Dataset Access
The OpenRoboCare dataset is available in multiple formats to suit different research needs. You can download specific modalities, individual OT demonstrator data, or request access to the full dataset.
Online Preview
Subset Download
A curated subset of the OpenRoboCare dataset is available for immediate download. This includes representative data from multiple modalities to help you get started.
Download Subset
Full Dataset Download
The complete OpenRoboCare dataset will be openly accessible on the AWS cloud through its Open Data Sponsorship Program. We are finalizing data hosting details to provide easier access and streamlined download instructions. In the meantime, contact the research team to receive early access.
Request Access
To access the full dataset, please contact the team through the project website. Full hosting details will be available soon.
Download Options
By Modality
Download specific sensor types across all occupational therapists:
- RGB Images - Anonymized camera footage from 3 viewpoints
- Depth Maps - Corresponding depth information
- Gaze Data - Eye-tracking coordinates and vectors
- Egocentric Videos - Head-mounted camera perspective from Pupil Labs eye tracker
- Third-Person Videos - GoPro camera footage
- Tactile Data - 44-sensor pressure readings
- 2D Pose Annotations - Keypoint annotations for pose estimation
- 3D Pose Tracking - Motion capture data for manikin and caregiver
- Action Labels - Temporal annotations of caregiving activities
By Occupational Therapist
Download all modalities for specific therapists (OT 1-21) to study individual techniques and approaches.
By Task
Download data for specific caregiving tasks (1-15) across multiple therapists to analyze task-specific patterns.
Full Dataset
Complete multi-modal dataset including all therapists, tasks, and modalities.
License Information
The OpenRoboCare dataset is released under the BSD-3-Clause license. By accessing the dataset, you agree to:
- Use the data solely for research and educational purposes
- Not attempt to re-identify anonymized individuals
- Cite the dataset in any publications using the data
- Not redistribute the dataset without permission
- Respect the privacy of all participants
Citation
Please cite the dataset using:
@inproceedings{Liang2025OpenRoboCare,
title={{OpenRoboCare: A Multimodal Multi-Task Expert Demonstration Dataset for Robot Caregiving}},
author={Xiaoyu Liang and Ziang Liu and Kelvin Lin and Edward Gu and Ruolin Ye and Tam Nguyen and Cynthia Hsu and Zhanxin Wu and Xiaoman Yang and Christy Sum Yu Cheung and Harold Soh and Katherine Dimitropoulou and Tapomayukh Bhattacharjee},
booktitle={IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
year={2025},
url={https://emprise.cs.cornell.edu/robo-care/}
}Support
For questions about dataset access or technical support, please visit the project website or contact the research team directly.