Résumé
This paper presents an extended adaptive control scheme via terminal sliding mode (TSM) for cable-driven parallel manipulators (CDPM). Compared with linear hyperplane-based sliding mode control, TSM is able to guarantee high-precision and robust tracking performances which arise from its main feature of finite-time convergence. This motivates applying TSM to robotic manipulators in general and, as presented in this paper, to CDPM in particular. The scheme pre-sented in this paper extends early developed TSM control schemes which are based on partial knowledge of system dynamics. Instead, making use of the property that the dynamic models of mechanical manipulators are linear in inertial parameters, an adaptive control law is synthesised based on an appropriate choice of Lyapunov function which guarantees finite-time convergence to neighborhood of sliding mode. A key challenge of the control of CDPM is that cable tensions must be admissible, i.e. lying in a non-negative range of admissible values. As long as cable tensions are admissible, the overall dynamics of CDPM can be easily written in either ac-tuator space or operational space which in turn facilitates control system design. The extended adaptive control scheme has been applied to a large redundantly ac-tuated CDPR prototype, CoGiRo. Simulation results show the effectiveness of the proposed control method.