Automatic Carrier Landing System For V/STOL Aircraft Using L1 Adaptive And Optimal Control

This thesis presents a framework for developing automatic carrier landing systems for aircraft with vertical or short take-off and landing capability using two different control strategies - gain-scheduled linear optimal control, and L1 adaptive control. The carrier landing sequence of V/STOL aircraft involves large variations in dynamic pressure and aerodynamic coefficients arising because of the transition from aerodynamic-supported to jet-borne flight, descent to the touchdown altitude, and turns performed to align with the runway. Consequently, the dynamics of the aircraft exhibit a highly non-linear dynamical behavior with variations in flight conditions prior to touchdown. Therefore, the implication is the need for non-linear control techniques to achieve automatic landing. Gain-scheduling has been one of the most widely employed techniques for control of aircraft, which involves designing linear controllers for numerous trimmed flight conditions, and interpolating them to achieve a global non-linear control. Adaptive control technique, on the other hand, eliminates the need to schedule the controller parameters as they adapt to changing flight conditions. A fully-non-linear high fidelity simulation model of the AV-8B is used for simulating aircraft motion, and to develop the two automatic flight control systems. Carrier motion is simulated using a simple kinematic model of a Nimitz class carrier subjected to sea-state 4 perturbations. The gain-scheduled flight control system design is performed by considering the aircraft's velocity, altitude, and turn-rate as scheduling variables. A three dimensional sample space of the scheduling variables is defined from which a large number of equilibrium flight conditions are chosen to design the automatic carrier landing system. The trim-data corresponding to each flight condition is extracted following which the linear models are obtained. The effects of inter-mode coupling and control-cross coupling are studied, and control interferences are minimized by control decoupling. Linear optimal tracking controllers are designed for each trim point, and their parameters are scheduled. Using just two linear models, an L1 adaptive controller is designed to replace the gain-scheduled controller. The adaptive controller accounts for matched uncertainties within a control bandwidth that is defined using a low-pass filter. Guidance laws are designed to command reference trajectories of velocity, altitude, turn-rate and slip-velocity based on the deviation of the aircraft from a predefined flight path. The approach pattern includes three flight legs, and culminates with a vertical landing at the designated landing point on the flight deck of the carrier. The simulations of automatic landing are performed using SIMULINK® envrironment in MATLAB®. The simulation results of the automatic carrier landing obtained using the gain-scheduled controller, and the L1 adaptive controller are presented.