Research Analyzer
← Back ICRA 2026

COSMO-Bench: A Benchmark for Collaborative SLAM Optimization

Daniel McGann, Easton Potokar, Michael Kaess

PDF

AI summary

Key figure (auto-extracted from paper)
The authors release COSMO-Bench, a suite of 24 realistic, open-source benchmark datasets that enable standardized evaluation of distributed multi-robot SLAM algorithms.
Collaborative SLAM Multi-robot benchmark Distributed optimization LiDAR SLAM Open-source dataset Communication modeling

Problem

Research on distributed optimization algorithms for multi-robot collaborative SLAM is hindered by a lack of standard, realistic benchmark datasets, forcing researchers to rely on unrealistic simulations or artificially partitioned single-robot data.

Approach

The authors derive 24 benchmark datasets by temporally synchronizing independent single-robot LiDAR trials from real-world environments, processing them through a baseline C-SLAM front-end, and simulating inter-robot communication using a data-driven network model.

Key results

  • Release of 24 open-source benchmark datasets derived from real-world LiDAR data
  • Development of a realistic communication model simulating inter-robot connectivity and bandwidth
  • Provision of high-quality reference solutions, temporal metadata, and outlier classifications
  • Validation that synchronized single-robot trials effectively simulate realistic multi-robot C-SLAM scenarios

Why it matters

It provides the robotics community with a standardized, realistic, and reproducible evaluation framework to accelerate the development and comparison of distributed multi-robot SLAM algorithms.

Abstract

Recent years have seen a focus on research into distributed optimization algorithms for multi-robot Collab- orative Simultaneous Localization and Mapping (C-SLAM). Research in this domain, however, is made difficult by a lack of standard benchmark datasets. Such datasets have been used to great effect in the field of single-robot SLAM, and researchers focused on multi-robot problems would benefit greatly from dedicated benchmark datasets. To address this gap, we de- sign and release the Collaborative Open-Source Multi-robot Optimization Benchmark (COSMO-Bench) – a suite of 24 datasets derived from a baseline C-SLAM front-end and real- world LiDAR data. Data DOI: 10.1184/R1/29652158. This work was partially supported by NASA award 80NSSC24CA020 and the NSF Graduate Research Fellowship Program. The authors are with the Robotics Institute, Carnegie Mellon University, Pittsburgh, PA, USA. {danmcgann, potokar, kaess}@cmu.edu 2026 IEEE International Conference on Robotics and Automation (ICRA 2026) June 1-5, 2026. Vienna, Austria 979-8-3315-8160-2/26/$31.00 ©2026 IEEE 8200

Index terms

Multi-Robot SLAM SLAM Multi-Robot Systems

Related papers