Proceedings:
Book One
Volume
Issue:
Proceedings of the AAAI Conference on Artificial Intelligence, 20
Track:
Agents / Multiagent Systems
Downloads:
Abstract:
We describe a decentralized algorithm for coordinating a swarm of identically-programmed mobile agents to spatially self-aggregate into arbitrary shapes using only local interactions. Our approach, called SHAPEBUGS, generates a consensus coordinate system by agents continually performing local trilaterations, and achieves shape formation by simultaneously allowing agents to disperse within the defined 2D shape using a Contained Gas Model. This approach has several novel features (1) agents can easily aggregate into arbitrary user-specified shapes, using a formation process that is independent of the number of agents (2) the system automatically adapts to influx and death of agents, as well as accidental displacement. We show that the consensus coordinate system is robust and provides reasonable accuracy in the face of significant sensor and movement error.
AAAI
Proceedings of the AAAI Conference on Artificial Intelligence, 20