Good vibrations. Drivers in a simulator drove safer with vibratory warnings of impending collisions.

Robert Gray/ASU

A Shake May Prevent a Crash

SAN FRANCISCO, CALIFORNIA--You're driving down the highway late at night, talking on your cell phone and fiddling with the radio. As you begin drifting into the lane on your right, your steering wheel suddenly vibrates, and you swing back to the left, narrowly missing a car you never saw. Such tactile warnings may be more effective than are audio or visual cues in preventing collisions, according to a study presented here on Sunday at the annual meeting of the Association for Psychological Science.

Automobile accidents are the leading cause of death for those between the ages of 3 and 35 in the United States, claiming more than 42,000 lives per year. The vast majority of accidents are caused by inattentive drivers, says Robert Gray, a psychologist at Arizona State University in Phoenix. Driving safely over great distances requires constant vigilance in a monotonous environment, he says, something for which stimulus-hungry humans are poorly equipped. Hazards usually come suddenly, leaving little time for distracted drivers to react.

To boost their safety records, many car manufacturers are designing systems to detect impending collisions, using radar or lasers, which warn the driver. Transportation psychologists broadly agree that a collision-warning system can help prevent accidents, but just what kind of warning is most effective remains unclear. For example, are flashing lights better than bells or beeps? Should the warning system tell drivers the direction of the hazard or the direction to steer to avoid it?

To put various warning systems to the test, a team led by Gray asked 32 volunteers to go for a hazardous cruise in a driving simulator. The drivers tried to avoid hitting pedestrians and other cars that appeared unexpectedly from either the left or the right. To make the challenge more realistic, the road was foggy and the drivers had to have conversations on hands-free cell phones. In different sessions, drivers were given warnings that came either 2 seconds or 4 seconds ahead of potential collisions, and the warnings varied by sensory input--visual, auditory, or tactile--and either indicated the direction of the hazard or the direction of safety. (The tactile warning came as vibrations in bands attached to each arm.) A control condition gave drivers no warning at all.

The results suggest that drivers in danger are best warned through their skin. Those who received visual warnings were no better at avoiding hazards than when they received no warning at all. Auditory warnings were about twice as good as the control, but tactile warnings were the best of all, allowing drivers to react correctly to incoming hazards about three times as quickly. And the direction of the warning made a significant difference, but it depended on the amount of time to react. With a leisurely 4 seconds before impact, drivers performed better if the warning indicated the direction of the hazard. But with only 2 seconds before impact, it was better to receive an indication of the direction of safety.

The system may work in a simulator, but it could be "very difficult to use tactile warnings in a car environment," says Neil Charness, a psychologist Florida State University in Tallahassee. For one thing, road vibrations may make it hard to detect a tactile warning. And then there's basic human psychology to contend with: "We don't have enormous success getting people to put on safety belts," says Charness, "let alone 'sensor belts.' "