Description du projet
Un nouvel assistant personnel qui permet aux pilotes de donner le meilleur d’eux-mêmes
L’expression «pilote automatique» est aujourd’hui utilisée à tout propos. Elle fait le plus souvent référence à l’exécution d’une séquence d’actions sans vraiment y penser, plutôt qu’à sa signification officielle: la commande automatique de la direction d’un navire ou d’un véhicule aérospatial. Bien que cet usage ait tendance à donner l’impression que le «pilotage automatique» est un processus simple et direct, remplacer la prise de décision compliquée d’un pilote par une machine reste une gageure des plus complexes. Le projet E-PILOTS, financé par l’UE, étudiera les possibilités d’assistants cognitifs intelligents pour les pilotes, en mettant l’accent sur les avantages d’une relation symbiotique interdépendante plutôt que sur le remplacement ou la séparation des tâches, et ce dans l’objectif d’améliorer à la fois la fiabilité et les performances.
Objectif
New technologies have allowed pilots to assume more responsibilities reducing the number of some mechanic and repetitive tasks through automation. The flight deck transition to digital computer-based flight management system has shown a reduction of the workload across a variety of tasks. However, there are still several
open issues that requires further research and a deeper understanding of pilot cognitive decision making process to guide next steps on aircraft control task sharing between human pilots and machines.
Present project proposal will analyse the benefits of cognitive computing assistant and the challenges in key areas such as the relationship between pilot with the surrounding traffic and the aircraft state evolution considering the pressure of the operational context factors such as task responsibilities, situational awareness, and the performance objectives of the mission among others. The dynamics of these interdependencies will be analysed by means of ML techniques identifying the different thresholds that actives a cognitive tasks which will be the baseline to elaborate new symbiosis dynamic structures to improve human-machine information sharing mechanisms.
Human-in-the-mesh modelling paradigm relying on cognitive computing technologies will drive the research and innovation efforts to identify the potential changes in pilot roles and task responsibilities that will improve aircraft system reliability and performance, the pathway to automation in the flight deck to simplify the crew interaction with the systems and make tacit knowledge explicit to be accessible and usable for automated, intelligent systems.
This proposal consider cockpit operations for “man-computer symbiosis” as a metaphor for designing decision support systems that enhance human cognitive performance. Furthermore, it will be explored the Strong Cognitive Symbiosis approach which analyses true interdependence rather than simply cooperation between human and machines
Champ scientifique
Programme(s)
Régime de financement
CS2-RIA - Research and Innovation actionCoordinateur
08193 Cerdanyola Del Valles
Espagne