Abstract
In this paper, we propose a new method addressing the robot-image Jacobian approximation of image-based visual servoing (IBVS) for a redundant manipulator. The robot-image Jacobian is approximated iteratively and is a model free. A linearised model of the robot-image Jacobian is applied, based on the first order Taylor series approximation. A weighted least norm solution is induced in a pseudo inverse computation of the approximated robot-image Jacobian. The resulting control law then can be used for visual servoing tasks with joint limit avoidance capability using both a static target or moving target. The self-motion of the robot joints resolves the redundancy during visual servoing tasks when one or more joints are approaching their joint limits. A design and stability analysis of the proposed method is discussed in this paper. Simulated and real-time experiments using a 7 DOF PowerCube robot manipulator are conducted. The IBVS is configured using a monovision eye-in-hand system configuration. The system behaviour and performances of the proposed method are presented and analysed.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call