ETD

• adaptive control
• control systems
• controllability
• detectability
• e-modification
• flight dynamics
• model reference adaptive control
• mrac
• non-minimum phase systems
• observability
• partial-state feedback control
• stabilizability

This thesis, developed in collaboration with MBDA Italia S.p.A., investigates the application of Partial-State Feedback Model Reference Adaptive Control (MRAC) to an aircraft acceleration autopilot with the objective of increasing its robustness without drastically sacrificing its performance. “Adaptive Control” refers to a category of control system that have the ability to adjust control design parameters such as control gains online based on inputs received by the plant in order to accommodate system uncertainty. In MRAC the controller parameters are adapted in such a way that the system behavior matches a reference model whose performances are defined a priori and are representative of the desired behavior. The choice of a Partial-State Feedback MRAC, as opposed to a conceptually simpler Full-State Feedback MRAC, is due to the fact that in most practical applications not all the states are available for measurement. The control strategy studied in this thesis will be applied to an aircraft controlled by a pre-existing optimized autopilot in order to enhance vertical and lateral acceleration tracking at the same time. The system is neither observable nor controllable, conditions generally deemed necessary for MRAC usage, but only detectable and stabilizable.
The system is also non-minimum phase, a characteristic that can cause stability issues. For this reason, a robustness modification has been introduced and its stability has been proven through the Lyapunov stability theorem. A series of Monte Carlo analyses, carried out in multiple flight conditions, demonstrate the MRAC effectiveness in reducing the error with respect to the nominal behavior in the presence of parametric uncertainties and disturbances that arise from the lateral-longitudinal flight dynamics coupling. However, stability issues occur when the plant to be controlled has poles with low damping.