Observer-based neural adaptive fixed-time tracking control for multi-input multi-output nonlinear systems with actuator faults

This study developed a new strategy for adaptive fixed-time output feedback control of multi-input multi-output (MIMO) nonlinear systems involving unmeasurable external disturbances and actuator faults. Two actuator faults, loss of effectiveness and lock-in-place were simultaneously considered. Furthermore, a composite observer was used in MIMO systems to estimate unmeasurable states and external disturbances simultaneously, and radial basis function neural networks were employed to identify unknown internal nonlinearities. Additionally, the command-filtered backstepping approach was applied to avoid tedious analytic calculations of the backstepping framework, and a compensation item was constructed to attenuate the filter error. A fault-tolerant controller was developed to ensure that the overall states of the controlled system were practically fixed-time bounded while the tracking error was regulated to the equilibrium region having a fixed-time convergence ratio. Illustrative studies showed the validity and practicability of the proposed theory.