CoWear synchronizes on-device measurements from a mobile phone, smartwatch, and Rokid glasses with 6-DoF ground-truth trajectories. This release contains 488 aligned sessions, packaged as 11 downloadable tar.zst shards totaling 6.78 GiB compressed. The dataset was authored by zyshe and last updated on July 20, 2026.
Use Cases
- Studying human motion patterns based on synchronized multi-device sensor streams.
- Developing cross-device sensor fusion algorithms based on aligned mobile phone, smartwatch, and glasses data.
- Researching localization techniques based on 6-DoF ground-truth trajectories.
- Benchmarking activity recognition models based on multi-modal wearable sensor data.
Strengths
- 488 aligned sessions provide a substantial collection for analysis.
- Data includes synchronized measurements from three distinct device types (mobile phone, smartwatch, Rokid glasses).
- Ground-truth 6-DoF trajectories offer a precise reference for localization and motion tasks.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
- Source
- huggingface
- Collection Method
- On-device measurements synchronized across multiple wearables.
- Freshness
- Last updated 2026-07-20 15:27:00; freshness should be verified.