Learning to Estimate the Pose of a Peer Robot in a Camera Image by Predicting the States of its LEDs