A dynamic and uncalibrated method to visually servo-control elastic deformations by fully-constrained robotic grippers

In this paper, we address the set-point deformation control of elastic objects by fully-constrained grippers. We propose an uncalibrated Lyapunov-based algorithm that iteratively estimates the deformation Jacobian matrix, with no prior knowledge of the deformation and camera models. With this new method we show how, by combining pose information of the grippers with several visual measurements, we can independently control elastic deformations of unknown objects. We report experiments with a 6-DOF robot manipulator to validate this control approach.