Abstract

Spatial cognition is an essential process for the survival and adaptation of biological organisms and for effective behaviors in artificial agents, and it has been extensively investigated using artificial intelligence (AI) techniques. In this study, we present two distinct AI-driven approaches to study spatial cognition: the first approach leverages AI to extract behavioral data through direct observation in a natural environment, specifically studying the orientation strategies of animals; AI applied to ecological observation enables precise tracking of movements and decisions, providing detailed data on their trajectories and use of environmental cues. The second approach uses simulation with artificial agents in controlled environments to model and test navigation strategies. This dual application of AI demonstrates its versatility and complexity, highlighting how it can be employed in complementary approaches to fully exploit its potential in the study of spatial behavior.