M2DGR: Multi-Modal and Multi-Scenario Ground Robot SLAM Data

SJTU-ViSYS developed M2DGR, a multi-modal and multi-scenario dataset for ground robot navigation, published in RA-L 2021 and ICRA 2022. It provides synchronized sensor data across diverse environments to support Simultaneous Localization and Mapping (SLAM) research.

Use Cases

• Benchmarking SLAM algorithms using multi-modal sensor inputs
• Testing localization accuracy across different environmental scenarios
• Developing sensor fusion models for ground robot navigation

Strengths

• Multi-modal sensor synchronization
• Multi-scenario environmental coverage
• Peer-reviewed via RA-L and ICRA publications
• Permissive MIT license

Limitations

• Lack of explicit sensor metadata in the provided description
• Potential high storage requirements typical of multi-modal robotics data

Provenance

Source

SJTU-ViSYS (RA-L 2021 & ICRA 2022)

Collection Method

Sensor-based data collection from ground robots

Freshness

Last updated July 2025; based on research published in 2021-2022.