Understanding and Imitating Human-Robot Motion with Restricted Visual Fields