Learning Interpretable Models of Aircraft Handling Behaviour by Reinforcement Learning from Human Feedback