“That’s a bit of a tight squeeze!”: A thematic analysis of narrow passage driving interactions using the Perceptual Cycle model

• Limited research exists on the relationship between what drivers perceive, how they interpret it, and their actions during narrow passage interactions. • To rectify this, an on-road study was conducted with participants verbalising their thoughts, followed by a retrospective interview. • Drivers consider four aspects of the environment: the road narrowing, the interaction partner, give way gaps & surrounding vehicles. • Drivers use this information to evaluate five key questions, considered in tandem, to safely resolve narrow passage interactions. • The study also suggests that an informal “convoy” rule and three possible outcomes exist during narrow passage interactions. The ability to accurately represent driver behaviours in microscopic traffic models and for autonomous vehicles to discern human driving behaviours is reliant on having a deep understanding of the behaviour that is being modelled. In spite of this, the decision-making processes undertaken during narrow passage interactions, a high risk and relatively unregulated situation in which a road narrowing means that two opposing vehicles cannot pass through at the same time, remain under investigated. To rectify this limitation, an on-road “think aloud” study was conducted with participants in the UK and supplemented by a video-cued retrospective interview using a shortened version of the Schema World Action Research Method. 175 decision points were then analysed using Neisser’s Perceptual Cycle Model, which highlights the relationships between the information perceived from the narrow passage environment, a driver’s schema, and the actions taken. It was found that drivers conduct four key assessments, including evaluating their interaction partner’s intention and the suitability of give way gaps on both sides of the road, when deciding what actions to take during a narrow passage interaction. These results provide a clearer understanding of how driver behaviour is influenced by different contextual/situational factors in a safety critical driving situation and helps ensure that future mathematical models better reflect driver decision-making during narrow passage interactions. This, in turn, can be used to ensure that autonomous vehicles are able to safely interact with human drivers at narrow passages and that microscopic traffic models are able to produce more accurate outputs.