MR-SLAM lets operators supervise multi-robot fleets via mixed reality
A Meta Quest 3 gives real-time control of three robots at 8.83 Hz scanning.
MR-SLAM tackles the scalability problem of monitoring multi-robot SLAM (simultaneous localization and mapping) through traditional 2D interfaces. The system uses a Meta Quest 3 headset to provide a passthrough view with real-world occlusion, allowing the operator to teleoperate up to three simulated TurtleBot3 robots while spatially anchored dashboard panels show mapping progress in situ. Each robot runs an independent SLAM Toolbox instance, with occupancy grids merged in real time on a ROS 2 back end. Across five 9-minute sessions, the system delivered scans at 8.83 ± 0.16 Hz, mapping 17.9 ± 0.8 m² of merged occupancy, and reached 94.7 ± 0.5% cross-instance occupancy consistency. Additional benchmarks recorded 6.3 ms median transform jitter and 26.7 m² coverage of a 41 m² grid.
Accepted to the ICRA 2026 MM-SpatialAI workshop, MR-SLAM is positioned as a reference implementation for combining passthrough mixed reality supervision with multi-robot SLAM on consumer hardware. The research demonstrates that commercial VR headsets like the Quest 3 can effectively replace expensive custom setups for spatial awareness tasks in building inspection, warehouse monitoring, and other industrial robotics applications. By offloading spatial reasoning to the human operator via an intuitive 3D interface, the system reduces cognitive load while maintaining high mapping fidelity and low latency.
- Uses Meta Quest 3 for passthrough mixed reality with real-world occlusion and spatially anchored dashboards.
- Three TurtleBot3 robots run SLAM Toolbox on ROS 2, merging occupancy grids in real time at 8.83 Hz.
- Achieved 94.7% cross-robot occupancy consistency and 6.3 ms median transform jitter over a 41 m² grid.
Why It Matters
MR-SLAM shows consumer mixed reality headsets can drastically simplify multi-robot mapping for industrial inspection and logistics.