Motion Perception is not Direct with Indirect Viewing Systems

Research has shown that when a driver’s field of view is limited, judgments of coincidence with an external object are quite inaccurate under both real and simulated driving conditions. Two laboratory studies are reported in this paper that sought to investigate and improve performance on such tasks. In both studies, subjects looked at a series of vision-only video recordings taken from the driver’s eye position, which showed the bonnet of the car moving up to and past a stationary vehicle. Subjects were required to indicate when they, as the driver of the moving vehicle, felt they were directly in line with the front bumper of the approached vehicle. The time lag between indicated and true coincidence was re-expressed in terms of the distance from true coincidence. In the first study it was shown that this distance error increased with approach speed and more importantly, that when very little peripheral information was present, subjects were less accurate than when more peripheral information was available. These results replicate previous findings, which suggest that peripheral vision is important to speed and distance judgments. Even with the larger field of view, coincidence judgments were still poor. The second study showed that the accuracy of such judgments can be improved. When vehicle noise appropriate to the visually presented approach speed was added, subjects’ responses were closer to true coincidence. Furthermore, by mismatching vehicle noise and visual approaches it was shown that subjects’ responses were biased in the direction of the inappropriate noise cue. The findings appear not to support Gibson’s notion that perception is completely ‘direct’. They also show that driver performance using indirect viewing systems may be improved by the use of both increased peripheral visual information and sound which reflects vehicle speed.