Human performance in apparently simple tasks like walking in a crowd, bicycling and driving, is deeply impressive. We learn these kinds of tasks efficiently, and cope under a variety of conditions., but this adaptability is in contrast to current AI and robotics, where the complexity and ambiguity of the real world have turned out to be huge challenges. This suggests that the human brain has evolved techniques for organizing perception and action that may be different from current AI. This research explores the hypothesis that a highly efficient representation of space underlies human adaptability. It combines computational and experimental methods to investigate how our cognitive system creates this “sense of space”. The integrative approach that takes the research beyond the state-of-the-art is how data from multiple experimental tasks derived from analysis of natural task constraints are used to constrain the development of fully implemented unified models of cognition.

2020

2025