Naturalistic driving study is a life-saving reality show
By Kayla Czech, Virginia Tech Transportation Institute
Every 13 minutes somebody on America's roadways dies in a vehicle-related crash. The Virginia Tech Transportation Institute (VTTI) is working to lower that statistic through naturalistic driving research involving teen drivers, truck drivers, older drivers, and motorcyclists.
Naturalistic driving studies, which use sophisticated cameras and instrumentation in participants' personal vehicles, provide researchers with thousands to millions of hours of data on actual driver behavior and performance. These studies could be described as a reality TV show filmed inside a vehicle with the intention of saving lives. Participants are given no instructions other than to go about their normal daily driving activities while they are monitored for anywhere from six months to three years. Data is collected through multiple channels of digital compressed video, radar sensors, machine vision-based lane trackers, cell phones, GPS, and instruments that record such vehicle information as braking, acceleration, and yaw.
The first large-scale instrumented-vehicle study undertaken by the transportation institute with the primary purpose of collecting pre-crash and near-crash naturalistic driving data began in 2003 and tracked the behaviors of drivers of 100 instrumented vehicles in the Washington, D.C., area for a year. Drivers' ages ranged from 18 to 73, with 60 percent of the participants being male. The study garnered 42,300 hours of data, 82 police-reported crashes, 761 near crashes, and 8,295 critical incidents.
The follow-up analysis of the 100-car study revealed that drivers who engage frequently in distracting activities are more likely to be involved in a crash or near crash. The rate of inattention-related crash and near-crash events was found to be four times higher for the 18-to-20-year-old age group compared to those over 35. Both the study and the follow-up analysis were co-sponsored by the National Highway Traffic Safety Administration (NHTSA), the Virginia Center for Transportation Innovation and Research (formerly the Virginia Transportation Research Council), and Virginia Tech.
What makes naturalistic driving studies superior to other methods? They're conducted in the real world. Drivers participating in traditional controlled experiments might alter their behavior to make themselves "look good on camera." But participants in naturalistic driving studies face real driving conditions and pressures, and make real decisions that have real consequences.
In addition, police accident reports must rely on information from those involved and from crash scenes from which the vehicles have often been moved. Drivers in crashes may be deceased or injured; pre-crash events occur so rapidly that the driver and/or passenger(s) often forget key elements; and witnesses frequently leave out important details. Data from naturalistic driving studies, however, provide great detail and accuracy regarding driver behavior, driver error, and vehicle performance in a format similar to an "instant replay" on a televised sporting event.
In August 2011, the institute launched the world's first large-scale naturalistic motorcycle riding study. The Motorcycle Safety Foundation's (MSF) 100 Motorcyclists Naturalistic Study is a partnership effort between the Virginia Tech Transportation Institute and MSF, the largest trainer of motorcycle riders in the world.
Each of the participant motorcycles is equipped with five color cameras, a GPS, accelerometers, a gyro, forward radar, a machine-vision lane tracker, and front and rear brake sensors. Participants were chosen based on their age and the model of motorcycle used. The study is tracking two age groups, 21 to 34 years and 45 to 64 years, and seven motorcycle models, including sport bikes, cruisers, and touring bikes. Three different locations are being used for outfitting, tracking, and data collection: VTTI in Blacksburg; MSF headquarters in Irvine, Calif.; and the Motorcycle Mechanics Institute in Orlando, Fla. Data will be collected until summer 2013. Researchers expect approximately 500,000 miles of riding data, which they will study for an additional year.
"We know of no other naturalistic study for motorcycles," says transportation institute director Tom Dingus. "We expect the study to be very valuable to the Motorcycle Safety Foundation's work, since we also will examine where and how crash avoidance is successful. With so much information bandwidth coming from the cameras and instrumentation on each bike, we'll be able to examine details for years, and the findings will be relevant for decades."
"This research offers much more valuable data than going to a crash site where the bike may have been removed," says Sherry Williams, MSF director of quality assurance and research. "With this research, we will get a very rich picture, where you can see the input from the rider and how the bike is reacting."
VTTI is also in the planning stages of a naturalistic motorcycle study with NHTSA in which a total of 160 motorcycles will be instrumented. According to NHTSA, while fatality rates for other road users have been decreasing, fatalities from motorcycle crashes increased 128 percent from 1997-2007. In 2010, AAA Mid-Atlantic reported 78 motorcyclist fatalities and 1,981 injuries in Virginia.
Motorcycle riding garnered increased national attention as it rose to the number-one noncombat safety concern across the U.S. armed services. In 2008, 126 service members were killed in noncombat motorcycle crashes, an increase from 93 fatalities in 2007.
"VTTI has a vested interest in motorcycle safety, and we see the rising number of motorcycle crashes as a national problem," Dingus says. "Through naturalistic research, we can collect data that will provide guidance to help enthusiasts, industry, and others move toward positive safety outcomes."
In June 2009, the results of a VTTI naturalistic driving study titled "Driving Behavior of Commercial Vehicle Drivers" sparked national and international debate over the dangers of texting while driving, resulting in new laws for commercial vehicle drivers.
The study combined data from two naturalistic truck-driving studies previously conducted by VTTI to evaluate the types, frequency, and impact of driver distraction in commercial motor vehicle operations. Analysis of 3 million miles of motion and video data collected from 203 commercial vehicle drivers revealed that drivers who text while driving were 23.2 times more likely to be involved in a crash or near-crash event.
Another key finding, and one that provides balance in the debate, was that talking or listening on a cell phone or CB radio did not statistically increase crash risk. However, dialing, reaching, and other intensive visual-manual cell phone sub-tasks did.
The results from this study prompted a two-day Distracted Driving Summit in Washington, D.C., and an executive order by the president of the United States banning texting while driving for government workers operating government vehicles. On Jan. 26, 2010, the U.S. Department of Transportation announced a federal ban on texting while driving by truckers and bus drivers. Nearly a year later, a Notice of Proposed Rulemaking was released that would restrict, but not ban, cell phone use for truck drivers.
Commercial vehicle drivers are not the only people on the road impacted by Virginia Tech Transportation Institute's naturalistic driving research, as it has provided impetus in the passage of state laws banning texting while driving for all motorists.
"Our research has sparked national debate over the dangers of distracted driving and, if nothing else, has made and will continue to make drivers more aware of how quickly situations can change when they take their eyes off the forward roadway," says Charlie Klauer, research scientist in VTTI's Center for Automotive Safety Research.
Eye glance data gathered from more than 6 million miles of driving during VTTI's various naturalistic driving studies revealed that text messaging, which had the highest risk — more than 20 times worse than driving while using a phone — also had the longest duration of eyes not on the road: 4.6 seconds over a six-second interval. This equates to a driver traveling the length of a football field at 55 mph without looking at the roadway. "The danger increases exponentially when drivers take their eyes off the forward roadway for extended periods of time for any distracting task," says Klauer. Currently, 35 states and Washington, D.C., ban texting while driving. In Virginia, texting while driving is considered a secondary offense.
"The U.S. Department of Transportation, along with other organizations, including the Transportation Research Board, is active in utilizing the naturalistic data collection approach," says Richard Hanowski, director of the Center for Truck and Bus Safety at VTTI. "Large-scale naturalistic driving studies will continue to be an important method for understanding safety issues, in general, and human factors issues, specifically, that are key to reducing crashes on our nation's highways."
Several naturalistic driving studies with teens are being conducted. An 18-month study sponsored by the National Institutes of Health will determine issues associated with newly licensed teenage drivers that place them at a much higher crash risk when compared to other drivers.
The study is divided into two parts. The Smart Road at VTTI will allow researchers to manipulate task difficulty. Teenagers and parents will be tested on the Smart Road once at zero months and again at 12 months. During the naturalistic driving part of the study, teenage drivers will be recorded continuously during the first 18 months after they receive their license.
The study's results will provide transportation researchers with a more complete understanding of how teens learn to drive during the first 18 months of independent driving, and will possibly provide information to legislators who are working on graduated driver licensing laws in an effort to reduce teen fatalities.
NHTSA has provided additional funds to analyze the naturalistic driving data from the study of new drivers to help determine which performance parameters and situations should be used to provide feedback to newly licensed teens. By investigating which driving performance and behavior measures were most highly associated with crash and near-crash involvement, NHTSA and VTTI will be able to program a training module to provide feedback to the teens or their parents to make them aware of unsafe driving performance parameters so teens will be able to avoid those behaviors in the future.
A New Practice Driving Study, sponsored by the Eunice Kennedy Shriver National Institute of Child Health and Human Development, aims to assess important factors during the nine-month learner's permit phase and the first six months of independent driving.
Past research has indicated that crash rates of novice teenage drivers are significantly higher during the first six months and 1,000 miles of independent driving. Virginia Tech Transportation Institute researchers seek to determine the effect of greater and lesser amounts of supervised driving on the driving performance of newly licensed teens. In their analysis, researchers will analyze the amount and variety of practice provided by the teen's parent as well as compliance with the state laws.
VTTI has also been conducting research with drivers who are 75 and older. In a 2007 study sponsored by the National Surface Transportation Safety Center for Excellence, 20 senior drivers had their vehicles instrumented for one year. The study was one of the first long-term naturalistic observations of older drivers. During the initial phases of the study, 40 participants, made up of 20 senior drivers and 20 seniors who had given up driving within the last two years, underwent a battery of assessment testing including a driving history and tests of health, vision, physical strength, and reaction time. While results are still being tabulated, the study was renewed and continues. The data analyses will aid older drivers in making an educated decision about whether or not they should stop driving.
1,500 vehicles instrumented
VTTI is currently managing data collection from more than 1,500 cars, pickups, and sport utility vehicles in the largest light-vehicle naturalistic driving study ever conducted: the Second Strategic Highway Research Program. The nationwide study was authorized by the U.S. Congress to address the critical needs of the nation's highway system. The study is administered by the Transportation Research Board of the National Academies under a memorandum of understanding with the Federal Highway Administration and the American Association of State Highway and Transportation Officials.
Data is being collected at Buffalo, N.Y.; Tampa, Fla.; Seattle, Wash.; State College, Pa.; Bloomington, Ind.; and Durham, N.C.
The Virginia Tech Transportation Institute is responsible for designing hardware and software for collecting the naturalistic data, and training contractors (Battelle, Penn State University, University of Indiana, and Westat) to instrument vehicles and conduct operations at the sites.
VTTI monitors the health of the systems through cellular connection to each of its data acquisition systems. By the time the study is complete, researchers anticipate having gathered more than 2 million hours of driving data that will be used to explore formerly unexamined transportation research questions.
Roadway users of all ages and in all vehicle types face risks while driving as part of their everyday activities. But with the continual help of naturalistic driving research, the frequency of fatalities and injuries can be greatly reduced.