Compact Keyframe-Optimized Multi-Agent Gaussian Splatting SLAM

arXiv cs.RO / 4/2/2026

💬 OpinionSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper introduces a compact, keyframe-optimized multi-agent RGB-D Gaussian Splatting SLAM framework aimed at enabling efficient 3D mapping and map exchange over limited-bandwidth communication links.
It reduces communication load by adding a compaction step that removes redundant 3D Gaussians without degrading rendering quality.
For loop closure, the method performs centralized loop-closure computation without needing an initial guess, supporting two operation modes: rendered-depth only and camera-depth (with lightweight depth images).
In the camera-depth mode, additional Gaussian pruning improves registration accuracy while further cutting transmitted data.
Experiments on synthetic and real-world datasets show a reported 85–95% reduction in transmitted data versus prior state-of-the-art methods, with code released publicly.

Abstract

Efficient multi-agent 3D mapping is essential for robotic teams operating in unknown environments, but dense representations hinder real-time exchange over constrained communication links. In multi-agent Simultaneous Localization and Mapping (SLAM), systems typically rely on a centralized server to merge and optimize the local maps produced by individual agents. However, sharing these large map representations, particularly those generated by recent methods such as Gaussian Splatting, becomes a bottleneck in real-world scenarios with limited bandwidth. We present an improved multi-agent RGB-D Gaussian Splatting SLAM framework that reduces communication load while preserving map fidelity. First, we incorporate a compaction step into our SLAM system to remove redundant 3D Gaussians, without degrading the rendering quality. Second, our approach performs centralized loop closure computation without initial guess, operating in two modes: a pure rendered-depth mode that requires no data beyond the 3D Gaussians, and a camera-depth mode that includes lightweight depth images for improved registration accuracy and additional Gaussian pruning. Evaluation on both synthetic and real-world datasets shows up to 85-95\% reduction in transmitted data compared to state-of-the-art approaches in both modes, bringing 3D Gaussian multi-agent SLAM closer to practical deployment in real-world scenarios. Code: https://github.com/lemonci/coko-slam