Learning the inverse map for a robot hitting task

We describe our approach to the robot's Hanetsuki task (Japanese badminton), i.e. to return the incoming ball to the human opponent with a racket. A learning algorithm that consists of updating action commands and smoothing them based on a Gaussian kernel is proposed to compensate for the insufficiency in a non-adaptive model-based approach. Experimental results obtained by using the developed Hanetsuki robot are also shown.