AI augmented pilot

点击翻译按钮获取中文摘要

One of the key elements in aircraft design is the interface with pilots. The human-machine interaction is done both by the instruments that are present in the cockpit, by the communication between the pilot and the traffic control tower, the physical feeling of the acceleration of the plane but also by the flight computers that transform the pilots' commands in displacements of the surfaces control of the aircraft. A cognitive overload due to the number of interactions can lead to misinterpretations and dramatic accidents. The aircraft control and management environment must be designed for use by a wide range of pilots, with uneven morphologies, skills and emotional and cognitive responses, and to ensure the safe operation of the plane. In order to further improve human-machine interfaces in aeronautics and ensure their safety operability, regardless of the situations and type of pilot, this project addresses the pilot modeling problem, and the ability to simulate its behavior based on factors such as cognitive load, mission, type of failure, etc. for a variety of representative pilot profiles. Such a pilot model will be based on Artificial Intelligence techniques using the actions and interactions made by the pilot, compared to the expected actions obtained throught simulators and the asessment of the pilot workload. It will help to elaborate the detection of possible mistakes and the set up of new evaluation rules of the aircraft operations. To carry out the proposed modeling, we have assembled a team composed of an aircraft manufacturer with experience in aircraft design (Bombardier), a company specializing in aircraft modeling and simulation and pilot training (CAE), a specialized company in the evaluation of the cognitive load of individuals by analysis, thanks to the artificial intelligence applied to the neurophysiological data of these individuals (BMU Labs), as well as two researchers recognized for their expertise in modeling the users interactions, emotional intelligence and brain-environments interactions.