Home // COGNITIVE 2012, The Fourth International Conference on Advanced Cognitive Technologies and Applications // View article

Learning Cognitive Human Navigation Behaviors for Indoor Mobile Robot Navigation

Authors:
Luz Abril Torres Méndez
Roberto Cervantes Jacobo

Keywords: human navigation; cognitive conceptual maps.

Abstract:
We present a framework to transfer cognitive human navigation behaviors to an artificial agent so it can generate route directions similar to those created by humans. Our method is based on a spatial conceptual map that attempts to emulate the cognitive process carried on by living beings during the navigation process. This conceptual map is modeled as a three-level of interconnected graphs to simulate human spatial reasoning. We based some of our ideas of spatial reasoning on qualitative definitions of neighborhood, distance and orientation. The first level of the conceptual model contains the approximated metric of the environment and the physical obstacles that influence the navigation trajectory. In the second level, we define the abstract characteristics that give information about the ambient, such as the areas of influence and key features. Finally, in the third level, the navigation route obtained from the first two levels is stored. The visual and cognitive skills of each person in the experiments are captured in terms of the space-time perception while navigating. Our experimental results demonstrate that the inference of the route directions can be easily obtained and transferred to an agent from this spatial conceptual map.

Pages: 37 to 45

Copyright: Copyright (c) IARIA, 2012

Publication date: July 22, 2012

Published in: conference

ISSN: 2308-4197

ISBN: 978-1-61208-218-9

Location: Nice, France

Dates: from July 22, 2012 to July 27, 2012