| Abstrakt: |
In this paper, a model-free fractional-order prescribed performance finite-time controller (MFF-PPFTC) is proposed for mechatronic systems subject to uncertainties, external disturbances, and actuator failure. The MFF-PPFTC integrates an ultra-local model (ULM)-based time-delay estimation (TDE), a Proportional-Differential (PD) controller, and a fractio-nal-order prescribed performance finite-time controller (FO-PPFTC) that employs a sliding mode disturbance compensation (SMDC). The ULM is utilized to approximate the complex mechatronic system within an extremely short sliding time window, while TDE is used to observe and eliminate the lumped disturbances. A PD controller is then designed to stabilize the closed-loop system, leading to the development of the TDE-based PD (TDE-PD) controller. However, estimation errors can occur using TDE technique, and potentially compromise tracking performance. To avoid this, SMDC is designed to eliminate the estimation error of TDE within finite time. The FO-PPFTC is then developed and integrated into the TDE-PD controller to achieve finite-time stabilization of the closed-loop system and to bound the tracking error within predefined performance limits, allowing to specify convergence time and precision. Finally, the MFF-PPFTC is proposed. The finite-time stability and convergence of the closed-loop system under the MFF-PPFTC are analyzed using the Lyapunov theorem. To validate the proposed method, numerical simulations for a 2-DOF robotic manipulator, and co-simulations for a virtual prototype of a 7-DOF iReHave upper limb exoskeleton, are conducted. The comparative simulation results with a supervising switching technique-based adaptive model-free controller (SST-AMFC) and a robust fault tolerant controller (RFTC) demonstrate the superior trajectory tracking performance of the proposed controller. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink