In this video, Pi Robot uses his head mounted camera and range sensors to
locate the target (green or red balloon) in 3-dimensional space. The
coordinates of the balloon in the head-centered frame of reference are
then transformed into coordinates relative to reference frames
attached to each shoulder joint. From there we use spherical coordinates to figure out where to point the hands in anticipation of grasping the balloon.
The behavior demonstrated in the video takes place using five
independent threads: head tracking, arm tracking (x2), grasping,
and orienting.
