Skeleton based gesture to control manipulator

Abstract

The key aspect to control the motion of manipulator intuitively is ability to map both joints on human and machine. This research proposes a method to interact with arm robot by using gesture based on skeleton image of user that is read from Kinect sensor. Velocity of each joint is defined by measuring angle rotation of every bone with respect to previous linked bone. Experiments has been carried out to manipulate SCORBOT-ER 9Pro with ability to handle 4 of 5 axises.