# Policies over Poses: Reinforcement Learning based Distributed Pose-Graph Optimization for Multi-Robot SLAM

arXiv: https://arxiv.org/abs/2510.22740 · HTML: https://mlsanity.com/paper/2510.22740/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in RL for robot navigation and UAV control, one of 15 directions in Deep reinforcement learning methods.
- Builds on 21 of its 37 references that are in the map.
- First posted on arXiv 26 Oct 2025.

## Placed in

- Deep reinforcement learning methods / [RL for robot navigation and UAV control](https://mlsanity.com/deeprl/rl-for-robot-navigation-and-uav-control/)

## Builds on (most cited first, 21 of 21)

- [The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables](https://mlsanity.com/paper/1611.00712/) (2016)
- [Learning Sparse Neural Networks through L0 Regularization](https://mlsanity.com/paper/1712.01312/) (2017)
- [Dynamic Edge-Conditioned Filters in Convolutional Neural Networks on Graphs](https://mlsanity.com/paper/1704.02901/) (2017)
- [Actor-Attention-Critic for Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/1810.02912/) (2018)
- [DeepV2D: Video to Depth with Differentiable Structure from Motion](https://mlsanity.com/paper/1812.04605/) (2018)
- [BA-Net: Dense Bundle Adjustment Network](https://mlsanity.com/paper/1806.04807/) (2018)
- [Kimera-Multi: Robust, Distributed, Dense Metric-Semantic SLAM for Multi-Robot Systems](https://mlsanity.com/paper/2106.14386/) (2021)
- [NeuRoRA: Neural Robust Rotation Averaging](https://mlsanity.com/paper/1912.04485/) (2019)
- [Message Passing Least Squares Framework and its Application to Rotation Synchronization](https://mlsanity.com/paper/2007.13638/) (2020)
- [Distributed Certifiably Correct Pose-Graph Optimization](https://mlsanity.com/paper/1911.03721/) (2019)
- [Differentiable Factor Graph Optimization for Learning Smoothers](https://mlsanity.com/paper/2105.08257/) (2021)
- [Multi-Robot Cooperative Socially-Aware Navigation Using Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2309.15234/) (2023)
- [Decomposed Soft Actor-Critic Method for Cooperative Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2104.06655/) (2021)
- [Graph-CoVis: GNN-based Multi-view Panorama Global Pose Estimation](https://mlsanity.com/paper/2304.13201/) (2023)
- [Majorization Minimization Methods for Distributed Pose Graph Optimization](https://mlsanity.com/paper/2108.00083/) (2021)
- [Multi S-Graphs: An Efficient Distributed Semantic-Relational Collaborative SLAM](https://mlsanity.com/paper/2401.05152/) (2024)
- [RL-PGO: Reinforcement Learning-based Planar Pose-Graph Optimization](https://mlsanity.com/paper/2202.13221/) (2022)
- [DGORL: Distributed Graph Optimization based Relative Localization of Multi-Robot Systems](https://mlsanity.com/paper/2210.01662/) (2022)
- [Rotation Averaging with Attention Graph Neural Networks](https://mlsanity.com/paper/2010.06773/) (2020)
- [SPACE: 3D Spatial Co-operation and Exploration Framework for Robust Mapping and Coverage with Multi-Robot Systems](https://mlsanity.com/paper/2411.02524/) (2024)
- [MGPRL: Distributed Multi-Gaussian Processes for Wi-Fi-based Multi-Robot Relative Localization in Large Indoor Environments](https://mlsanity.com/paper/2506.23514/) (2025)

## Built on by (most cited first, 1 of 1)

- [MARVO: Marine-Adaptive Radiance-aware Visual Odometry](https://mlsanity.com/paper/2511.22860/) (2025)

## Abstract (arXiv)

We consider the distributed pose-graph optimization (PGO) problem, which is fundamental in accurate trajectory estimation in multi-robot simultaneous localization and mapping (SLAM). Conventional iterative approaches linearize a highly non-convex optimization objective, requiring repeated solving of normal equations, which often converge to local minima and thus produce suboptimal estimates. We propose a scalable, outlier-robust distributed planar PGO framework using Multi-Agent Reinforcement Learning (MARL). We cast distributed PGO as a partially observable Markov game defined on local pose-graphs, where each action refines a single edge's pose estimate. A graph partitioner decomposes the global pose graph, and each robot runs a recurrent edge-conditioned Graph Neural Network (GNN) encoder with adaptive edge-gating to denoise noisy edges. Robots sequentially refine poses through a hybrid policy that utilizes prior action memory and graph embeddings. After local graph correction, a consensus scheme reconciles inter-robot disagreements to produce a globally consistent estimate. Our extensive evaluations on a comprehensive suite of synthetic and real-world datasets demonstrate that our learned MARL-based actors reduce the global objective by an average of 37.5% more than the state-of-the-art distributed PGO framework, while enhancing inference efficiency by at least 6X. We also demonstrate that actor replication allows a single learned policy to scale effortlessly to substantially larger robot teams without any retraining. Code is publicly available at https://github.com/herolab-uga/policies-over-poses.