Uncalibrated Image‐Based Visual Servoing of Rigid‐Link Electrically Driven Robotic Manipulators

Abstract

Abstract In this paper, the uncalibrated fixed‐camera visual servoing problem of robot manipulators will be addressed by considering its full motor dynamics. A new adaptive image‐space visual servoing strategy without both the joint and visual acceleration measurements is presented, which can handle uncertainties in the camera intrinsic and extrinsic parameters, the robot kinematic and dynamic parameters, and the motor dynamic parameters. The proposed scheme is developed based on the depth‐independent interaction matrix in order to deal with the nonlinear dependence of image Jacobian matrix on the unknown parameters, which allows the camera and robot kinematic parameters in the closed‐loop dynamics to be linearly parameterized. In this way, adaptive laws for the online estimation of the unknown camera, kinematic, rigid dynamic, and motor dynamic parameters can be developed very efficiently. Furthermore, a joint velocity observer will also be presented to solve the problem without both the joint and visual acceleration measurements. To show asymptotic convergence of image errors, stability analysis based on both the rigid‐link robot dynamics and full motor dynamics will be performed by using Lyapunov theory. Simulation results will be given to validate the performance of the proposed scheme.