Driving Analytics: Comparison of Teen and Adult Naturalistic Car-following Patterns (Multiple Year Project)

Principal Investigator: Helen Loeb, PhD, Children’s Hospital of Philadelphia

Below is an executive summary of this line of research. Please note that each summary describes results and interpretation that may not be final. Final interpretation of results will be in the peer-reviewed literature.



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. However, many crashes and most near crashes are never reported. The Strategic Highway Research Program 2 (SHRP2) driving database contains extensive real world data on teen and adult driving. Large-scale naturalistic driving studies offer the opportunity to compute crash rates using a reliable methodology to capture crashes and driving exposure.


The broad, long term objective of this study is to inform the development of Advanced Driver Assistance Systems (ADAS) such as Forward Collision Warning (FCW) or Brake Assist (BA), Dynamic Brake Support (DBS) and Crash Imminent Brake (CIB) This study will address the driving and braking deficits of teens, the most vulnerable driving population (Loeb et al., 2015).


A subset from the SHRP2 database consisting of 355 crashes, 252 near crashes for novice teens (16-19 years, 549 drivers), and 180 crashes, 253 near crashes for experienced adults (35-54 years, 591 drivers) was used. Onboard instrumentation such as scene cameras, accelerometers, and GPS logged time series data at 10 Hz. Videos were reviewed for all events to identify rear-end crashes.  Dynamic variables such as acceleration and velocity were analyzed for crashes and near crashes. Number of crashes, crash rates, crash severity, impact velocity, and near crashes were compared between novice teens and experienced adults.


Video review of the SHRP2 crashes identified significantly more crashes among the teen group than the adult group. Four categories of crash were reviewed for this study. We observed that teens had 14 times more road departures than adults, 7 times more crashes at intersections. We found that adults experienced 5 times more crashes with pedestrians than teens and that teens experienced five times more crashes with large animals than adults (need further exploration for causes, such as time of day, type of road…). When using airbag deployment as a metrics, we found that teens experienced 5.52 airbag deployment (non rear-end related) per million miles, compared to 2.82 airbag deployment (non-rear end related) for adults.


SHRP2 is the largest naturalistic driving database ever undertaken. Our careful comparison of teen and adult crashes and near crashes has alled us to better understand the frequency and circumstances of these crashes. Further study of the actions teens and adults take or do not take when they are faced with an emergency situation need to be assessed. Our crash rates are much higher than what is typically reported in the literature, as they reliably control for crash type and driving exposure through the exact vehicle mileage. These results confirm previous findings that novice teens exhibit increased crash rates compared to experienced adults. We anticipate novel technology such as FCW and Automatic Braking could greatly benefit teens on the road and should be made available to that most vulnerable population.


Motor vehicle crash (MVC) rates for teen and adult drivers are traditionally based upon fatal crashes, police-reported crashes, and estimated miles driven (EMD). Yet, nearly 30 percent of all crashes–particularly those that do not result in injury or death–are not reported to police. Teen drivers are much more likely to crash than adults; according to the Insurance Institute for Highway Safety, in 2014 the risk of MVCs per miles driven was nearly three times greater among 16- to 19-year-olds than among any other age group.

Naturalistic driving studies offer a unique opportunity to measure crash rates inclusive of all crashes, not only those reported to police, and to provide an exact quantification of miles driven among study participants. In this study, investigators used the Strategic Highway Research Program 2 (SHRP2), a large-scale naturalistic driving database administered by the Virginia Tech Transportation Institute, to study motor vehicle crashes. Over 3,000 volunteer drivers had their vehicles fitted with cameras, radar, and other sensors to capture data as they drove.

For this project, investigators analyzed 353 crashes involving 549 novice 16- to 19-year-old drivers and 185 crashes involving 591 experienced 35- to 54-year-old drivers.

This analysis used logged SHRP2 data of 1,000 actual motor vehicle crashes and near crashes to understand in depth why teen drivers crash.

Scene videos were reviewed for all events to identify rear-end crashes, and dynamic variables, such as acceleration and velocity, were analyzed for rear-end events. The investigators compared rear-end crash rates, crash severity, and impact velocity between novice teen drivers and experienced adult drivers. The teen group crash rate was more than seven times higher than the adult group, with 12.8 rear-end crashes occurring per 1 million miles driven for novice teens compared to 1.8 rear-end crashes occurring per 1 million miles driven for experienced adult drivers.

Further analysis of the SHRP2 dataset is needed to determine why novice teens experience rear-end crashes at such a high rate as compared to experienced adult drivers. To our knowledge, this is the first study to compare rear-end crash rates between teens and adults using a large-scale naturalistic driving database. This new knowledge of drivers’ reactions in emergency situations will help manufacturers design Active Safety systems, such as Forward Collision Warnings or Brake Assist systems, such as Dynamic Brake Support and Crash Imminent Brake systems, to address the braking deficits of newly licensed teen drivers

Project Team Members

Aditya Belwadi, PhD, Children's Hospital of Philadelphia; Thomas Seacrist, MBE, Children's Hospital of Philadelphia (Y1, Y2)


Abhiti Prabahar, The University of Pennsylvania; Sam Chamberlain, Drexel University; James Megariotis, Drexel University (Y1, Y2); David McDonnell, Drexel University; Ayya Elzarka, University of Pennsylvania; Charles Gottlieb, Jack M. Barrack Hebrew Academy; Leora Haber, Princeton University; Abe Kashem, University of  Pennsylvania; Taryn McKinney, University of Pennsylvania;  Danielle Mendoza, Widener University; Abhiti Prabahar, University of Pennsylvania; Adin Solomon, Jack M. Barrack Hebrew Academy

IAB Mentors

Doug Longhitano, American Honda Motor Co., Inc.; Anthony Rossetto, FCA US LLC; MaryAnn Beebe, General Motors Holdings LLC; Dan Glaser, General Motors Holdings LLC; Melissa Miles, State Farm Mutual Automobile Insurance Company (Y1, Y2); Zine Ben Aoun, Fiat Chrysler Automobiles; MaryAnn Beebe, General Motors Holdings LLC; Julie Kleinert, General Motors Holding LLC; Mark Neal, General Motors Holding LLC; Uwe Meissner, Technical Advisor