In this paper results are presented from a pilot study that explores validity of looking behaviour in a driving simulation. Four subjects completed test rides in a simulator and on the road, in both conditions on a similar two-lane test route that included curves. Each trial (real versus simulator) was performed three times in daylight conditions. During the sessions eye movement behaviour was recorded via the Dornier Eye-Tracking-System (ETS). The results show that subjects use the same cues (lane markers) and very similar systematic viewing strategies during curve negotiation in reality and in simulation. However, the difference in ranges of measured viewing angles in the driving simulator compared with reality cannot yet be fully explained. The cause for greater angles as found in the simulator requires further data collection and calibration. Nevertheless, the results indicate that the driving simulator is a useful tool for further studies concerning the investigation of drivers' visual behaviour. (A)
Abstract