Visually servoed pickup of stationary objects with a kinematically controlled manipulator

• Main Authors: Kai Wang, Fuquan Dai, Jize Li, Shaofeng Zeng
• Format: Conference Proceeding
• Language: English
• Subjects: Cameras; End effectors; Estimation; Visual servoing
• ISSN: 2161-8089
• DOI: 10.1109/COASE.2016.7743414

The visually servoed pickup of stationary objects (or regulation, etc.) has attracted many academic researchers and industrial engineers over the past two decades, and numerous control strategies are proposed for this challenging problem. The controller proposed in this paper follows the Position Based Visual Servoing (PBVS) paradigm. Most of the PBVS controllers, if not all, are designed assuming that the position of the manipulator end effector and that of the stationary object have been accurately recovered with the vision-based estimators. Therefore the rigorous stabilities of such controller/estimator strategies cannot be promised due to the separation principle does not hold for the general nonlinear systems. Accordingly, a new PBVS controller is proposed in this paper for the visually servoed pickup of stationary objects by embedding two novel adaptive estimators into this new controller to estimate the position of the manipulator end effctor and that of the stationary object online. It is proved by Lyapunov theory that this adaptive controller/estimator strategy gives rise to the asymptotic pickup of stationary objects and the convergence of the estimated positions to the actual ones. An experiment is conducted to validate the effectiveness of the proposed controller/estimator strategy.