Peer-reviewed research details a vision-guided reinforcement-learning controller that coordinates locomotion with a racket arm.