InterJournal Complex Systems, 906
Status: Accepted
Manuscript Number: [906]
Submission Date: 2004
A Multi-Modeling Approach to the Study of Animal Behavior
Author(s): Jeffrey Schank ,Sanjay S. Joshi

Subject(s): CX.41



We describe a multi-modeling approach using individual-based probabilistic, dynamic system, and robotic models for the study of sensorimotor development in Norway rat pups. These studies are aimed at informing both animal behavior and autonomous robotics research. For the study of behavior, the crucial idea is that by simulating animal behavior by instantiating actual sensorimotor rules, complex individual and social behaviors can be seen as emerging from group dynamics and the interactions sensors, bodies, and physical properties of the environment. For robotics, studying infant mammals can facilitate the study of development of sensorimotor rules from the simple to the complex, providing insight on the development of autonomous robots, which face increasingly complex tasks. Our research is unique because we use animal, simulation, and robot studies in parallel and to inform each other. Our primary focus in this paper is on developing methods for this multi-modeling approach. Towards this end, we describe our simulation models, robotic rat pups, tracking algorithms, data analysis techniques, and initial results.

Retrieve Manuscript
Submit referee report/comment

Public Comments: