TOPICS IN THIS ARTICLE

Drivers were studied while driving their own cars, under normal traffic conditions

Similar to an airplane's "black box," video-enhanced electronic data recorders can provide data on human factors involved in auto crashes

Fatigue, distraction, and failure to pay attention ranked as the top three crash-causers


LINKS

The Virginia Tech Transportation Institute

Virginia Department of Transportation

Virginia Transportation Research Council


SUBJECTS OF OTHER ARTICLES
IN THIS ISSUE

Multicultural mindfulness

Infectious disease research

How kids deal with parents’ deployment

Jaguars

New ways to fight the war on cancer

Driverless vehicles

Hurricane-related research

Editorial: All giants welcome


 

 

 

Drivers everywhere everyday are multitasking while driving. They read mail, eat lunch, put on makeup, and talk on the phone. At the same time, car crashes result in property damage and injuries across the country. What causes all these crashes? Could the distractions of activities unrelated to driving have something to do with it? That is what researchers at the Virginia Tech Transportation Institute (VTTI) are working to find out from their 100-car naturalistic driving study.

With vehicle crashes as the leading cause of injury-related death in the United States for people between 1 and 65 years old, high-quality transportation research is essential to public safety. Researchers are working to provide information that would help prevent the more than 40,000 deaths and 2 million injuries, and the $150 billion cost of crashes each year.

Until recently, research has been confined to analysis of police-reported crash data and studies conducted on test tracks and in simulators. While these methods can be effective, there is no substitute for collecting data in a real-world, or naturalistic, environment.

So, for a year, researchers at VTTI observed the actual daily driving habits of 241 drivers. It was the first instrumented vehicle study undertaken with the primary purpose of collecting large-scale naturalistic driving data.

“Due to the unpredictability of driver performance and the random nature of automobile crashes, the collection of naturalistic data gives a more accurate perspective of why crashes occur,” says Tom Dingus, director of VTTI and program manager for the study.

A naturalistic study is one that takes place within a natural environment, providing researchers with a more realistic perspective of what causes a crash. “Our extensive naturalistic driving study became a reality thanks to advances in technology and data collection systems,” says Andy Peterson, director of the VTTI Center for Technology Development.

The National Highway Traffic Safety Administration (NHTSA) has long recognized the need for, and importance of, collecting naturalistic on-road driving data, especially for pre-crash and near-crash driving situations. In 2000, VTTI was awarded $3.7 million by NHTSA in partnership with the Intelligent Transportation Systems (ITS) Joint Program Office, acting through the Intelligent Vehicle Initiative, the Virginia Department of Transportation (VDOT), the Virginia Transportation Research Council (VTRC), and Virginia Tech to study driver performance and behavior leading up to crashes and near crashes.

Drivers were observed in their own vehicles in real traffic conditions

More than 100 leased and privately owned cars in the Northern Virginia/Washington, D.C., area were volunteered to be equipped with specialized instrumentation for one year. Using these vehicles in their normal daily routines, drivers were given no special instructions. No researchers were present and the data collection instrumentation was unobtrusive. One of the major advantages of naturalistic driving studies is the ability to observe drivers in their natural settings as they drive to and from work, to the grocery store, out to eat, and so on.

“This is the largest instrumented vehicle study ever attempted and will provide a wealth of new information to help understand, and eventually reduce, vehicle crashes,” says Dingus. “The goal is to save lives and this 100-car study is a first step. Our data sources up to this point were pretty limited. If you don’t have good information about why crashes occur and why fatalities occur, you can’t really solve the problem.”

“This project provided a unique opportunity to study drivers’ performance in their own vehicles in real traffic conditions,” says Vicki Neale, leader of the institute’s Center for Automotive Safety Research.

Gary Allen, chief of technology, research, and innovation at VTRC, was interested in co-sponsoring the work because of its potential contribution to VDOT’s overall mission of making driving safer in Virginia. “We see this study as a complement to our ongoing research in the area of advancing transportation safety,” says Allen.

There are two traditional approaches to collecting and analyzing human factors data related to driving. The first approach uses data gathered from large population studies (often collected on a national level). These databases, however, lack sufficient detail to be helpful for many applications, such as the development of countermeasure systems or the assessment of interactions between contributing factors that lead to crashes, says Dingus.

The second approach uses data gathered through controlled experiments, including newer, high-fidelity driving simulators and test tracks. However, these studies cannot avoid a certain level of artificiality and do not always capture the complexities of the driving environment or of natural behavior, says Dingus.

“Test subjects are often more alert and more cautious in a simulation environment or when an experimenter is present in a research vehicle than when they are driving alone in their own cars,” says Shelia Klauer, project manager for the study.

Although controlled experiments are useful in other contexts, they provide a limited picture of the likelihood of a crash in a given situation or the potential reduction of that likelihood by a given countermeasure. Controlled experiment approaches can only assess the relative safety of various countermeasures or scenarios. They cannot be used to predict the effect of a safety device or policy change on the crash rate.

Cars used in the study were equipped wtih video-enhanced electronic data recorders

Advances in sensor, data storage, and communications technology have led to the development of a hybrid approach to data collection and analysis that uses vehicle-based technology, such as electronic data recorders (EDRs) introduced into research several years ago.

EDRs collect a variety of vehicular dynamic and status data, but lack sufficient measurement capabilities to assess many human factors issues, says Neale. However, field tests that use video-enhanced EDRs in a naturalistic setting make clear the conditions and driver behaviors that precipitate crashes, as well as support the development and refinement of crash countermeasures.

The data acquisition system that was installed in the 100 cars is often compared to an airplane’s black box, but it’s actually much more advanced because it allows researchers to see what is taking place before and during a crash. This unique visual element of data provides another dimension for understanding how and why crashes occur.

The vehicle instrumentation gathered information through five channels of digital compressed video and many vehicle status and motion variables. The extensive data collection resulted in approximately 2 million vehicle miles of driving, more than 42,000 hours of data with 241 primary and secondary driver participants over a 12- to 13-month period for each vehicle, and an 18-month total data collection period. Results included:

• 15 police-reported and 82 total crashes and collisions (any contact between the subject vehicle and another vehicle, fixed object, pedestrian, cyclist, or animal)

• 761 near crashes (a rapid, severe evasive maneuver to avoid a crash)

• 8,295 incidents (an evasive maneuver of less magnitude than a near-crash)

"The number of crashes that occurred during the study surprised us, as did the fact that many went unreported," says Klauer.

“Driver error” was found to be a contributing factor in more than 90 percent of all vehicular crashes, while 80 percent of all crashes and 65 percent of all near-crashes involved the driver looking away from the forward roadway just prior to (within three seconds) the onset of the event, according to the study.

An “event” database was created with video and vehicle performance data included.

"The participants appeared to disregard the presence of vehicle instrumentation quickly,” says Neale. As a result, the event database contains many extreme cases of driving behavior and performance, including severe fatigue, impairment, judgment error, risk taking, secondary task engagement (multitasking), aggressive driving, and traffic law violation.

Inattention, distraction, and driver fatigue shown to be factors in crashes

The most significant and widely publicized result of the study regarded inattention to the forward roadway. The "secondary task distraction" that most frequently resulted in a crash or near crash was hand-held wireless devices, the researchers report.

However, the most dangerous contributing factor leading to crashes was fatigue. Fatigue was a factor in 12 percent of all crashes and 10 percent of all near-crashes, the study showed. Previous database estimates placed fatigue-related crashes at 2 to 4 percent of total crashes. The VTTI study revealed that drivers suffering from moderate to severe fatigue were 4.7 times more likely to be involved in a crash than an alert, attentive driver.

In addition to distraction and fatigue, inattention includes driving-related tasks and non-specific glances away from the road. Inattention to the forward roadway was found to be the primary contributing factor in most crashes and collisions. For instance, 93 percent of rear-end collisions involved the driver of the following vehicle looking away within three seconds of the crash.

The event database also allowed researchers to gather valuable information about general driving response skills. Data indicated that drivers generally have sufficient awareness and ability to perform evasive maneuvers when responding to typical traffic conditions. However, drivers were found to have difficulty responding appropriately when other vehicles performed unexpected or unanticipated maneuvers, such as suddenly stopping or changing lanes.

Researchers also found that age was clearly a factor in the likelihood of being involved in an inattention–related crash or near crash. Such events decreased dramatically with age, with the rate being as much as four times higher for 18- to 20-year-olds than older age groups. Drivers younger than 18 years old were not tested.

Data from crash and near-crash events from this study may provide additional insight into effective defensive driving techniques and factors, as well as potential countermeasures for these driving situations. It was found that near crashes occurred 15 times more often than crashes and every near-crash demonstrated a driver successfully performing an evasive maneuver.

"The data analysis results have greater application and accuracy relative to the big picture of driving than do experimental methods, such as test tracks or simulators," says Dingus.

Results from the study present a wide variety of information about driving habits, driving skills, and crash causation. As additional analyses are completed with the event database, researchers will be able to extract even more knowledge about crashes and driving habits in general.

"The results of the 100-car study are particularly significant because of the future safety implications," says Dingus. "By understanding the details of how and why crashes occur, researchers will be able to present findings to policy makers to help them make informed decisions about transportation safety, recommend improvements to manufacturers of safety features in automobiles, make drivers more aware of what causes crashes and near-crashes, and perhaps even provide recommendations for driver training."

Ultimately, sound transportation policy and safer vehicles combined with safer drivers will lead to fewer crash-related fatalities and injuries, says Dingus. "The findings of this study will provide a wealth of information, including insights to drivers of what causes accidents and hopefully lead to improving their overall driving ability."

VTTI researchers anticipate that more comprehensive studies will be funded to build on the results of the 100-car study. With additional research involving more drivers on various road types in a range of geographies, researchers will have the opportunity to delve even deeper into real-world driving behavior with the goal of improving transportation safety and saving lives.

— Chris Bagg, Virginia Tech Transportation Institute