Dynamic inversion- and neural network- based five-stage cascaded control for automatic carrier landing subject to multiple disturbances and uncertainties

This work proposes a novel control scheme for automatic carrier landing to enhance the landing precision and improve both the tracking of the reference trajectory and the touchdown accuracy. First, the nonlinear dynamics of the carrier-based aircraft is written under a five-stage cascaded strict feedback form. Then, a robust automatic carrier landing system (ACLS) is proposed by employing a novel nonlinear dynamic inversion (NDI)-based control approach, disturbance observers (for estimating the airwake and wind type disturbances), radial basis function neural networks (RBFNNs) for the estimation of the parametric uncertainties, a block based on a recursive least squares algorithm (for the deck motion prediction), a deck motion compensation subsystem, a block for updating the aircraft reference trajectory, and first-order command filters (for computing the required reference signals). The NDI- and RBFNN- based automatic carrier landing system consists of a guidance subsystem, attitude control subsystems, an angular rates’ control subsystem, and an approach power compensation system. The stability properties of the disturbance observers, controllers, RBFNNs, and command filters are integrated into a Lyapunov-based analysis to prove the stability of the closed-loop control architecture. Finally, the effectiveness of the presented automatic carrier landing system is verified by means of comparative numerical simulations.