Conformal fusion algorithm restores coverage guarantees in multi-robot mapping

Accurate mapping with formal reliability guarantees is a critical challenge for multi-robot autonomy. Conformal prediction offers a distribution-free way to equip each robot's local map with finite-sample coverage guarantees, but those guarantees degrade in practice: temporal correlation along a robot's trajectory breaks the exchangeability assumption, and each robot only observes a spatially limited, non-uniform portion of the environment. To address this, Ritvik Mahajan and colleagues from KTH Royal Institute of Technology propose a distributed fusion algorithm that takes these degraded per-agent guarantees as given and recovers the desired coverage across the team. Robots exchange lightweight scalar e-values with their neighbors, and a receiver fuses them using a per-neighborhood miscoverage budget and an uncertainty-attenuated fusion operator. They prove the fused set-valued map recovers the target user-specified coverage level regardless of communication graph topology or sensor noise distribution.

A notable limitation is that where fused evidence is insufficient, the map declines to commit and returns both labels (free and occupied), leaving a significant fraction of the domain unclassified rather than thresholded into a single decision. In simulated multi-agent mapping experiments, the fused predictor reliably meets its theoretical coverage bounds, and denser communication topologies significantly enhance map efficiency by shrinking the unclassified fraction. This work provides a theoretically grounded, practical solution for deploying teams of robots with formal guarantees on their environmental understanding — essential for safety-critical applications like search-and-rescue, autonomous exploration, or warehouse logistics where unreliable maps can lead to collisions or missed targets.