AIMY: An Open-source Table Tennis Ball Launcher for Versatile and High-fidelity Trajectory Generation

To approach the level of advanced human players in table tennis with robots, generating varied ball trajectories in a reproducible and controlled manner is essential. Current ball launchers used in robot table tennis either do not provide an interface for automatic control or are limited in their capabilities to adapt speed, direction, and spin of the ball. For these reasons, we present AIMY, a three-wheeled open- hardware and open-source table tennis ball launcher, which can generate ball speeds and spins of up to 15.4 ms−1 and 192.0 s−1, respectively, which are comparable to advanced human players. The wheel speeds, launch orientation and time can be fully controlled via an open Ethernet or Wi-Fi interface. We provide a detailed overview of the core design features, and open-source the software to encourage distribution and duplication within and beyond the robot table tennis research community. We also extensively evaluate the ball launcher’s accuracy for different system settings and learn to launch a ball to desired locations. With this ball launcher, we enable long-duration training of robot table tennis approaches where the complexity of the ball trajectory can be automatically adjusted, enabling large-scale real-world online reinforcement learning for table tennis robots.

A rich dataset was included as part of the evaluation of AIMY and the training of the target shooting algorithm.

Accuracy evaluation data set:

• Accuracy evaluation of TTmatic 303A for four different target distances.
• Accuracy evaluation of AIMY with MN4004 KV300 for four different target distances.
• Accuracy evaluation of AIMY with MN5008 KV170 for four different target distances.

System parameter evaluation data set:

• Accuracy evaluation of AIMY with MN5008 KV170 for different stroke gain values.
• Accuracy evaluation of AIMY with MN5008 KV170 for different ball pinching settings.
• Accuracy evaluation of AIMY with MN5008 KV170 with and without prior orientation jump.
• Accuracy evaluation of AIMY with MN5008 KV170 for different ramp-up times.
• Accuracy evaluation of AIMY with MN5008 KV170 for different stroke gain values.
• Accuracy evaluation of AIMY with MN5008 KV170 with the best evaluated system parameters.

Grid search data set:

• Training data based on grid search without trajectories missing the table.
• Training data based on grid search with trajectories missing the table.

Multiple trajectories (usually belonging to one measurement series) are stored with the file format HDF5 (Hierarchical Data Format). HDF5 allows efficient storing of large amounts of data and grouping of data sets. In our experiments, we stored all data in one group called "originals". For easier usage of the data, a small Python script is attached. All trajectories are filtered and roughly visually inspected before evaluation and publication. For more information about the filter process, checkout the appendix of the paper on arXiv.