GRAND-SLAM: Local Optimization for Globally Consistent Large-Scale Multi-Agent Gaussian SLAM

2025-06-23

AI Summary

Existing 3D Gaussian Splatting SLAM methods are confined to small-scale indoor scenes and struggle to scale to large-scale outdoor multi-robot environments. To address this, this paper introduces 3D Gaussian lattices into a large-scale collaborative multi-robot SLAM framework for the first time. We propose a joint mechanism integrating implicit tracking, subgraph-local optimization, and cross-robot loop closure detection, embedded within a pose-graph optimization pipeline to achieve globally consistent dense reconstruction. Key contributions include: (1) a scalable distributed Gaussian representation; (2) a lightweight implicit tracking module enhancing robustness; and (3) a multi-level loop closure detection strategy ensuring global consistency. Experiments demonstrate a 28% PSNR improvement on the Replica dataset and a 91% reduction in multi-robot localization error on the large-scale outdoor Kimera-Multi dataset, with rendering quality significantly surpassing state-of-the-art methods.

Technology Category

• Intelligent Robots: Localization, Mapping, and Navigation
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Computer Vision: Motion & Tracking

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

3D Gaussian splatting has emerged as an expressive scene representation for RGB-D visual SLAM, but its application to large-scale, multi-agent outdoor environments remains unexplored. Multi-agent Gaussian SLAM is a promising approach to rapid exploration and reconstruction of environments, offering scalable environment representations, but existing approaches are limited to small-scale, indoor environments. To that end, we propose Gaussian Reconstruction via Multi-Agent Dense SLAM, or GRAND-SLAM, a collaborative Gaussian splatting SLAM method that integrates i) an implicit tracking module based on local optimization over submaps and ii) an approach to inter- and intra-robot loop closure integrated into a pose-graph optimization framework. Experiments show that GRAND-SLAM provides state-of-the-art tracking performance and 28% higher PSNR than existing methods on the Replica indoor dataset, as well as 91% lower multi-agent tracking error and improved rendering over existing multi-agent methods on the large-scale, outdoor Kimera-Multi dataset.

Problem

• Extends Gaussian SLAM to large-scale multi-agent outdoor environments
• Improves tracking and rendering in collaborative multi-agent SLAM
• Addresses limitations of current methods in scalability and accuracy

Innovation

• Local optimization for submap tracking
• Inter-intra robot loop closure integration
• Pose-graph optimization framework enhancement