To operate autonomously, minimal robot swarms must make timely and reliable collective decisions despite noisy individual sensing and severe constraints on communication, computation, and memory.
Achieving this capability could expand their use in applications such as healthcare, disaster response, and environmental monitoring.
Researchers compared two decentralized decision rules in minimal robot swarms: direct switching between opinions (“change your mind immediately”) and cross-inhibition (“first become undecided”), a mechanism found in biological systems including neural populations and insect colonies.
A useful analogy would be a committee choosing between A and B:
Direct switch: “Mary says B is good, so I’m voting B now.” Cross-inhibition: “Mary’s evidence makes me withdraw my vote for A. I’ll remain undecided until I’m persuaded.”
Cross-inhibition separates losing confidence in your current state from adopting somebody else’s state. That extra intermediate state gives the collective considerably better dynamics.
