We propose a method of controlling a table tennis robot so as to return the incoming ball to a desired point on the table with specified flight duration. The proposed method consists of the following three input-output maps implemented by means of locally weighted regression: 1) a map for predicting the impact time of the ball hit by the paddle and the ball position and velocity at that moment according to input vectors describing the state of the incoming ball; 2) a map representing a change in ball velocities before and after impact; and 3) a map giving the relation between the ball velocity just after impact and the landing point and time of the returned ball. We also propose a feed-forward control scheme based on iterative learning control to accurately achieve the stroke movement of the paddle as determined by using these maps. Experimental results including rallies with a human opponent are also reported to demonstrate the effectiveness of our approach.

We propose a method of controlling the paddle so as to return the ball to a desired point on the table with a specified flight duration. The proposed method consists of the following three input-output maps implemented by means of Locally Weighted Regression (LWR): (1) A map for predicting the impact time of the ball hit by the paddle and the ball position and velocity at that moment according to input vectors describing the state of the incoming ball, (2) A map representing a change in ball velocities before and after impact and (3) A map giving the relation between the ball velocity just after impact and the landing point and time of the returned ball. We also propose a novel control scheme based on iterative learning control to accurately achieve the stroke movement of the paddle as determined by using these maps.

This paper describes how a table tennis robot with a flat paddle coordinates its movement in order to achieve efficient strokes for any given ball. We propose a method of generating stroke movement based on virtual targets that means the point at which the ball should be struck and the paddle velocity just before hitting the ball. These targets are predicted using inputs-outputs maps implemented efficiently by means of a k-d tree (k dimensional tree). The paddle approaches these targets by using a visual feedback control scheme similar to the mirror law proposed by Koditschek. The results of the implementation are also given to show the effectiveness of the proposed method.

In this paper, we consider ”wall-bouncing ” task in which we repeat the process of hitting the ball that rebounds from the wall and the table as a typical example of dexterous batting tasks. Through the stability analysis of a discrete system, we reveal that required dexterity varies depending on the location of equilibrium points, that is, the shape of the ball trajectory. A feedback control scheme for the paddle at the moment of impact is also proposed to continue ”wall-bouncing ” task under disturbances.