{
  "name": "CafeTac",
  "url": [
    "https://huggingface.co/datasets/cheng000/CafeTac"
  ],
  "version": "1.1.0",
  "description": "CafeTac is a four-task robotic manipulation dataset with 110 complete episodes and 120321 timesteps. It records seven arm joints including base rotation and one gripper joint in radians, head and wrist RGB views, and two Xense pads on the left and right fingers of one gripper. XenseRGB denotes tactile RGB images; XenseDepth denotes depth matrices; XenseGrid denotes three-channel force-grid matrices. Tasks are hard-drive removal, light-bulb placement, marker placement, and pipe-clamp transfer.",
  "type": "gripper",
  "sensor_image": [
    "XenseRGB"
  ],
  "sensor_array": [
    "XenseDepth",
    "XenseGrid"
  ],
  "sensor_state": [],
  "task": 4,
  "trajectory": 110,
  "frame": 120321,
  "label": [],
  "contact": [
    "pancheng000@gmail.com"
  ],
  "reference": {
    "paper": "Robot-setup reference only (no dedicated CafeTac dataset paper): Online Imitation Learning for Manipulation via Decaying Relative Correction through Teleoperation (2025), arXiv:2503.15368. Paper authors: Cheng Pan, Hung Hon Cheng, Josie Hughes.",
    "website": "https://arxiv.org/abs/2503.15368",
    "authors": "CafeTac dataset authors: Cheng Pan; Hung Hon Cheng (Paul); Yanyuan Qiao; Dacheng Zhou; Josie Hughes"
  },
  "code_trigger": "FTP_1_Source"
}