Neural networks adaptive control for fractional-order non-linear system with unmodelled dynamics and actuator faults

The fault-tolerant control (FTC) problem for fractional-order (FO) non-linear systems with unmodelled dynamics and actuator faults is studied. First, the author uses neural networks (NNs) to identify unknown non-linear functions and apply a FO dynamic signal to control unmodelled dynamics. Then, fractional-order dynamic surface control (FODSC) is introduced in the design process of the adaptive backstepping control algorithm to avoid complex explosion problems. In addition, an adaptive NNs FTC algorithm using the FO Lyapunov stability criterion is designed. Importantly, the author shows that the proposed system is stable, and the tracking error could be converged to a small neighbourhood of zero. Finally, a simulation example is used to verify the effectiveness of the proposed control scheme.