Table tennis robot with stereo vision and humanoid manipulator I: System design and learning-based batting decision

Abstract

In this paper, a 7-DoF robot table tennis system is presented, which adopts a stereo vision system as its perceptional sensor and a humanoid robot arm as its manipulator. The batting policy, which can return various incoming balls to a desired location, is learned through empirical data based on e-support vector regression(e-SVR). Two experiments, playing with a launcher machine and rallying with a human player, were carried out and the results verified the effectiveness of the proposed approach.