Adaptive finite-time tracking control for robotic manipulators with funnel boundary

The finite-time tracking control problem with the output-constraint property of robotic manipulators subjected to system uncertainties is addressed. Specifically, the radial basis function neural network is employed to compensate for system uncertainties. The finite-time stability theorem is used for the backstepping design process, by which the limit of the settling time is set. A funnel boundary is used to limit the output overshoot. The proposed controller guarantees that all the signals are semi-globally practically finite-time bounded, while the tracking errors are enveloped by the funnel boundary. The performance of the proposed control method is illustrated by a numerical simulation of a 3-DOF manipulator. It is shown that the tracking errors are bounded by prescribed funnel boundaries. In the meantime, the manipulator is stabilized within a finite period of time.