UDON: Uncertainty-weighted Distributed Optimization for Multi-Robot Neural Implicit Mapping under Extreme Communication Constraints

Accepted by ICRA 2026

Hongrui Zhao*, Xunlan Zhou*, Boris Ivanovic, Negar Mehr
University of Illinois Urbana-Champaign
Nanjing University
NVIDIA Research
University of California Berkeley

In a challenging real-world experiment under severe communication deterioration (e.g., a communication success rate as low as 1%), our method UDON enables each turtlebot to successfully map the full scene while only physically visiting half of the scene (explored areas and trajectories are colored accordingly). UDON maintains high-fidelity reconstructions with accuracy comparable to the ground truth, whereas the baseline method (RAMEN) fails to converge under such extreme disruptions.

Abstract

Multi-robot mapping with neural implicit representations enables the compact reconstruction of complex environments. However, it demands robustness against communication challenges like packet loss and limited bandwidth. While prior works have introduced various mechanisms to mitigate communication disruptions, performance degradation still occurs under extremely low communication success rates. This paper presents UDON, a real-time multi-agent neural implicit mapping framework that introduces a novel uncertainty-weighted distributed optimization to achieve high-quality mapping under severe communication deterioration. The uncertainty weighting prioritizes more reliable portions of the map, while the distributed optimization isolates and penalizes mapping disagreement between individual pairs of communicating agents. We conduct extensive experiments on standard benchmark datasets and real-world robot hardware. We demonstrate that UDON significantly outperforms existing baselines, maintaining high-fidelity reconstructions and consistent scene representations even under extreme communication degradation (as low as 1% success rate).

Qualitative Results

Visual comparison of the reconstructed maps. Under extreme communication constraints, baseline methods such as Di-NeRF and RAMEN fail to converge or produce highly noisy representations. In contrast, UDON maintains a consistent, high-fidelity scene representation, accurately recovering structural details (highlighted in orange) comparable to the ground truth.

Qualitative comparison of UDON against baselines

Quantitative Results

Robustness evaluation across varying communication success rates. As communication severely degrades (down to a 1% success rate), UDON dramatically outperforms prior methods, sustaining low artifacts and holes while maintaining a near-optimal completion ratio.

Quantitative evaluation charts

BibTeX

@article{zhao2025udon,
  title={UDON: Uncertainty-weighted Distributed Optimization for Multi-Robot Neural Implicit Mapping under Extreme Communication Constraints},
  author={Zhao, Hongrui and Zhou, Xunlan and Ivanovic, Boris and Mehr, Negar},
  journal={arXiv preprint arXiv:2509.12702},
  year={2025}
}