Two Wheels and a Crowd: Cyclist Decision-Making Dynamics in Shared Spaces

Abstract

This thesis aimed to investigate the decision-making behavior of cyclists in shared spaces, specifically examining the potential of dynamic gap acceptance models in predicting cyclists' responses to pedestrian crossings. By extending previous research on Diffusion Decision Models (DDM), this study incorporated the influence of pedestrian density on cyclists' decision processes. By doing a Continue / Brake cycling task in a simulated city environment, the effects of different environmental factors on decision outcomes and reaction times where tested and significant findings emerged. Time-to-Arrival (TTA) was found to negatively affect the likelihood of cyclists braking, indicating that a greater time to arrival influences decreases the probability of braking decisions. Although distance did not individually show a significant effect on decision outcomes, pedestrian density did. This suggests that higher pedestrian density increases the likelihood of braking, likely due to perceived risk. Reaction Times (RTs) were significantly influenced by both distance and density, with higher distance and density leading to shorter RTs. The interaction between TTA and pedestrian density notably affected RTs, with higher density conditions increasing decision complexity and cognitive load. Based on the statistical results, a baseline DDM based on previous research is compared to three models with different variations of density integration. The models with density integrated showed worsened performance, with the model where density influences urgency showing the most promise. This worsened performance highlights the need for further model refinement to capture cautious behaviors accurately.