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

TL;DR

This paper compares black-box and gray-box methods for predicting table tennis ball trajectories with spin and impacts, demonstrating the superiority of the physics-informed gray-box approach in long-term accuracy and robotic return performance.

Contribution

The paper introduces a hybrid gray-box model combining physical principles with data-driven parameter learning for improved trajectory prediction.

Findings

Gray-box approach outperforms black-box methods in prediction accuracy.

Neural network initialization of spin improves long-term predictions.

Achieved 97.7% success rate in robotic ball return.

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%).