The present study investigates what visual information a driver needs in order to be able to steer his car on a straight road. In chapter 1 this question is addressed by reviewing the literature on control models of drivers and on the perception of egomotion. In chapter 2 a detailed analysis is presented of the possibilities a moving observer has to perceive his own movement. The general formulation of this problem allows for a more unified discussion of earlier formulations. In chapter 3 a laboratory experiment is described in which thresholds for the perception of two relevant dynamic visual cues are determined. These are the optical effects of lateral speed and of heading rate. In chapter 4 an experiment is described which determines whether the proposed visual cues are actually used during straight road driving. By manipulating the visual environment and the available visual field the perceptability of the various cues is varied and their effects on driving registered. In chapter 5 a method is developed and tested that uncouples the perception of lateral position from that of lateral speed. The literature on motion perception suggests that direct movement perception is still possible at a 25 hz stroboscopic occlusion rate, but not at a 5 hz rate. However, at the latter rate the perception of position cues should still be intact. In chapter 6 an experiment is described in which the method of stroboscopic occlusion is used during actual driving. The main result of this experiment is that information about lateral position is by itself insufficient for keeping the car between the lane boundaries. In the final chapter a general discussion of the findings is presented. The possibilities for integrating these findings in control models are discussed. Suggestions are given for the application of the results to problems of traffic safety.
Abstract