"The Driver's Only Human … " M.E. Online Exclusive #2, February 2008

February 2008

To read
our previous
ME Online Exclusives,
click here.

The Driver's Only
Human …

… And future advances in traffic safety will have to take that into account, the experts say.

The reduction in traffic deaths in the United States over the past four decades has been inspiring. We in the U.S. can expect fewer than 1.5 deaths for every 100 million vehicle miles traveled this year. That stacks up against 5.5 deaths per 100 million VMT in 1966.

At the same time, the rate of injury has been reduced about 40 percent—from 169 injuries per 100 million vehicle miles traveled in 1966 to about 100 injuries today.

Even so, 40,000 people or more are likely to lose their lives in traffic accidents in the U.S. this year, and another 3 million will suffer traffic-related injuries.

The number of deaths seems to be a plateau that we haven't been able to move past in more than 20 years. In only one year of the past 40, 1992, did traffic deaths decline to as few as 39,250.

If you talk to John D. Lee, a scholar of traffic safety at the University of Iowa, he will point out that, as impressive as the gains in auto safety have been, they fail to measure up to the safety record for commercial aviation. More than one million deaths a year are attributable worldwide to traffic accidents, and that is a major public health problem, he said. Lee is a professor in the university's Department of Mechanical and Industrial Engineering whose research interests include human factors and vehicle safety.

Crumple zones, airbags, and other protective features in addition to seat belts have made major contributions to auto safety. Automobiles are much safer environments than they once were.

A spokesperson for the Department of Transportation attributes the reductions made since the '60s to the use of seat belts and the crackdown on driving while intoxicated. DWI 20 years ago or so was often considered a peccadillo. It is taken much more seriously now, at least in part because of the efforts of public-interest groups like Mothers Against Drunk Driving. Although the problem is far from eliminated from the highway, increased enforcement and stiff penalties have brought about a significant change in behavior.

It took the law, too, to put seat belts into all cars. Then it took more legislation to convince drivers to use them. The National Highway Traffic Safety Administration estimates that the use of seat belts saves as much as $50 billion a year in medical care, lost productivity, and other injury-related costs in the United States.

Refinements in current systems and improvements in the balance of some top-heavy vehicles can probably save more lives. But Lee and others believe that, given the physics of auto crashes, there may not be many opportunities for substantial improvement to the crashworthiness of vehicles. According to Lee, future gains in traffic safety must come from improving the ability of drivers to avoid accidents.

Newer technologies, including electronic stability control, adaptive cruise control, and antilock brakes, help drivers avoid crashes. Other collision-avoidance systems are being developed.

Lee sees promise in new control technology. Electronic stability control uses computer algorithms to minimize skidding during quick turns and lane changes, as in passing on a two-lane highway. (Electronic stability control is the subject of an article, "Intelligent Safety," by Alan S. Brown in the December 2007 issue. You can find it among the back issues of Mechanical Engineering Online.)

Adaptive cruise control uses sensors to detect a vehicle in front, and will automatically adjust speed to maintain a safe following distance. Current systems, Lee said, can deliver warnings that alert the driver to take over when distance closes faster than the automatic system can handle.

Studded tires can also contribute to crash avoidance.

The new technologies contribute to auto safety when they are used properly, and therein lies one of the dangers associated with the human factor.

"Drivers typically are not passive recipients of technology," Lee said. "Sometimes they drive differently. Some safety benefits don't materialize."

According to Lee, people have said that they feel free to drive faster on snow with studded tires. Others have said they would drive faster in a car equipped with antilock brakes.

It seems that protective systems can make some drivers feel too well protected for their own good. "The driver can rely too heavily on technology," Lee said. It is conceivable, for instance, that a driver might try reading the paper in a moving car if he feels that the adaptive cruise control has the situation under control.

What's more, there is the question of getting drivers to use protective technology when they have a choice.

When seat belts were first introduced on a broad scale, there was a wisecrack about them: "I wouldn't mind seat belts if they weren't so uncomfortable to sit on."

Traffic laws eventually changed the behavior of the public by making it a ticketing offense to drive without using a safety belt. But even today, when there is no question that belts save lives, it seems that people cannot be relied on to use them. There are two kinds of seat belt laws in the U.S. In the first case—the states where the billboards tell us, "Click it or ticket"—a police officer can stop a vehicle and issue a summons if a driver fails to use a seat belt. In the second, a driver can be ticketed for not using the seat belt only if the vehicle has been stopped for another traffic offense. According to Lee, seat belts are more widely used in states where failure to do so is a primary offense than in those states where it is secondary.

Airbags were once thought to be a way around the problem. They do not require a driver's cooperation. Airbags alone, however, have been found to be less effective than seat belts. After all, the object is to keep the human body from being battered as it bounces around the cabin. The airbag is most effective in a head-on crash. The addition of side airbags improves the odds, but reliance on front and side bags together is still not believed to be as secure as the use of seat belts.

Collision alert systems can alert drivers to rapidly closing distances to objects. There are cruise control systems that do the same thing, but they must be intentionally engaged by the driver in order to work. Collision alert is working whenever the car is operated.

Lee has researched the results of different types of collision alert systems. There are some that can be designed to provide graded warnings, which continue to grow in urgency as the danger of collision increases. They can produce false alarms that can annoy drivers. But it seems that they are better at helping drivers avoid crashes than are systems that alert drivers only when collisions are imminent.

A haptic signal—say, an increase in force from the gas pedal—produced the quickest responses in drivers. An unexpected observation, Lee said, is that a system combining several sensory signals—auditory, visual, and haptic together—drew the slowest response times.

Meanwhile, there are some interesting indications that feedback—especially evidence of an individual driver's mistakes—can have a beneficial effect on behavior. Lee said the Public Policy Center at the University of Iowa had conducted a study in which high school age drivers had event-triggered video systems installed in their cars.

According to Daniel V. McGehee, director of the human factors and vehicle safety research program at the center, car crashes are the number-one cause of death among teenagers every year in the United States. As many as 40 percent of the deaths of people age 16 to 19 are traffic related. There are diseases that kill far fewer kids each year that get more research money and attention than the epidemic of car crashes among teens, McGehee said.

The center's video study was funded by an insurer, American Family Insurance, and it studied 25 drivers from a rural Iowa high school. The group included 16- and 17-year-olds. (The study started with 26 volunteers, but one driver moved out of state before the project was finished.)

The video system was provided by DriveCam Inc. of San Diego. The system was always on in the background, with one lens trained inside the vehicle on the driver and another on the area in front of the vehicle.

Audio and video information was collected in a buffer. The system kept a permanent recording only when an event threshold was exceeded. The system was triggered by accelerometers, which could detect a sudden change in speed or direction, and recorded the 10 seconds before and the 10 after each trigger event.

An LED indicated when the system was recording and also served to warn drivers when they had exceeded an acceleration threshold. Data was automatically downloaded via a secure wireless system when students drove into the school parking lot.

The teenagers and their parents received a weekly report that compared their number of events to the group's average and included video of safety-related events. (In a 2007 paper, "Extending parental mentoring using an event-triggered video intervention in rural teen drivers," published in the Journal of Safety Research, McGehee, Lee, and their colleagues refer to the recorded events as "teachable moments.")

The study identified 18 of the drivers who averaged 2.6 warnings every thousand miles. The other seven drivers, however, averaged about 23 warnings in the same traveling distance. The most frequent incidents were improper turning or curve handling, and abrupt braking.

After nine weeks of driving with the LED and getting a weekly report, the high-incident group had reduced its number of events to 6.4 per thousand miles. After nine more weeks, they were averaging 2.6 warnings every thousand miles, a reduction in safety-related events of 89 percent.

A synopsis describing the test is on the center's Web site at http://ppc.uiowa.edu/dnn4/HumanFactorsbrVehicle
Safety/TeenDrivingStudy/tabid/203/
Default.aspx.

According to McGehee, the center has received funding from the Centers for Disease Control to conduct a new study. This one will recruit 14-year-old drivers. Six mostly rural states, including Iowa, permit people to drive at that age, McGehee said. The study will track the drivers for four years.

McGehee said the technology had a very high acceptance rate among the teens. All of them recommended the program to other teens; all but one did not think it was an invasion of privacy. "Technology acceptance is key to the success of any product or process, so we were quite gratified with the teens' opinions," he said.

Rusty Weiss, the director of DriveCam's consumer division, pointed out that drivers who consider the video system invasive can avoid having their activities recorded. All they have to do is drive in a fashion that doesn't trigger a recording.

DriveCam's system has been used for a few years with commercial customers—primarily operators of vehicle fleets—to alert drivers to risky behavior and help them learn to avoid it. The company's Web site, http://www.drivecam.com/, shows clips of accidents and near-misses recorded in taxis, trucks, and buses in various countries. Weiss said he is a former DriveCam customer himself.

Sometime around the beginning of February 2008, the company expects to complete a rollout of the system as a consumer product. When he spoke to Mechanical Engineering, Weiss said some details were still being worked out, but gave a general idea of the business model for the service. The cost of the hardware may be about $450, and the company will work with the Best Buy electronics retail chain, which will provide installation services. There will be an additional subscription fee of $35 or so a month.

The device with the two lenses is installed in the car, and it records incidents in which accelerations exceed a threshold. This much is similar to the university's study. This version keeps a video record of eight seconds before the trigger incident and another four seconds after it. A report, including the video file, is sent over the Sprint network and will be available on DriveCam's server within 48 hours, Weiss said. Parents can log on to the server to review their child's driving record.

DriveCam will also monitor the driving records of its charges. "If we see two or three weeks of high-risk behavior with no improvement, we may contact the parents," Weiss said. The company may suggest resources for parents of high-risk drivers who fail to improve.

The company recommends that the system be used for a year, although a family can extend its subscription.

The product is already available to some families. American Family Insurance, which sponsored the University of Iowa's original study, makes it available to customers.

According to Weiss, the results have been significant. He said that, after six weeks of using the system, young drivers on average have changed their habits to the point where they stop triggering the device.

Technology is attempting to make up for driver error or misjudgment—the heavy foot, the too-sharp turn. It can also teach us what we did during the close call or the fender-bender so we can do better in the future. That appears to be having a beneficial effect on some of the highest-risk drivers.

Maybe enough of us drivers of all ages will be doing better behind the wheel in the future. Perhaps we'll all improve to the point where fewer people lose their futures on the road.

SIDEBAR: Technology vs. Fatigue

Inattention and fatigue are leading causes of accidents and close calls among drivers.

Fatigue is a very serious safety hazard for all drivers, including some who have plenty of mileage behind them. A third or more of the heavy truck crashes in the United States are attributed at least in part to driver fatigue. These are the long-haul truckers—who drive for a living. Sometimes these professional drivers see pushing for a few hours more as a way to make a little more profit.

Sometimes it puts them out of business.

Systems have been developed to detect signs of fatigue or inattention among drivers and alert them to their danger. DriveCam in San Diego makes a video system triggered by sudden changes in acceleration that creates a record that drivers can use to learn by their mistakes.

Others are focusing specifically on systems to detect fatigue and warn drivers. Seeing Machines Ltd. in Canberra, Australia, has taken technology from a human performance tracking device called faceLab and adapted it to monitor driver fatigue and attention.

The faceLAB system is designed to monitor the human face. According to Nick Langdale-Smith, the company's director of sales and marketing, faceLAB "tracks faces, eyes, pupils, eyelids, whole heads, and various other facial features to determine human attention, inattention, fatigue—and a whole raft of various other states." It has been used in psychology research, military training and simulation, and other areas in which it is useful to track and understand the gaze direction of the subject. It has been used to study how drivers read road signs.

A camera monitors the person's head and records various telling signs such as head pose, eyelid movement, and gaze direction. If the face is close enough, the system can pick up the dilation of pupils and the movement of the eyes as they maintain focus. A computer analyzes and records the information.

Seeing Machines packages a similar technology in its Driver State Sensor. A small camera mounted on the dashboard monitors the driver's face. By analyzing the image, particularly the closure of the eyelids, the system can detect symptoms of fatigue. A speaker may issue a warning to the driver.

According to Langdale-Smith, "Better warning systems are being trialled by various automotive manufacturers, and these include vibrating the seat, changing the vehicle dynamics, or adjusting parameters of various other safety systems. For instance, if the driver is detected as being distracted, the vehicle may autonomously put more distance between the vehicle and the one in front, or pre-tension the seatbelts."

Potential customers for the Driver State Sensor include operators of truck fleets. They can choose an option to have data transmitted to a central location. The system also can be used in other applications where fatigue is a serious hazard— say, in aircraft cockpits, at the consoles of unmanned aerial vehicles, or at the controls of heavy machinery.

The original faceLAB is a research tool that demands a certain amount of subject cooperation. "The DSS has been designed specifically to cope with real-world situations on the fly," Langdale-Smith said. "The system employs its own infrared illumination."

Langdale-Smith said that the Driver State Sensor has been available for about six months and that some vehicle fleets in the United States are using it on a trial basis. The system is described on the company's Web site at www.seeingmachines.com/dss.html.

Another driver fatigue monitoring system was developed by engineers at Triodyne Inc., a mechanical engineering consulting firm in Northbrook, Ill. Designed with heavy trucks in mind, it uses off-the-shelf parts, including the sensing devices used in antilock braking systems, to gauge driver alertness by measuring the activity of the steering shaft. It can also be used in passenger vehicles.

The system was the subject of a technical paper delivered by the inventors, Christopher W. Ferrone and Charles Sinkovits, at ASME's 2005 International Mechanical Engineering Congress in Orlando. According to Ferrone, when they are alert, drivers continually steer to make small corrections in the direction of the vehicle. Without that micro-steering, an automobile will drift. So when the micro-steering stops, it is likely that the driver isn't paying attention, and may even have fallen asleep at the wheel.

The paper by Ferrone and Sinkovitz, "Driver Fatigue/ Inattention Monitoring Device—An Integrated System for Heavy Trucks," said the main components of the system are an ABS tone ring, ABS sensor, wiring, and a networked input-output module. The module contains a microcontroller.

According to Ferrone, the ABS sensor, which is designed to monitor wheel rotation in an antilock braking system, is used in the fatigue monitoring device to look for movement of the steering shaft. The sensor's output is proportional to the number of motions in the shaft over a designated period. If the signal drops below a threshold for a predetermined number of seconds, the system kicks in. It can be made to deactivate cruise control and to initiate other actions, including activating an engine brake. It can also control hazard lights to warn nearby motorists. "We can make it static, or make it so it can learn and adapt to different drivers," he said.

A proof of concept was installed on a 1996 Peterbilt 379 tractor. Ferrone worked for Triodyne when the device was developed. He has formed an engineering firm, Areca Inc. in Pittsburgh, which includes Triodyne among its clients.

Triodyne received a patent (No. 7,138,923) for the driver fatigue montoring system in November 2006.