Decoding wayfinding: analyzing wayfinding processes in the outdoor environment

Abstract

Navigating complex environments is crucial for human life, yet understanding the cognitive processes involved in its wayfinding component remains challenging. One theoretical model that explains these processes is Downs and Stea’s four-step model. Our study builds on this model to empirically analyze its steps, focusing particularly on the monitoring step. Machine learning models were trained on gaze behavior and head/body movement data from over 300 routes walked by 56 participants in a real-world outdoor study, predicting three of these wayfinding steps: self-localization, route planning, and goal recognition. Applying this trained model to the respective monitoring segment of the same routes suggests that monitoring includes micro-versions of these three steps, indicating it operates as a recursive process rather than a distinct cognitive step. By bridging theoretical frameworks with empirical evidence, these findings enhance our understanding of spatial cognition and can inform the design of navigational tools and urban spaces.