Carnegie Mellon University
|Topic:||Artificial and Natural Colonies|
|Date:||Tuesday, April 10, 2001|
|Place:||Gould-Simpson, Room 701|
Multi-agent and multi-robot systems promise more reliable and effective solutions to many tasks than individual agent solutions can provide. But how can we build, organize and control large-scale multi-robot systems? To investigate these issues my research is focused along two major thrusts: (1) building large-scale multi-robot systems, and (2) observing and modeling live multi-agent systems. By building physical multi-agent systems, we are able to empirically explore critical issues like communication, distributed sensor fusion and cooperation. By observing and modeling live multi-agent systems (such as social insect colonies), we can gain insight into how successful multi-agent systems are organized.
In this talk I will describe recent research in these areas, including: the development of a robot team that uses novel communication protocols for collaboration and distributed sensing, robot control strategies for object manipulation in dynamic environments, and a fully automated system for observing and tracking the activities of live social insect colonies composed of thousands of individuals.