UDON: Uncertainty-Weighted Distributed Optimization for Multi-Robot Neural Implicit Mapping under Extreme Communication Constraints
Hongrui Zhao, Xunlan Zhou, Boris Ivanovic, Negar Mehr
AI summary
Problem
Multi-robot neural implicit mapping struggles to maintain reconstruction quality and consensus when communication is severely disrupted by packet loss or limited bandwidth, often causing prior consensus algorithms to diverge or produce incomplete maps.
Approach
UDON replaces the single aggregate dual variable per agent with independent, uncertainty-weighted dual variables for each communication link, allowing agents to dynamically prioritize reliable data from active neighbors while ignoring stale data from disconnected peers.
Key results
- Maintains high-fidelity reconstructions at 1% communication success rate
- Reduces artifacts and holes compared to RAMEN and MACIM baselines
- Prevents optimization divergence in highly sparse communication graphs
- Validated across synthetic benchmarks and real-world TurtleBot hardware experiments
Why it matters
Provides a robust foundation for collaborative 3D mapping in multi-robot systems operating in bandwidth-constrained or high-interference environments.
Abstract
Multi-robot mapping with neural implicit rep- resentations enables the compact reconstruction of complex environments. However, it demands robustness against com- munication challenges like packet loss and limited bandwidth. While prior works have introduced various mechanisms to mitigate communication disruptions, performance degrada- tion 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 consis- tent scene representations even under extreme communication degradation (as low as 1% success rate). The codes can be found at https://iconlab.negarmehr.com/UDON/. 1 Department of Aerospace Engineering, University of Illinois Urbana- Champaign hongrui5@illinois.edu 2 School of Intelligent Science and Technology, Nanjing University wyattzhouxl@smail.nju.edu.cn 3 National Key Laboratory for Novel Software Technology, Nanjing University 4 NVIDIA Research bivanovic@nvidia.com 5 Department of Mechanical Engineering, University of California, Berkeley negar@berkeley.edu ∗These authors contributed equally to this work.