ABSTRACT
This chapter develops a finite-time fault-tolerant control scheme for a fixed-wing unmanned aerial vehicle in the presence of actuator faults and input saturation by using a fractional-order backstepping iterative design strategy. To compensate for the lumped disturbance induced by the actuator faults, a neural network disturbance observer with finite-time observation capability is first developed as the fault diagnosis unit. Then, based on the diagnosed fault information, fractional-order calculus is artfully utilized to enhance the fault-tolerant control performance within the backstepping design architecture. The salient feature of the developed control scheme is that the finite-time neural network disturbance observer and fractional-order calculus are simultaneously used to significantly increase the fault-tolerant control performance against unexpected actuator faults. Moreover, to address the input saturation problem, the faulty unmanned aerial vehicle dynamics is augmented by an auxiliary system. Furthermore, a Nussbaum function is incorporated into the fault-tolerant control scheme to further avoid the calculation of the inverse gain matrix involved within the auxiliary system. It is shown by Lyapunov analysis that the tracking errors are convergent in finite time. Finally, comparative simulation results and hardware-in-the loop experimental results are presented show the effectiveness of the developed fault-tolerant control scheme.
