A vision-based generic dynamic model of PKMs and its experimental validation on the Quattro parallel robot

In this paper, we present a dynamic modeling method for parallel kinematic manipulators. This method is built upon the postures (i.e., 3D orientation vectors, lengths and self rotations) of the kinematic elements which form the kinematic chains of the robot. Regarding the structure of the above method, computer vision emerges as a good option to obtain the postures of these kinematic elements. This method is then validated experimentally on the commercial Adept Quattro parallel robot.