Adaptive actor-critic network-based appointed-time attitude stabilization under actuator faults and dual-state constraints

Abstract

In this article, adaptive dynamic programming (ADP)-based attitude stabilization is achieved in the presence of external disturbances, actuator failures, and kinematic constraints. First, we design a new appointed-time convergence performance constraint function to constrain the transient and steady-state performance of attitude and angular velocity. Next, the attitude control system with state constraints is transformed into an unconstrained system through state mapping. Then, combining the extended state observer (ESO) and the integral sliding mode (ISM), an adaptive integral sliding mode controller is designed to eliminate disturbances and actuator failures to ensure that the system can move toward the nominal surface. Finally, the actor-critic network employs the adaptive weight update law to approximate the optimal value function and the optimal controller of the nominal system, respectively, without relying on the persistent excitation (PE) condition. The uniformly ultimately bounded (UUB) stability of the nominal system is obtained via the Lyapunov method. Moreover, it can be found from the numerical simulation that under the proposed control scheme, the attitude control system's attitude angle and attitude angular velocity both meet the performance constraints, and the cost consumption is much lower than other control schemes.