Navigation Content
University of Wisconsin Madison College of Engineering
You are here:
  1. Home > 
  2. News > 
  3. News archive > 
  4. 1997 > 

Device May Help Prevent "Falling Asleep at the Switch"

In the blinking of your eye, Ron Leder wants to tell whether you are about to have what he terms an "alertness outage." If his PhD research project succeeds, you may find yourself wearing a sensor that weighs less than a gram on your eyeglasses to measure the movement of your eyelid. The sensor will have a wire or wireless connection to a microprocessor unit on your belt that will interpret the signals and sound an alarm if you are no longer alert enough to do your job.

If you're a truck driver, that could prevent an accident that might cost your life. If you're a chemical-plant operator, it could prevent something even worse.

Eyeglass-mounted sensor.

Eyeglass-mounted sensor. (large image)

Leder, a University of Wisconsin-Madison graduate student in the biomedical area of the Department of Electrical and Computer Engineering (ECE), admits that he's only part way along to his goal. His device, which was developed with support from the Center for Human Performance and Risk Analysis, has been tested on a few subjects in Anatomy Professor Dwayne Yamasaki's lab at UW-Madison, but the technology is a long way from the shelves of a truck stop or a safety equipment supplier.

Developed with help from his advisor, Professor John G. Webster in ECE, Leder's sensor system can tell whether an eye is open, closed, or anywhere in between. The tiny device uses a light-emitting diode (LED) to shine invisible infra-red light toward the eye. A photodiode mounted alongside the LED measures the light reflected from the eye and determines the position of the eyelid (oddly enough, an open eye reflects less light than a closed eye).

Rod Leder with alert eyeglasses

Ron Leder wears "alert" eyeglasses. (large image)

Leder's device works as well as other methods for measuring eyelid movement. Why bother reinventing the wheel? Webster says that the existing systems are either too intrusive (they rely on sensors attached to the eyelid) or do not provide a complete picture of the eyelid's motion. "We reviewed what was out there and decided on an optimal compromise between unobtrusiveness and quality of data," Webster says. "We wanted to find out if it's possible to determine vigilance by a sensor that's not intrusive." That ruled out another technology -- attaching the electrodes needed for electroencephalo-graphy (EEG), which is the so-called "gold standard" of alertness measurement.

The easier part of Leder's project has been developing the equipment, the harder part is figuring out what the equipment is measuring. Obviously, Leder says, eyelid movement has some relation to alertness, since an eye that is closed is "visually not alert." But if the eyes are already shut, it's a little late to warn of an alertness outage, since the damage may already have happened. In fact, Leder observes, people may be in the twilight zone long before their eyelids close.

Leder says his goal is "to be predictive. We'd like something that says, the probability is X percent that in the next 10 minutes you're going to have an alertness outage. So from your blinking, we know what you're thinking."

The next step is to test the apparatus in a situation where there are likely to be "varying levels of alertness," as Leder delicately puts it. That's a euphemistic description for an experiment that Claudio Stampi is running at the Institute for Circadian Physiology in Cambridge, Massachusetts, where subjects are getting by on three hours of sleep per night for 15 days at a stretch.

The experiment is intended to find out whether this limited amount of sleep is most effective in one chunk, or as a series of naps. To Leder, it offers a chance to test his eyelid sensor on a lot of drowsy and partially refreshed volunteers, and to do the four kinds of tests required for alertness studies: EEG; psychomotor (e.g., reaction time) tests; mental function tests (for example, mental arithmetic); and self-ratings of mood and alertness.

Leder hopes to use data from the Massachusetts experiment to identify patterns of eyelid movement that precede loss of alertness. From that, he hopes to construct a mathematical model and algorithm so the device can teach itself to interpret individual eyeblink patterns.

"Falling asleep at the switch" is big business, Leder says; estimates of the cost of lost productivity and accidents due to reduced alertness run as high as $70 billion per year. "The ultimate measure of alertness is performance," he adds. "If you feel lousy but can still do the job safely and accurately, that's the bottom line according to most alertness investigators." Alternatively, many people may feel fine but still be too drowsy to perform well on the job.

Would anybody use the kind of alertness system Leder is developing? Nobody knows for sure, but Leder says trends are positive. For example, Nissan Motor Co. is offering an alertness sensor in its heavy trucks, and a 1995 workshop on fatigue in transportation sponsored by the National Aeronautics and Space Administration and the National Transportation Safety Board revealed a lot of interest in alertness issues.