Modeling lost person behavior with spatial and temporal decision points

Efficient search and rescue (SAR) operations depend on tools and strategies that can understand and predict the behavior of lost individuals. Traditional methods, like the distance-ring model, assume uniform speed and direction and neglect the effect of the landscape on decision-making. In this paper, we introduce an agent-based model of lost person behavior that incorporates both spatial and temporal decision points to reflect how an individual might adjust their behavior over time and due to the landscape. We also introduce a novel behavior called “contouring” where the person will maintain their current elevation, and have expanded the map to include more detailed terrain data. Using real-world lost person incident data from the International Search and Rescue Incident Database, we find a best-fit behavioral profile for a hiker in the wilderness and validate using actual routes taken from real SAR incidents. Ultimately, this lost person behavior model is intended to improve SAR outcomes by leveraging empirical data and more accurately representing the dynamics of lost person movement.