Max Robotics — Global Service Robot Directory + US Certification

About

Michael Crosscombe is a researcher focused on collective decision-making and distributed learning in autonomous systems, particularly within swarm robotics. His primary contributions lie in advancing the best-of-n problem, where a group of agents must identify the optimal option from a set of alternatives through decentralized interaction. Crosscombe’s most cited work (2017, 27 citations) introduces a novel approach to robust distributed decision-making by incorporating a third truth state—representing uncertainty or indifference—into the weighted voter model, significantly improving resilience in robot swarms. He further extends this line of inquiry with his 2021 paper on collective preference learning (10 citations), which explores how decentralized systems can collaboratively learn preferences without centralized control. Crosscombe also investigates epistemic sets (2019, 8 citations) and imprecise evidence in social learning (2024, 3 citations), proposing models where agents represent beliefs as sets of possible states to handle ambiguity. His recent work on interaction constraints (2024, 2 citations) examines how limiting communication can enhance system performance. Crosscombe’s research is notable for bridging theoretical frameworks with practical applications in swarm robotics, offering scalable solutions for autonomous decision-making under uncertainty.