Adaptive Fault-Tolerant Tracking Control for Nonlinear Systems With Unknown Actuator and Multiple Sensor Faults

This paper formulates fault-tolerant tracking control for high-order nonlinear systems with actuator faults, and each state signal is measured by a single sensor which possibly suffers partial failure. The unknown fault parameters of sensors and actuator are assumed to be time-varying and their upper and lower bounds are unknown. A parameter separating tactic is employed to tackle the coupling issues in view of real state signals and unknown fault parameters. The actuator fault with high-order power $u^{\kappa _{n}}(t)$ is transformed into a linear function of $v^{\kappa _{n} }(t)$ plus a bounded term $dv(t)$ to overcome the restriction of high-order power and separate the fault parameters, such that the control signal can be constructed directly. Neural networks (NNS) are utilized to identify the unknown coexisting nonlinear uncertainties. By formulating adaptive bound estimation scheme to construct the NNS-based fault-tolerant tracking controller, which can eliminate the effects of multiple sensor faults, actuator faults, external disturbances and identify errors. Finally, the simulations verify the feasibility of the proposed tactic. Note to Practitioners—HNSs has a more general system structure. In practice, many systems can be described or transformed as HNSs, such as space vehicles and mechanical systems. Owing to the influence of environment and the physical components of the practical engineering systems, sensor faults and actuator faults are inevitable in engineering systems. Due to the restrictions of high-order power, the actuator faults cannot be converted into a friendly form such that the controller cannot be designed directly. Thus, the conventional results often assume the powers equals to 1 or the fault parameters are known.