A Unified Mathematical Approach for Foraging and Construction Systems in a 1,000,000 Robot Swarm

Abstract

Automation in construction is possible with systems designed using the swarm robotic principles of scalability, flexibility, robustness, and emergence. We derive quantitative measurements of these principles in 10,000 robot swarms as a first step in achieving this goal. We summarize our recent task allocation work in the context of an object gathering task and demonstrate its feasibility in the context of automated construction tasks. We present a trajectory to extend our current task allocation methodology using stochastic processes in order to present a unified approach to task allocation in swarm-robotic construction.

Publication
Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence, IJCAI-19