TartanGround is a large-scale, multi-modal dataset for ground robot perception and navigation. It was created by theairlabcmu and collected across 63 photorealistic simulation environments. The dataset was last updated on October 14, 2025.
Use Cases
- Training visual perception models based on data from diverse simulated environments.
- Developing navigation algorithms based on multi-modal sensor streams.
- Benchmarking robotic autonomy performance across indoor and outdoor simulation categories.
Strengths
- Data collected across 63 distinct simulation environments.
- Designed for multi-modal robotic tasks.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count and file formats are unknown, which may limit suitability assessment.
Provenance
- Source
- theairlabcmu
- Collection Method
- Collected across photorealistic simulation environments.
- Time Range
- null
- Freshness
- Last updated 2025-10-14 17:35:06; freshness should be verified.
- Geography
- null