# A Distributed Gaussian Process Model for Multi-Robot Mapping

arXiv: https://arxiv.org/abs/2603.07351 · HTML: https://mlsanity.com/paper/2603.07351/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 24 references that are in the map.
- First posted on arXiv 7 Mar 2026.

## Builds on (most cited first, 9 of 9)

- [Loopy Belief Propagation for Approximate Inference: An Empirical Study](https://mlsanity.com/paper/1301.6725/) (2013)
- [Convergence of Sparse Variational Inference in Gaussian Processes Regression](https://mlsanity.com/paper/2008.00323/) (2020)
- [FutureMapping 2: Gaussian Belief Propagation for Spatial AI](https://mlsanity.com/paper/1910.14139/) (2019)
- [DiNNO: Distributed Neural Network Optimization for Multi-Robot Collaborative Learning](https://mlsanity.com/paper/2109.08665/) (2021)
- [A Robot Web for Distributed Many-Device Localisation](https://mlsanity.com/paper/2202.03314/) (2022)
- [Di-NeRF: Distributed NeRF for Collaborative Learning with Relative Pose Refinement](https://mlsanity.com/paper/2402.01485/) (2024)
- [A Distributed Multi-Robot Framework for Exploration, Information Acquisition and Consensus](https://mlsanity.com/paper/2310.01930/) (2023)
- [Distributed NeRF Learning for Collaborative Multi-Robot Perception](https://mlsanity.com/paper/2409.20289/) (2024)
- [RAMEN: Real-time Asynchronous Multi-agent Neural Implicit Mapping](https://mlsanity.com/paper/2502.19592/) (2025)

## Abstract (arXiv)

We propose DistGP: a multi-robot learning method for collaborative learning of a global function using only local experience and computation. We utilise a sparse Gaussian process (GP) model with a factorisation that mirrors the multi-robot structure of the task, and admits distributed training via Gaussian belief propagation (GBP). Our loopy model outperforms Tree-Structured GPs citebui2014tree and can be trained online and in settings with dynamic connectivity. We show that such distributed, asynchronous training can reach the same performance as a centralised, batch-trained model, albeit with slower convergence. Last, we compare to DiNNO citeyu2022dinno, a distributed neural network (NN) optimiser, and find DistGP achieves superior accuracy, is more robust to sparse communication and is better able to learn continually.