Flight scenario via behavior-emergence reinforcement learning for assessment of airworthiness and evaluation of aircraft

Flight scenarios for assessment of airworthiness and evaluation of aircraft are imperative for the aircraft industry. A flight scenario developing method which takes both aircraft airworthiness and evaluation into consideration was proposed. Four aspects of techniques, including configuration of dynamical relations among flight environment, weather conditions and aircraft considerations, decomposition of flight crew tasks, mapping of workload functions and factors onto operational tasks, and specification of window events and data items, were provided and integrated based on methodological guide. Especially, the decomposition of flight crew tasks into operationally related single manipulations was realized via behavior-emergence reinforcement learning. Operationally related single manipulations were obtained as control policies searched from the knowledge of flight trajectories which were collected by interaction with the aircraft system. An automated planning and scheduling mechanism was adopted to facilitate recurrence of flight conditions and detection of working status of pilot, so that flight scenarios developed could be utilized in various flight tests for fulfillment of flight tasks, implementation of workload measurement, and establishment of minimum flight crew, which were useful for assessment of airworthiness, analysis of human factors, human-centered design of aircraft, virtual prototyping and engineering of aircraft.