Abstract
This article focuses on the adaptive output feedback fault‐tolerant control (FTC) problem for nonlinear stochastic systems with output constraint. Different from the existing FTC results, the actuator replacement time is considered, which may cause the tracking error to violate the constraint boundaries. First, because only inaccurate output information is available, a dynamic gain‐based reduced‐order K‐filter is designed to reconstruct unmeasurable state variables. Second, by means of a shifting function, a funnel‐like constraint function is proposed to constrain the tracking error into the pre‐given asymmetric constraint boundaries within a predefined time. Third, monitoring function is proposed to detect actuator failures. Once the tracking error violates the monitoring function, it is determined that the actuator failure has occurred, and then the backup actuator driven by the recovery controller will replace the failed actuator. With the help of Lyapunov stability theory, it is strictly proved that all closed‐loop system signals are bounded in probability. Finally, the simulation result is presented to verify the validity of the proposed algorithm.