Fault-tolerant flight control faces challenges as developing a model-based controller for each unexpected failure is unrealistic, and online learning methods can handle limited system complexity due to their low sample efficiency. In this research, a model-free coupled-dynamics flight controller for a jet aircraft able to withstand multiple failure types is proposed. An offline-trained cascaded Soft Actor-Critic Deep Reinforcement Learning controller is successful on highly coupled maneuvers, including high-bank coordinated climbing turns. The controller is robust to six unforeseen failure cases, including the rudder jammed at -15°, the aileron effectiveness reduced by 70%, a structural failure, icing and a backward c.g. shift as the response is stable and the climbing turn is completed successfully. Robustness to biased sensor noise, atmospheric disturbances, and to varying initial flight conditions and reference signal shapes is also demonstrated.

Online Adaptive Flight Control is interesting in the context of growing complexity of aircraft systems and their adaptability requirements to ensure safety. An Incremental Approximate Dynamic Programming (iADP) controller combines reinforcement learning methods, optimal control and Online identified incremental model to achieve optimal adaptive control suitable for Nonlinear Time-Varying systems. The main contribution of this thesis is twofold. Firstly, the iADP controller is designed to achieve automatic Online rate control to track pilot commands via setpoints provided by the manual outer loop on Citation II Aircraft model. Secondly, to assess the controller performance in the presence of sensor dynamics and actuator dynamics, an analysis is carried out to identify causes of any performance degradation. The simulation results from iADP longitudinal control using full state feedback indicate that the discretization of sensor signals, sensor bias and transport delays did not have any significant effect on the controller performance or on the incremental model identification. However noisy signals and sensors delays are found to cause controller performance degradation. Appropriate filtering of signals resulted in better estimation of the incremental model subsequently improving the controller performance due to noisy signals. Control performance degradation due to sensor delays should be addressed in future before conducting flight tests on Citation II Aircraft.

The control of aircraft can be carried out by Reinforcement Learning agents; however, the difficulty of obtaining sufficient training samples often makes this approach infeasible. Demonstrations can be used to facilitate the learning process, yet algorithms such as Apprenticeship Learning generally fail to produce a policy that outperforms the demonstrator, and thus cannot efficiently generate policies. In this paper, a model-free learning algorithm with Reinforcement Learning in the loop, based on Apprenticeship Learning, is therefore proposed. This algorithm uses external measurement to improve on the initial demonstration, finally producing a policy that surpasses the demonstration. Efficiency is further improved by utilising the policies produced during the learning process. The empirical results for simulated quadrotor control show that the proposed algorithm is effective and can even learn good policies from a bad demonstration.