Published February 21, 2022
| Version v1
Dataset
Open
Dataset associated to the "ADHERENT: Learning Human-like Trajectory Generators for Whole-body Control of Humanoid Robots" paper (manuscript DOI: 10.1109/LRA.2022.3141658)
Authors/Creators
- 1. Artificial and Mechanical Intelligence, Istituto Italiano di Tecnologia, Genoa, Italy, and DIAG, Sapienza Università di Roma, Roma, Italy
- 2. Laboratory for Computational and Statistical Learning - IIT@MIT, Istituto Italiano di Tecnologia, Genoa, Italy, and Massachusetts Institute of Technology, Cambridge, MA, USA
- 3. Artificial and Mechanical Intelligence, Istituto Italiano di Tecnologia, Genoa, Italy, and DIBRIS, Università degli Studi di Genova, Genoa, Italy
- 4. Artificial and Mechanical Intelligence, Istituto Italiano di Tecnologia, Genoa, Italy, and Machine Learning and Optimisation, University of Manchester, Manchester, UK
- 5. Artificial and Mechanical Intelligence, Istituto Italiano di Tecnologia, Genoa, Italy
- 6. DIAG, Sapienza Università di Roma, Roma, Italy
- 7. Laboratory for Computational and Statistical Learning - IIT@MIT, Istituto Italiano di Tecnologia, Genoa, Italy, and Massachusetts Institute of Technology, Cambridge, MA, USA, and DIBRIS, Università degli Studi di Genova, Genoa, Italy, and Center for Brains, Minds and Machines, MIT, Cambridge, MA, USA
Description
This dataset contains data accompanying the work:
@ARTICLE{9676410,
author={Viceconte, Paolo Maria and Camoriano, Raffaello and Romualdi, Giulio and Ferigo, Diego and Dafarra, Stefano and Traversaro, Silvio and Oriolo, Giuseppe and Rosasco, Lorenzo and Pucci, Daniele},
journal={IEEE Robotics and Automation Letters},
title={ADHERENT: Learning Human-like Trajectory Generators for Whole-body Control of Humanoid Robots},
year={2022},
volume={7},
number={2},
pages={2779-2886},
doi={10.1109/LRA.2022.3141658}}
The dataset is organized in folders, whose content can be summarized as follows:
- mocap: motion capture data collected from human motion
- retargeted_mocap: motion capture data retargeted on the robot
- IO_features: input and output features extracted from the retargeted mocap data to train the trajectory generator
- training_D2_D3_subsampled_mirrored_4ew_98%: training data
- inference: data collected while generating trajectories
- trajectory_control_simulation: data collected while controlling trajectories in simulation
- trajectory_control_real_robot: data collected while controlling trajectories on the real robot
- additional_figures: additional data to reproduce some figures in the paper and portions of the supplementary video
A more detailed description of the content of each folder is provided in the README.txt file included in the dataset.