Adaptive Visual Tracking Control of Uncertain Rigid-Link Electrically Driven Robotic Manipulators with an Uncalibrated Fixed Camera

The adaptive image-based trajectory tracking control problem of rigid-link electrically driven (RLED) robotic manipulators is addressed in this paper. A fixed camera configuration is considered and the camera intrinsic and extrinsic parameters are assumed to be unknown. Furthermore, the manipulator dynamic and motor dynamic parameters are assumed to be uncertain and the depths of feature points can be allowed to be time varying. The depth-independent interaction matrix framework is used such that unknown parameters in the closed-loop dynamics can be parameterized linearly and adaptive laws can be derived to estimate them online. By simultaneously taking into consideration the mechanical and electrical subsystem dynamics of RLED robotic manipulators, the backstepping technique is used to design control voltage inputs to guarantee the tracking of desired image trajectories in the presence of parameter uncertainties and time-varying depth information. Asymptotical convergence of image tracking errors to zeros is proved by using Lyapunov stability theory. Simulation results based on a 3-DOFs anthropomorphic robotic manipulator are given to demonstrate the performance of the proposed adaptive visual tracking scheme.