A Sample-Efficient Adjustment-Learning (SEAL) Method for Table Tennis Robot Serve
Table tennis robots have significantly advanced in intelligence owing to the rapid progress in deep learning and reinforcement learning technologies.
However, these advancements often require a large number of training samples, and research focused on the robot serve task remains relatively limited.
In response to these problems, this paper introduces a sample-efficient adjustment-learning (SEAL) method for the serve task inspired by human experience in table tennis, which can inherently augment the available training samples without the need for additional sample collection.
The adjustment learning demonstrates superior performance in planning and prediction tasks, and can adapt to different serve styles with rapid convergence to the target position within a few iterations.
In addition, the random average method during dataset making stage is introduced, and the effectiveness of co-learning in joint space and Cartesian space is also demonstrated.
Movie
If you want to use the original video, please send an e-mail for copyright permission to yamakawalabo
