Decentralized Pose Graph Riemannian Optimization for Object-based Multi-Robot SLAM

2026-06-23

AI Summary

This work addresses the strong coupling between trajectory and shared object pose estimation in object-level multi-robot SLAM, a challenge exacerbated under sparse, intermittent, or time-varying communication where conventional decentralized methods suffer performance degradation. To overcome this, we propose a fully decentralized Riemannian optimization framework that decouples the joint estimation via a consensus mechanism and introduces, for the first time, a distributed Riemannian approximate Newton method for decentralized pose-graph optimization. Operating on the SE(d) manifold, our approach leverages local second-order information to accelerate convergence and adapts to arbitrary communication topologies. We theoretically prove convergence to a Riemannian first-order stationary point and elucidate the role of second-order information in improving problem conditioning. Extensive experiments on public datasets, large-scale simulations, and real multi-robot systems demonstrate significant gains in accuracy, efficiency, and scalability, with fewer iterations, reduced communication overhead, and strong robustness to communication failures.

Abstract

Pose graph optimization (PGO) is a key back-end component for state estimation in networked multi-robot simultaneous localization and mapping (SLAM). In object-based multi-robot SLAM, the problem becomes more tightly coupled because robots must jointly estimate both their trajectories and the poses of persistent objects observed by multiple agents. Existing decentralized solutions often assume that the communication graph closely matches the physical interaction topology, which is restrictive in realistic deployments where communication is sparse, intermittent, or time-varying. This paper presents a fully decentralized Riemannian optimization framework for object-based multi-robot PGO that decouples the coupled estimation problem via a consensus mechanism, enabling flexible communication topologies. To improve convergence under limited communication budgets, we further develop a distributed approximate-Newton scheme that exploits local second-order information while operating directly on the SE(d) manifold to preserve geometric consistency, and we establish the convergence to Riemannian first-order stationary points and provide a local condition-number analysis explaining the benefit of approximate second-order information over first-order Riemannian descent. The resulting method reduces iteration count and communication overhead without sacrificing estimation accuracy. Extensive evaluations on public benchmarks, large-scale simulations, and real-world multi-robot experiments demonstrate improved accuracy, runtime efficiency, scalability across network topologies, and robustness to communication failures.

Problem

• multi-robot SLAM
• pose graph optimization
• object-based SLAM
• decentralized optimization
• communication constraints

Innovation

• decentralized optimization
• Riemannian manifold
• object-based SLAM
• pose graph optimization
• approximate Newton method