Agile Catching with Whole-Body MPC and Blackbox Policy Learning
Abstract
We address a benchmark task in agile robotics: catching objects thrown at high-speed. This is a challenging task that involves tracking, intercepting, and cradling a thrown object with access only to visual observations of the object and the proprioceptive state of the robot, all within a fraction of a second. We present the relative merits of two fundamentally different solution strategies: (i) Model Predictive Control using accelerated constrained trajectory optimization, and (ii) Reinforcement Learning using zeroth-order optimization. We provide insights into various performance tradeoffs including sample efficiency, sim-to-real transfer, robustness to distribution shifts, and whole-body multimodality via extensive on-hardware experiments. We conclude with proposals on fusing "classical" and "learning-based" techniques for agile robot control.
Cite
Text
Abeyruwan et al. "Agile Catching with Whole-Body MPC and Blackbox Policy Learning." Proceedings of The 5th Annual Learning for Dynamics and Control Conference, 2023.Markdown
[Abeyruwan et al. "Agile Catching with Whole-Body MPC and Blackbox Policy Learning." Proceedings of The 5th Annual Learning for Dynamics and Control Conference, 2023.](https://mlanthology.org/l4dc/2023/abeyruwan2023l4dc-agile/)BibTeX
@inproceedings{abeyruwan2023l4dc-agile,
title = {{Agile Catching with Whole-Body MPC and Blackbox Policy Learning}},
author = {Abeyruwan, Saminda and Bewley, Alex and Boffi, Nicholas Matthew and Choromanski, Krzysztof Marcin and D’Ambrosio, David B and Jain, Deepali and Sanketi, Pannag R and Shankar, Anish and Sindhwani, Vikas and Singh, Sumeet and Slotine, Jean-Jacques and Tu, Stephen},
booktitle = {Proceedings of The 5th Annual Learning for Dynamics and Control Conference},
year = {2023},
pages = {851-863},
volume = {211},
url = {https://mlanthology.org/l4dc/2023/abeyruwan2023l4dc-agile/}
}