Actuator fault detection and adaptive accommodation control of flexible-joint robots

This study presents a fault detection and adaptive accommodation control scheme for flexible-joint (FJ) robots with model uncertainties and multiple actuator faults. It is assumed that the value and occurrence time of multiple actuator faults are unknown. A fault detection scheme is designed to detect the first actuator fault of uncertain FJ robots and the fault detectability characterising the class of actuator faults is analysed. Then, we design an adaptive fault accommodation control scheme with prescribed performance bounds, which characterise the convergence rate and maximum overshoot of tracking errors to accommodate the actuator faults and to compensate the effect caused by them. The proposed accommodation controller is developed based on the dynamic surface design and the adaptive technique. From the Lyapunov stability theorem, it is proved that all signals of the closed-loop system are semiglobally uniformly ultimately bounded and tracking errors are preserved within prescribed performance bounds regardless of actuator faults that occur after the first fault detection. A simulation example of a three-link FJ robot is provided to illustrate the effectiveness of the proposed scheme.