Learning Hybrid Position/Force Control and Its Application

This paper combines the robot position/force hybrid control algorithm proposed by Raibert and Craig and the robot dynamic hybrid control algorithm proposed by Yoshikawa. Taking account of the difficulty to identify the environment, especially the effect of mechanical disturbance, and the vibrations of mechanism, a PI learning hybrid control algorithm for using robot to do assembly task was presented. This algorithm was realized in a peg insertion by a three DOF robot. The experiment results showed that this algorithm could effectly realize the peg-and-hole task, and restrain and weaken the vibrations of mechanism.