Black-Box vs. Gray-Box: A Case Study on Learning Table Tennis Ball Trajectory Prediction with Spin and Impacts

Jan Achterhold, Philip Tobuschat, Hao Ma, Dieter Büchler, Michael Muehlebach, Joerg Stueckler
Proceedings of The 5th Annual Learning for Dynamics and Control Conference, PMLR 211:878-890, 2023.

Abstract

In this paper, we present a method for table tennis ball trajectory filtering and prediction. Our gray-box approach builds on a physical model. At the same time, we use data to learn parameters of the dynamics model, of an extended Kalman filter, and of a neural model that infers the ball’s initial condition. We demonstrate superior prediction performance of our approach over two black-box approaches, which are not supplied with physical prior knowledge. We demonstrate that initializing the spin from parameters of the ball launcher using a neural network drastically improves long-time prediction performance over estimating the spin purely from measured ball positions. An accurate prediction of the ball trajectory is crucial for successful returns. We therefore evaluate the return performance with a pneumatic artificial muscular robot and achieve a return rate of 29/30 (97.7%).

Cite this Paper

BibTeX
@InProceedings{pmlr-v211-achterhold23a, title = {Black-Box vs. Gray-Box: A Case Study on Learning Table Tennis Ball Trajectory Prediction with Spin and Impacts}, author = {Achterhold, Jan and Tobuschat, Philip and Ma, Hao and B\"uchler, Dieter and Muehlebach, Michael and Stueckler, Joerg}, booktitle = {Proceedings of The 5th Annual Learning for Dynamics and Control Conference}, pages = {878--890}, year = {2023}, editor = {Matni, Nikolai and Morari, Manfred and Pappas, George J.}, volume = {211}, series = {Proceedings of Machine Learning Research}, month = {15--16 Jun}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v211/achterhold23a/achterhold23a.pdf}, url = {https://proceedings.mlr.press/v211/achterhold23a.html}, abstract = {In this paper, we present a method for table tennis ball trajectory filtering and prediction. Our gray-box approach builds on a physical model. At the same time, we use data to learn parameters of the dynamics model, of an extended Kalman filter, and of a neural model that infers the ball’s initial condition. We demonstrate superior prediction performance of our approach over two black-box approaches, which are not supplied with physical prior knowledge. We demonstrate that initializing the spin from parameters of the ball launcher using a neural network drastically improves long-time prediction performance over estimating the spin purely from measured ball positions. An accurate prediction of the ball trajectory is crucial for successful returns. We therefore evaluate the return performance with a pneumatic artificial muscular robot and achieve a return rate of 29/30 (97.7%).} }
Endnote
%0 Conference Paper %T Black-Box vs. Gray-Box: A Case Study on Learning Table Tennis Ball Trajectory Prediction with Spin and Impacts %A Jan Achterhold %A Philip Tobuschat %A Hao Ma %A Dieter Büchler %A Michael Muehlebach %A Joerg Stueckler %B Proceedings of The 5th Annual Learning for Dynamics and Control Conference %C Proceedings of Machine Learning Research %D 2023 %E Nikolai Matni %E Manfred Morari %E George J. Pappas %F pmlr-v211-achterhold23a %I PMLR %P 878--890 %U https://proceedings.mlr.press/v211/achterhold23a.html %V 211 %X In this paper, we present a method for table tennis ball trajectory filtering and prediction. Our gray-box approach builds on a physical model. At the same time, we use data to learn parameters of the dynamics model, of an extended Kalman filter, and of a neural model that infers the ball’s initial condition. We demonstrate superior prediction performance of our approach over two black-box approaches, which are not supplied with physical prior knowledge. We demonstrate that initializing the spin from parameters of the ball launcher using a neural network drastically improves long-time prediction performance over estimating the spin purely from measured ball positions. An accurate prediction of the ball trajectory is crucial for successful returns. We therefore evaluate the return performance with a pneumatic artificial muscular robot and achieve a return rate of 29/30 (97.7%).
APA
Achterhold, J., Tobuschat, P., Ma, H., Büchler, D., Muehlebach, M. & Stueckler, J.. (2023). Black-Box vs. Gray-Box: A Case Study on Learning Table Tennis Ball Trajectory Prediction with Spin and Impacts. Proceedings of The 5th Annual Learning for Dynamics and Control Conference, in Proceedings of Machine Learning Research 211:878-890 Available from https://proceedings.mlr.press/v211/achterhold23a.html.

Related Material