceptance behavior. The results of this study have significant implications, which could be used in microsimulation model calibration and safety improvements in connected and autonomous vehicles (CAV).

Near crash characteristics among risky drivers using the SHRP2 naturalistic driving study

Problem: Previous research have focused extensively on crashes, however near crashes provide additional data on driver errors leading to critical events as well as evasive maneuvers employed to avoid crashes. The Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study contains extensive data on real world driving and offers a reliable methodology to study near crashes. The current study utilized the SHRP2 database to compare the rate and characteristics associated with near crashes among risky drivers. Methods: A subset from the SHRP2 database consisting of 4,818 near crashes for teen (16–19 yrs), young adult (20–24 yrs), adult (35–54 yrs), and older (70+ yrs) drivers was used. Near crashes were classified into seven incident types: rear-end, road departure, intersection, head-on, side-swipe, pedestrian/cyclist, and animal. Near crash rates, incident type, secondary tasks, and evasive maneuvers were compared across age groups. For rear-end near crashes, near crash severity, max deceleration, and time-to-collision at braking were compared across age. Results: Near crash rates significantly decreased with increasing age (p < 0.05). Young drivers exhibited greater rear-end (p < 0.05) and road departure (p < 0.05) near crashes compared to adult and older drivers. Intersection near crashes were the most common incident type among older drivers. Evasive maneuver type did not significantly vary across age groups. Near crashes exhibited a longer time-to-collision at braking (p < 0.01) compared to crashes. Summary: These data demonstrate increased total near crash rates among young drivers relative to adult and older drivers. Prevalence of specific near crash types also differed across age groups. Timely execution of evasive maneuvers was a distinguishing factor between crashes or near crashes. Practical Applications: These data can be used to develop more targeted driver training programs and help OEMs optimize ADAS to address the most common errors exhibited by risky drivers.

Predicting Buckled and Unbuckled Trips of Occasional Seat Belt Users in SHRP2 Data

Seat belts are the most effective injury prevention tool available in motor vehicles. However, researchers have a limited understanding of the factors underlying occasional seat belt use. Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study data provide researchers a way to develop a detailed understanding of occasional belt use behavior. The present study examined two predictors (trip distance and average trip speed) of buckled and unbuckled trips from occasional seat belt users using the SHRP2 data. The present study utilized both descriptive analysis (to confirm previous finding) and predictive analysis (to assess predictive power of the models) using trip distance and average trip speed. The results from a logistic regression model showed that unbuckled trips were associated with short trips and lower speed trips compared to buckled trips. Two Random Decision Forest models were applied to predict buckled/unbuckled trips of the occasional seat belt users, and the results showed that model performance was within a fair performance range. The models differentiated between buckled and unbuckled trips at an accuracy of 61-64%. Results suggest that the occasional seat belt users may decide their seat belt use prior to trips based on the distance/duration and average. The preliminary findings of the study shed light on strategies that can be used when designing countermeasures to promote seat belt use.

Kinematic risky driving behavior among younger and older drivers: Differences over time by age group and sex

Objective: This research examined the incidence rates of elevated gravitational force events (kinematic risky driving, KRD) among 16- to 17-year-old drivers compared to those of 18- to 20-year-old, 21- to 25-year-old, and 35- to 55-year-old drivers over a 12-month period.Methods: Data were sampled from the Strategic Highway Research Program 2 (SHRP2) naturalistic driving study that recruited a U.S. national sample of study participants. General linear mixed models (GLIMMIX) for recurrent events were used to estimate KRD incident rates for age cohorts in 3-month periods.Results: KRD incidence rates for 16- to 17-year-old drivers were higher than the rates for older drivers at each 3-month period. Analyses of individual differences for the 12-month period indicated that incidence rates for the 16- to 17-year-old group were 1.84 times higher than the rates for 18- to 20-year-old drivers, 2.86 higher than those for 21- to 25-year-old drivers, and 4.92 times higher than those for 35- to 55-year-old drivers. The incident rate for 16- to 17-year-old males was 1.9 times higher than that for same-aged females in the first 3 months and 2.3 times higher over 12 months. Over the study period, KRD rates of 16- to 17-year-old participants declined 24.5% among females and 18.0% among males.Conclusions: KRD rates were higher among younger relative to older, more experienced drivers and did not decline over time, consistent with a protracted period of risky driving behavior. The persistently higher KRD rate among young drivers suggests the need to enhance crash prevention approaches, such as feedback about abrupt maneuvering, to young drivers and their parents.

Validating the Adaptability of Travel Time Reliability Measurements using Probe Data

Travel time reliability (TTR) is considered a critical piece of information in highway performance evaluation. The L02 project from Strategic Highway Research Program 2 (SHRP2) has developed a holistic method using statistical probability functions of travel time as the TTR measure to build highway performance evaluation and monitoring systems. Compared with single-value reliability measures, the L02 measure is able to identify sources of unreliability and quantify their associated impacts. To validate the adaptability of L02 measure, TTR analysis on the I-15 freeway corridor in Salt Lake City, Utah using probe data has been conducted. The result is compared against output from the quadrant-based TTR measure that is currently used by the Utah Department of Transportation. Through cross-validation, it is determined that the two suites of measures demonstrate good consistency in relation to reliability assessment and unreliability source diagnoses. In addition, the study provides a method to calibrate the quadrant-based TTR measure, and new critical values were developed based on the cross-validation.

Driver’s Response Prediction Using Naturalistic Data Set

&lt;div class="section abstract"&gt;&lt;div class="htmlview paragraph"&gt;Evaluating the safety of Autonomous Vehicles (AV) is a challenging problem, especially in traffic conditions involving dynamic interactions. A thorough evaluation of the vehicle’s decisions at all possible critical scenarios is necessary for estimating and validating its safety. However, predicting the response of the vehicle to dynamic traffic conditions can be the first step in the complex problem of understanding vehicle’s behavior. This predicted response of the vehicle can be used in validating vehicle’s safety.&lt;/div&gt;&lt;div class="htmlview paragraph"&gt;In this paper, models based on Machine Learning were explored for predicting and classifying driver’s response. The Naturalistic Driving Study dataset (NDS), which is part of the Strategic Highway Research Program-2 (SHRP2) was used for training and validating these Machine Learning models. Various popular Machine Learning Algorithms were used for classifying and predicting driver’s response, such as Extremely Randomized Trees and Gaussian Mixture Model based Hidden Markov Model, which are widely used in multiple domains.&lt;/div&gt;&lt;div class="htmlview paragraph"&gt;For classifying driver’s response, longitudinal acceleration vs lateral acceleration plot (Ax-Ay plot) was divided into nine different classes and selected Machine Learning models were trained for predicting the class of driver’s response. Performances of models for classification were tabulated and it is observed that Extremely Randomized Trees based model had better prediction accuracies in comparison with other models when fit using SHRP2 NDS data. The input features were reduced using dimension reduction techniques to reduce the computation time by over 70%.&lt;/div&gt;&lt;/div&gt;

A preliminary investigation of the effectiveness of high visibility enforcement programs using naturalistic driving study data: A grouped random parameters approach

This paper seeks to assess the effectiveness of high-visibility enforcement (HVE) programs in terms of reducing aggressive driving behavior. Using Strategic Highway Research Program 2 (SHRP2) Naturalistic driving study (NDS) data, behavioral reactions of drivers before, during, and after the conduct of high-visibility enforcement programs are analyzed, in order to identify the potential effect of high-visibility enforcement in driving behavior. In this context, two fundamental aspects of aggressive driving behavior (speeding and tailgating) are employed and analyzed. To simultaneously explore the intensity and the duration of these behavioral patterns, novel metrics are defined and used in the analysis. To investigate the effect of high-visibility enforcement programs, and at the same time, to control for the effect of driver-, trip-, vehicle-, and weather-specific characteristics on the extent of speeding and tailgating, univariate grouped random parameters linear regression models are estimated. In addition, likelihoods of speeding and tailgating occurrences are analyzed simultaneously, within a grouped random parameters bivariate probit modeling framework. The results of this preliminary analysis show that even though the implementation of the high-visibility enforcement has mixed effects on the extent and the likelihood of the driving behavior metrics, it demonstrates a promising potential in modifying driving behavior.

Overestimation of self-reported driving exposure: Results from the SHRP2 Naturalistic Driving Study

Objectives: The accuracy of self-reported driving exposure has questioned the validity of using self-reported mileage to inform research questions. Studies examining the accuracy of self-reported driving exposure compared to objective measures find low validity, with drivers overestimating and underestimating driving distance. The aims of the current study were to (1) examine the discrepancy between self-reported annual mileage and driving exposure the following year and (2) investigate whether these differences depended on age and annual mileage.Methods: Two estimates of drivers’ self-reported annual mileage collected during vehicle installation (obtained via prestudy questionnaires) and approximated annual mileage driven (based upon Global Positioning System data) were acquired from 3,323 participants who participated in the Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study.Results: A Wilcoxon signed rank test showed that there was a significant difference between self-reported and annual driving exposure during participation in SHRP 2, with the majority of self-reported responses overestimating annual mileage the following year, irrespective of whether an ordinal or ratio variable was examined. Over 15% of participants provided self-reported responses with over 100% deviation, which were exclusive to participants underestimating annual mileage. Further, deviations in reporting differed between participants who had low, medium, and high exposure, as well as between participants in different age groups.Conclusions: These findings indicate that although self-reported annual mileage is heavily relied on for research, such estimates of driving distance may be an overestimate of current or future mileage and can influence the validity of prior research that has utilized estimates of driving exposure.

Investigation of Design Speed Characteristics on Freeway Ramps using SHRP2 Naturalistic Driving Data

Freeway ramp design guidance has existed in the United States for many decades, coinciding with the advent of the nation’s freeway network and the Interstate Highway system. Some principles associated with ramp design are largely unchanged since their inception, and a review of those principles in the context of today’s drivers and vehicles is beneficial for identifying potential updates to existing guidance. The process of collecting the necessary data may consist of a variety of methods, each with limitations on the number of ramps, vehicles, and trips that can be studied. A current research project is exploring the feasibility of using data from the Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study (NDS) to identify relationships between ramp design speed characteristics and drivers’ choices of operating speeds on those ramps. The NDS data provides a dataset that is unprecedented in its size and detail, but its suitability for this type of analysis is largely unknown. This paper summarizes the activities and findings of the current research project, including basic models for estimating vehicle speeds on freeway ramps based on the NDS data; these models may be used in conjunction with other ongoing related research efforts to suggest material for potential updates to existing ramp design guidance.

Analysis of near crashes among teen, young adult, and experienced adult drivers using the SHRP2 naturalistic driving study

ABSTRACTObjective: Motor vehicle crashes are the leading cause of death among young drivers. Though previous research has focused on crash events, near crashes offer additional data to help identify driver errors that could potentially lead to crashes as well as evasive maneuvers used to avoid them. The Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study (NDS) contains extensive data on real-world driving and offers a reliable methodology to quantify and study near crashes. This article presents findings on near crashes and how they compare to crash events among teen, young adult, and experienced adult drivers.Methods: A subset from the SHRP2 database consisting of 1,653 near crashes for teen (16–19 years, n = 550), young adult (20–24 years, n = 748), and experienced adult (35–54 years, n = 591) drivers was used. Onboard instrumentation including scene cameras, accelerometers, and Global Positioning System logged time series data at 10 Hz. Scene videos were reviewed for all events to classify near crashes based on 7 types: rear-end, road departure, intersection, head-on, side-swipe, pedestrian/cyclist, and animal. Near crash rates, incident type, secondary tasks, and evasive maneuvers were compared across age groups and between crashes and near crashes. For rear-end near crashes, vehicle dynamic variables including near crash severity, headway distance, time headway, and time to collision at the time of braking were compared across age groups. Crashes and near crashes were combined to compare the frequency of critical events across age.Results: Teen drivers exhibited a significantly higher (P <.01) near crash rate than young adult and experienced adult drivers. The near crash rates were 81.6, 56.6, and 37.3 near crashes per million miles for teens, young adults, and experienced adults, respectively. Teens were also involved in significantly more rear-end (P <.01), road departure (P <.01), side-swipe (P <.01), and animal (P <.05) near crashes compared to young and experienced adults. Teens exhibited a significantly greater (P <.01) critical event rate of 102.2 critical events per million miles compared to 72.4 and 40.0 critical events per million miles for young adults and experienced adults, respectively; the critical event rate ratio was 2.6 and 1.8 for teens and young adults, respectively.Conclusions: To our knowledge, this is the first study to examine near crashes among teen, young adult, and experienced adult drivers using SHRP2 naturalistic data. Near crash and critical event rates significantly decreased with increasing age and driver experience. Overall, teens were more than twice as likely to be involved in critical events compared to experienced adults. These data can be used to develop more targeted driver training programs and help manufacturers design active safety systems based on the most common driving errors for vulnerable road users.

Comparison of crash rates and rear-end striking crashes among novice teens and experienced adults using the SHRP2 Naturalistic Driving Study

ABSTRACTObjective: Motor vehicle crashes are the leading cause of death for teens. Previous teen and adult crash rates have been based upon fatal crashes, police-reported crashes, and estimated miles driven. Large-scale naturalistic driving studies offer the opportunity to compute crash rates using a reliable methodology to capture crashes and driving exposure. The Strategic Highway Research Program 2 (SHRP2) Naturalistic Driving Study contains extensive real-world data on teen and adult driving. This article presents findings on the crash rates of novice teen and experienced adult drivers in naturalistic crashes.Methods: A subset from the SHRP2 database consisting of 539 crash events for novice teens (16–19 years, n = 549) and experienced adults (35–54 years, n = 591) was used. Onboard instrumentation such as scene cameras, accelerometers, and Global Positioning System logged time series data at 10 Hz. Scene videos were reviewed for all events to identify rear-end striking crashes. Dynamic variables such as acceleration and velocity were analyzed for rear-end striking events. Number of crashes, crash rates, rear-end striking crash severity, and rear-end striking impact velocity were compared between novice teens and experienced adults.Results: Video review of the SHRP2 crashes identified significantly more crashes (P < 0.01) and rear-end striking crashes (P < 0.01) among the teen group than among the adult group. This yielded crash rates of 30.0 crashes per million miles driven for novice teens compared to 5.3 crashes per million miles driven for experienced adults. The crash rate ratio for teens vs. adults was 5.7. The rear-end striking crash rate was 13.5 and 1.8 per million miles driven for novice teens and experienced adults, respectively. The rear-end striking crash rate ratio for teens vs. adults was 7.5. The rear-end striking crash severity measured by the accelerometers was greater (P < 0.05) for the teen group (1.8 ± 0.9 g; median = 1.6 g) than for the adult group (1.1 ± 0.4 g; median = 1.0 g), suggesting that teen crashes tend to be more serious than adult crashes. Increased rear-end striking impact velocity (P < 0.01) was also observed for novice teens (18.8 ± 13.2 mph; median = 18.9 mph) compared to experienced adults (3.3 ± 1.2 mph; median = 2.8 mph).Conclusion: To our knowledge, this is the first study to compare crash rates between teens and adults using a large-scale naturalistic driving database. Unlike previous crash rates, the reported rates reliably control for crash type and driving exposure. These results conform to previous findings that novice teens exhibit increased crash rates compared to experienced adults.

Creation of the Naturalistic Engagement in Secondary Tasks (NEST) distracted driving dataset

Distracted driving has become a topic of critical importance to driving safety research over the past several decades. Naturalistic driving data offer a unique opportunity to study how drivers engage with secondary tasks in real-world driving; however, the complexities involved with identifying and coding relevant epochs of naturalistic data have limited its accessibility to the general research community. This project was developed to help address this problem by creating an accessible dataset of driver behavior and situational factors observed during distraction-related safety-critical events and baseline driving epochs, using the Strategic Highway Research Program 2 (SHRP2) naturalistic dataset. The new NEST (Naturalistic Engagement in Secondary Tasks) dataset was created using crashes and near-crashes from the SHRP2 dataset that were identified as including secondary task engagement as a potential contributing factor. Data coding included frame-by-frame video analysis of secondary task and hands-on-wheel activity, as well as summary event information. In addition, information about each secondary task engagement within the trip prior to the crash/near-crash was coded at a higher level. Data were also coded for four baseline epochs and trips per safety-critical event. 1,180 events and baseline epochs were coded, and a dataset was constructed. The project team is currently working to determine the most useful way to allow broad public access to the dataset. We anticipate that the NEST dataset will be extraordinarily useful in allowing qualified researchers access to timely, real-world data concerning how drivers interact with secondary tasks during safety-critical events and baseline driving. The coded dataset developed for this project will allow future researchers to have access to detailed data on driver secondary task engagement in the real world. It will be useful for standalone research, as well as for integration with additional SHRP2 data to enable the conduct of more complex research.