For interacting with objects, it is necessary for the robot to have vision.  The program depicted in the embedded video uses colour based object recognition, allowing objects of specific colours to be segmented from a video feed.