OpenRoboCare Multi-Modal Expert Demonstration Dataset for Robot-Assisted Caregiving

A comprehensive multimodal dataset capturing real-world caregiving routines from 21 occupational therapists performing 15 daily caregiving tasks. The dataset includes synchronized RGB-D video, tactile sensing, eye-gaze tracking, pose annotations, and action labels across 315 sessions totaling 19.8 hours of expert demonstrations. Data modalities include anonymized RGB images, depth maps, 44-sensor tactile readings, 2D/3D pose tracking, temporal action annotations, and first/third-person videos, enabling research in robot learning from demonstration, multimodal perception, and safe human-robot i...本数据集为多模态综合数据集，收录了21名作业治疗师（Occupational Therapists）执行15项日常照护任务过程中的真实场景照护日常流程。该数据集涵盖315段会话（总时长19.8小时的专家演示内容）中的同步RGB-D视频、触觉传感数据、眼动追踪数据、姿态标注与动作标签。数据模态包含匿名化RGB图像、深度图、44传感器触觉读数、2D/3D姿态追踪结果、时序动作标注以及第一/第三人称视频，可支撑演示学习（Robot Learning from Demonstration）、多模态感知以及安全人机交互（原文截断为i...）等方向的研究。