eth_agent_affordances

FeaturesDict({
    'episode_metadata': FeaturesDict({
        'file_path': Text(shape=(), dtype=string),
        'input_point_cloud': Tensor(shape=(10000, 3), dtype=float16, description=Point cloud (geometry only) of the object at the beginning of the episode (world frame) as a numpy array (10000,3).),
    }),
    'steps': Dataset({
        'action': Tensor(shape=(6,), dtype=float32, description=Robot action, consists of [end-effector velocity (v_x,v_y,v_z,omega_x,omega_y,omega_z) in world frame),
        'discount': Scalar(shape=(), dtype=float32, description=Discount if provided, default to 1.),
        'is_first': bool,
        'is_last': bool,
        'is_terminal': bool,
        'language_embedding': Tensor(shape=(512,), dtype=float32, description=Kona language embedding. See https://tfhub.dev/google/universal-sentence-encoder-large/5),
        'language_instruction': Text(shape=(), dtype=string),
        'observation': FeaturesDict({
            'image': Image(shape=(64, 64, 3), dtype=uint8, description=Main camera RGB observation. Not available for this dataset, will be set to np.zeros.),
            'input_point_cloud': Tensor(shape=(10000, 3), dtype=float16, description=Point cloud (geometry only) of the object at the beginning of the episode (world frame) as a numpy array (10000,3).),
            'state': Tensor(shape=(8,), dtype=float32, description=State, consists of [end-effector pose (x,y,z,yaw,pitch,roll) in world frame, 1x gripper open/close, 1x door opening angle].),
        }),
        'reward': Scalar(shape=(), dtype=float32, description=Reward if provided, 1 on final step for demos.),
    }),
})

@inproceedings{schiavi2023learning,
  title={Learning agent-aware affordances for closed-loop interaction with articulated objects},
  author={Schiavi, Giulio and Wulkop, Paula and Rizzi, Giuseppe and Ott, Lionel and Siegwart, Roland and Chung, Jen Jen},
  booktitle={2023 IEEE International Conference on Robotics and Automation (ICRA)},
  pages={5916--5922},
  year={2023},
  organization={IEEE}
}