Dynamic Position-Based Visual Servoing of Robotic Manipulators

− This paper discusses dynamic visual servoing and proposes a position-based approach to visual servo control of robotic manipulators. Several dynamic visual servoing control architectures used for position-based are addressed in this paper, in order to motivate the proposed approach. In position-based visual servoing, the features used to control the robot manipulator, actual pose of the object related to the camera frame, are extracted from the 2D image data using a pose estimation algorithm. Using the stated features a Lyapunov stable controller is derived, inspired on classical position-based visual servoing and the Jacobian transpose controller approach. In the experimental setup the image is acquired from a CCD camera mounted on a planar robotic manipulator end-effector, and processed in a standard PC based vision system in order to close the visual servoing loop with Lyapunov stable controllers for dynamic position-based visual servoing. This proposed approach was tested experimentally in a 2 dof robot manipulator, but is also suitable for 6 dof robot manipulators. Comparison results, obtained experimentally, are also presented for the addressed dynamic visual servoing controllers.