Data on driver behavior and roadway features will improve safety for decades.
Naturalistic Driving Study (NDS)
Traffic crashes still result in tragic loss of life and serious injuries; in 2012 there were more than 33,000 fatalities and 2.2 million injuries in the United States. Driver behavior is a significant factor in more than 90 percent of these crashes. Key to reducing crashes, serious injuries, and fatalities is a deeper understanding of driver behavior and how drivers interact with and adapt to their vehicles (including vehicle information systems), traffic, roadway characteristics, traffic control devices, and other environmental features. Driver behavior research to date has been indirect; it has examined crashes and attempted to reconstruct the events that produced them. To progress further, detailed, direct observational data on driver behavior are needed.
Two robust and linked databases will provide invaluable insight into driving behaviors and how drivers react to their environment.
Naturalistic Driving Study Database
SHRP2 is carrying out the largest, most comprehensive naturalistic driving study (NDS) ever conducted to monitor how drivers interact with their vehicles and the highway environment. The study collects real-time data on more than 5.8 million trips from more than 3,100 participants, ages 16-80 in 6 study sites across the country. The data will include 33 million travel miles, more than 1.4 million driving hours, and will exceed more than 4 petabytes in size.
The NDS database contains continuous data from all trips taken by volunteers over one to two years. Volunteers’ vehicles are heavily instrumented and record vehicle location, forward radar, vehicle control positions, and many other data elements, including video of the forward roadway and of the driver’s face and hands. Crash investigations are conducted after certain crashes to gather more detailed data.
Roadway Information Database
The Roadway Information database (RID) is comprised of two broad sources of data: new roadway data gathered by automated data collection vehicles and existing data from agencies (state DOTs, MPOs, counties) in the six NDS site states (Florida, Indiana, New York, North Carolina, Pennsylvania, and Washington).
The new data approximates 12,500 centerline miles in and around the study sites and the data types being collected are highlighted in the box to the right.
Coverage for the existing data approximates 200,000 plus centerline miles and includes roadway inventory information, crash histories, traffic, weather, roadway improvements, work zones, safety laws, and enforcement campaigns. The RID will also provide a model for developing linked data sets for asset management purposes.
Data from both databases will be linked through a geographic information system (GIS) that will provide researchers with a uniquely powerful data source. The linking of the NDS database with the RID will make it possible to relate driver behavior to roadway characteristics. Correlating driving behaviors with particular highway design features will give traffic engineers critical information to design safer roads.
The knowledge gained from analyzing the NDS and RID data will support life-saving improvements in the development and deployment of new safety countermeasures such as:
- updated driver training programs that can demonstrate to young drivers appropriate behavior behind the wheel;
- improved vehicle design including connected vehicle technology that better accommodates vehicle information systems in a way that doesn’t distract the driver or vehicle on-board systems to detect and respond to driver distraction;
- updated design guides, associated practices, and infrastructure improvements related to intersections, curves, and other complex areas;
- public policy and enforcement, and many more possibilities yet to be discovered.
These foundational databases along with analysis support tools are being created so they can be used by public, private, and academic researchers to improve safety for the next 30 years. They will be available for broad use consistent with volunteer privacy protections and are expected to be definitive sources of information on driver behavior for decades to come.
In the Field
Field activities also performed in Florida, Indiana, New York, North Carolina, Pennsylvania and Washington during the research phase.