Eye movements the key to avoiding car accidents

Berlin (dpa) – Because your eyes move faster than your arms, an

optical device may one day steer your car around sudden, sharp curves

the instant you see the danger ahead, according to a team of German


“The ultimate goal of the project is to build a device into cars

that warns the driver if he is in danger of unintentionally departing

from the lane,” says Farid I. Kandil of the Department of Psychology,

University of Muenster, Germany.

It is well-documented that when negotiating winding roads, drivers

tend to look at a specific, well-defined point on the lane marking –

referred to as the tangent point.

The German researchers have found that the further drivers can

look ahead, generally in left-hand curves, wide curves and when

leaving a curve, the less they have to look at the tangent point.

Constructing a device which registers those subtle eye movements

could reduce the danger of losing control in a curve.

The findings of the scientists are reported in a recently

published article in the Journal of Vision. In the study, six drivers

test-drove a car repeatedly through a series of 12 right- and

left-hand bends, or curves, on real roads while their eye movements

were recorded.

The results confirmed that when moving into a curve, drivers rely

heavily on using the tangent point before turning the steering wheel.

The findings further revealed that a driver will look at the

tangent point 80 per cent of the time when there is a shorter sight

distance, such as with sharp, right-hand curves. In open bends such

as left-hand curves, and when leaving curves, drivers spent a third

of their time looking at the end of the curve and the straight road

that comes after.

The experiments were conducted in right-hand traffic as in

continental Europe and the United States. According to the

researchers, there are many hints suggesting that the results can

also be used to predict how drivers negotiate curves in left-hand


“The system we envision will look out for upcoming curves and

retrieve information about the eye movements the driver normally

performs,” explains Kandil. “If the driver does not show his typical

pattern of eye movements upon approaching a bend, then the system

will assume that he has not seen it and will warn him in time.”

The researchers plan to conduct additional experiments, using a

prototype to determine whether the warning system provides enough

time for the driver to react properly.