Open-H-Embodiment: A Large-Scale Dataset for Enabling Foundation Models in Medical Robotics
The current reference point for multi-institution, multi-robot video–kinematics data in healthcare robotics.
Living bibliography
Research Library
Not a paper dump. Each work is read through the same clinical-action lens: what the system observes, what it can do, how the patient responds, and which outcome matters.
08 worksState · Action · Response · Outcome
GR00T-H: A Foundation Vision-Language-Action Model for Healthcare Robotics
Shows how heterogeneous healthcare robot demonstrations can be normalized into a shared action representation.
ImitateCholec: A Multimodal Dataset for Long-Horizon Imitation Learning
A useful example of task-segmented video, kinematics, successful demonstrations, and recovery maneuvers.
MiDAS: A Multimodal Data Acquisition System for Robot-Assisted Surgery
A platform-agnostic way to capture hand, foot, video, and interaction signals without proprietary telemetry.
A Deep Learning Model for Assistive Decision-Making During Robot-Aided Rehabilitation
Learns when assistance is needed from therapist demonstrations and adapts that judgment to new users.
Collaborative AI for Precision Neurorehabilitation: A Roadmap
Frames patient-specific treatment policies around longitudinal models, constraints, and human collaboration.
Kinematic Data Profile and Clinical Outcomes in Robotic Inguinal Hernia Repairs
An early attempt to connect robotic kinematic indicators with patient-reported postoperative outcomes.
Autonomic Robotic Ultrasound Imaging System Based on Reinforcement Learning
A clear physical-action problem: infer probe motion from image state to autonomously acquire a target view.