Resilient Neural Control Based on Event-Triggered Extended State Observers and the Application in Unmanned Aerial Vehicles

In the real world, many systems can be written as multiple-input multiple-output (MIMO) systems, such as unmanned aerial vehicles (UAVs), unmanned vehicles, etc. Therefore, this is a problem for the generality of the research method, and this article aimming at the control problem of MIMO nonlinear systems with system uncertainties and external disturbances, a resilient control approach is proposed based on the event-triggered extended state observer (ETESO) with neural networks (NNs) in this article. Firstly, the constraint problem of the tracking error is transformed into an unconstrained problem based on the performance function, then the NNs are employed to approximate the uncertainties, and the ETESO with NNs is designed to estimate the combination of the approximation error and the external disturbance. Secondly, the perturbation of the controller gain is described by an external system, which is estimated by a fault observer. To avoid the problem of the differential explosion, the resilient controller is designed by the backstepping control method based on the command filter. Through the Lyapunov stability analysis, all signals of the entire system are uniformly ultimate bounded. Finally, the flight control experiment is shown to demonstrate the feasibility of the control scheme.