Should I stay or should I go? Perceiving and sampling vehicle motion at the roadside

Visual motion is an important source of information when acting in a dynamic world. In road traffic, the ability to infer and to predict motion based on visual information is vital to avoid collisions and to navigate safely. This ability, however, can be compromised by various factors. Pedestrians are among the most vulnerable road users to the consequences of misperceiving the dynamics of others in road traffic. The aim of this thesis therefore was to investigate motion perception from the perspective of pedestrians. Across three empirical studies, we tested pedestrians’ perceptual judgements of approaching vehicles in a virtual road-crossing scenario. We find that the ability to accurately discriminate the arrival times of vehicles relies on smooth tracking of the vehicles’ motion paths. Pedestrians also display inter-individual differences in the way visual information is sampled and used to derive judgements about the vehicles’ speed and time-to-arrival, with some of them relying on sub-optimal visual cues (CHAPTER 3). Furthermore, keeping gaze on or at least close to a vehicle’s path enhances motion prediction. Conversely, allocating visual attention away from the approach direction of traffic considerably impairs the ability to predict the arrival of a vehicle at the crossing location (CHAPTER 4). Finally, we show that vehicle visibility and visual features of the road environment affect the perception of vehicle speeds. Whether speed is under- or overestimated depends on the spatial distribution of visual contrast; whether the distribution mimics atmospheric conditions (fog) or low vision (CHAPTER 5). Based on the empirical findings, we derive conclusions on how traffic safety measures could promote pedestrian safety by increasing public awareness of perceptual biases, facilitating optimal visual sampling, and improving visibility at crossing locations. We note that further research is needed under more naturalistic viewing conditions and on the link between perceptual judgments and road-crossing actions.