Object persistence for synthetic creatures

  • Authors:
  • Damian A. Isla;Bruce M. Blumberg

  • Affiliations:
  • MIT Media Lab, Cambridge, MA;MIT Media Lab, Cambridge, MA

  • Venue:
  • Proceedings of the first international joint conference on Autonomous agents and multiagent systems: part 3
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present methods for anticipatory behavior in simulated graphical creatures. We discuss in general terms the importance of anticipatory behavior through explicit expectation formation. We present an in-depth description of a specific type of expectation-formation, namely location-expectation, or object persistence. A new representation - the Probabilistic Occupancy Map (POM) - is presented, and it is shown how this representation can be used to maintain estimations of the positions of mobile objects in the world based on both positive and negative knowledge provided by the creature's perceptual system. Finally a number of illustrative results are presented that show Duncan, our simulated dog, successfully performing a number of tasks that require a high degree of spatial common sense.