IMPROVE HUMAN-MACHINE INTERACTION IN THE AVIATION SECTOR

Available online :

Automation has disrupted the relationship between the crew and the aircraft. Research in neuroscience and artificial intelligence seeks to propose solutions to relieve the mental burden.

Expérience en simulation de vol avec mesure de l'activité cérébrale

The automation introduced in the aviation sector in the 1980’s has been very beneficial because a pilot cannot control everything. But today the number of tasks delegated to the machine is such that the pilots assume a new role, that of supervisor, which implies new learning,” said Mickaël Causse, engineer-researcher specialized in neuroergonomics in ISAE-SUPAERO’s Department of Design and Operation of Aerospace Vehicles (DCAS) and co-organizer of the International Conference on Cognitive Aircraft Systems (ICCAS).

“In the new generations of civil and military aircraft, operations will be increasingly complex. It is necessary to assist the crew in a relevant way,” added Jean-Louis Guéneau, of Dassault Aviation, Technical Manager of the Chair of Design and Architecture of Cognitive Aerial Systems (CASAC), which associates the industrialist with ISAE-SUPAERO and its foundation.

Assistance that involves detecting fatigue, decreased alertness, stress, distraction or even excessive confidence in automation. For this, systems such as eye-tracking, facial temperature measurements or electrocardiograms can be implemented. “We must quantify the human factor, measure what can be measured. This research, which we support in particular in the context of the CASAC Chair, is useful in designing the architecture of future systems and also in establishing appropriate regulations for certification,” Jean-Louis Guéneau went on to say. The aircraft – but also the drones and autonomous vehicles – of the future will have to be intelligent, sensitive to context, have the ability to learn and interact with operators in a natural way.

Photo: Flight simulation experiment with brain activity measurement

International Conference on Cognitive Aircraft Systems (ICCAS)

Mesure de la température du visage avec caméra thermique

However, measurement instruments and software tools are still far from being operational and major issues remain: reliability, safety, ergonomics, explicability of algorithms. This is why the second ICCAS conference, held on June 1 and 2 in Toulouse, brought together nearly 180 participants from 23 countries, from such sectors as research in artificial intelligence, neurosciences, aviation, and the world of industry.

“One of the challenges we discussed was how to interpret the data using algorithms. For example, this involves translating information, such as the pilot’s eye movements or heart rate, into operational solutions to temporarily relieve his/her mental burden,”explained Caroline Chanel, a teacher-researcher specializing in automatics and decision systems in the DCAS department and co-organizer of the conference.

Significant advances were presented: “Until now, physiological data were collected and processed after the fact. The most recent work shows that we are now able to process data online, in real time. You can fit out an operator with sensors, extract relevant information using algorithms, deduce an assistance solution and propose it to him/her. Thus, the loop is closed." Jean-François Bonnefon, researcher at the Toulouse School of Economics, specialist in ethical and mobility issues, stressed that artificial intelligences and the solutions implemented to assist humans must be compatible with cultural and ethical frameworks which may differ from one country to another.

Photo: Face temperature measurement with a thermal camera

Photos of the conference

ICCAS_1 & 2 juin 2022

Credits: ISAE-SUPAERO and Aude Lemarchand

***************

Design and Architecture of Cognitive Aviation Systems chair - CASAC

This Chair, supported by Dassault Aviation and the ISAE-SUPAERO Foundation since 2016 and renewed until 2025, aims to develop innovative technologies that contribute, on the one hand, to qualify the interaction between humans and machines to check if cooperation is effective and, on the other hand, to automatically decide what needs to be maintained, suggested or changed to enhance team performance. For this purpose, quantitative behavioral and physiological metrics (eye-tracking, heart rate, face temperature, etc.) are studied and merged with more qualitative metrics to evaluate the effectiveness of human-machine cooperation. ISAE-SUPAERO researchers design algorithms that come from the field of artificial intelligence in order to adapt and strengthen this cooperation.

Our news Research

For full functionality of this site it is necessary to enable JavaScript. Here are the instructions how to enable JavaScript in your web browser.
Choose an RSS feed
The entire RSS feed
RSS Feed by theme
Campus Academic programs Institute International Diversity program Research Companies Sustainable Development Innovation DEOS Alumni Doctorat DCAS Ingénieur DMSM Mastère Spécialisé DISC LACS Apprentissage Evénement DAEP