A Visual Servoing Approach Applied to Robotic Tasks

Abstract

We are interested in the design of control systems which work in closed loop with regard to the environment, achieving robotic tasks with the use of vision data. We apply a visual servoing approach to robotic tasks consisting in the positioning of the end effector relatively to a priori known 3D objects. The robotic configuration is constituted by a CCD camera coupled with one or two laser stripes mounted on the hand of a manipulator. In order to formalize the problem, we choose the sensor bused control approach [Espiau87], [Samson90], and applied to visual servoing [Rives87b]. [Rives89], [Chaumette90]. We express our approach (camera-light stripe coupling) with this formalism. In our case, the elementary signals are points of discontinuity. In order to demonstrate the feasibility of this approach, we implemented the control algorithm in a robotic environment built in our laboratory. We present simulation and experimental results obtained with two positioning tasks. The obtained results thus prove the robusmess and the stability of the algorithm.

Extracted Key Phrases

Cite this paper

@inproceedings{Urban1990AVS, title={A Visual Servoing Approach Applied to Robotic Tasks}, author={Jean-Philippe Urban and Guy Motyl and Jean Gallice}, booktitle={MVA}, year={1990} }