Research Of Pilot Fatigue Assessment Model And Application

Making control of the fatigue state of pilot is an important topic of air safety. A lotof researchers and airline-company focus on this topic, while they only use the form of suggestion or biology. The major content of this paper is to make use of management information system and machine learning algorithms to manage pilot fatigueCombine theory of software engineering, we make needs analysis according to the needs of airline company. And then we make overall design of our system and divide system into 4 sub-modules: information management, fatigue management,flight recommendation management and rest period management. And on design stage, first we refer to the company's operating specifications and manuals to summarize the influencing factors of pilot fatigue, then we combine various fatigue factors into linear regression to calculate the fatigue value. Next we use these fatigue values to build up the pilot-flight matrix, and then make use of PMF model to recommend the flight. Then we combine CAAC regulations and fatigue value to optimize the rest period of pilot after flight. Finally, we integrate this three models to our Pilot Fatigue Management Information System, and offer risk value to pilot, offer fatigue value to manager, offer flight recommendation to scheduler, in order to analysis and control pilot fatigue.We collect flight data from real airline company, and get fatigue estimation by flight expert. Through model train and verify, we finally apply our model into the management system and prove it by Spring MVC frame which is popular in industry and python algorithms library.