Learning anticipation policies for robot table tennis | IEEE Conference Publication | IEEE Xplore