Neural adaptive control of fractional-order nonlinear systems subject to actuator faults and input saturation: Application to a single-link manipulator

This article addresses the challenge of neural network-based adaptive control for fractional-order nonlinear systems in nonstrict-feedback form, subject to actuator faults and input saturation. Radial basis function neural networks are integrated into the recursive design to approximate unknown nonlinear dynamics of the system. To enhance robustness against actuator faults, a dedicated fault compensation mechanism is incorporated, where actuator fault severity is defined as the reduction in actuator effectiveness relative to normal operation, providing a clear and quantitative measure of the fault's impact, regardless of the number or type of faults. Moreover, a smooth nonaffine function is employed to effectively capture the nonlinearity caused by input saturation. Fractional-order adaptive laws are developed, and closed-loop stability is rigorously established using the fractional-order Lyapunov criterion. The proposed robust adaptive control strategy ensures reliable performance, and its effectiveness is validated through a practical example.