Archives Ouvertes HAL
Toutes les publications de l'ENAC en direct.
All ENAC publications.
[hal-05046627] Multi-Agent Reinforcement Learning based Drone Guidance for N-View Triangulation
Cet article présente une nouvelle approche pour le contrôle d'une flotte de drones qui peuvent suivre la position d'une cible volante à l'aide de caméras omnidirectionnelles embarquées. Les drones utilisent l'apprentissage par renforcement multi-agents (MARL) pour apprendre des stratégies décentralisées qui optimisent leur formation et leur mouvement autour de la cible, en minimisant l'incertitude sur la position triangulée. Nous définissons une fonction de récompense qui encourage les suiveurs à minimiser la trace de la matrice de covariance de la position triangulée, dérivée d'un modèle analytique de propagation de l'incertitude. Nous utilisons Multi-Agent PPO (MAPPO)...
[hal-05044705] EEG-based Adaptive Visual Alerts to Mitigate UAS Operator Mental Fatigue: A Technical Proof of Concept
Uncrewed Aerial Systems (UAS) mishaps are partially attributable to mental fatigue during long endurance missions. Mental fatigue decreases cognitive flexibility resulting in a decreased performance when switching between sub-tasks. Solutions include adaptive interfaces. We propose a technical proof of concept that launches visual alerts to mitigate operator mental fatigue by estimating future performance based on brain activity (EEG). The proposed solution uses flying and detection tasks, and includes an online classification pipeline. The method is designed to work in an intra-user and inter-session manner, i.e. the solution is tuned to a user and works without...
[hal-05039992] EVOLVED AND TRANSPARENT PIPELINES FOR BIOMEDICAL IMAGE CLASSIFICATION
<div><p>This article presents an interpretable approach to binary image classification using Genetic Programming (GP), applied to the PatchCamelyon (PCAM) dataset, which contains small tissue biopsy patches labeled as malignant or benign. While Deep Neural Networks (DNNs) achieve high performance in image classification, their opaque decision-making processes, prone to overfitting behavior and dependency on large amounts of annotated data limit their utility in critical fields like digital pathology, where interpretability is essential. To address this, we employ GP, specifically using the Multi-Modal Adaptive Graph Evolution (MAGE) framework, to evolve end-to-end image...
[hal-05039610] Navigating Partial Automation in Firefighting with Drones: Trust, Take-Over, and Human-Drone Teaming
In the past years, the use of drones has been increasingly introduced to firefighting operations. Drawing on concrete examples from interviews with Thai firefighting professionals and recent field trials, as well as prior research, this paper examines the challenges of integrating (partially) autonomous, AI-enhanced drones into firefighting operations. Our findings reveal that, despite the promise of automation, on-field operators still prefer communication via a dedicated drone pilot-a preference driven by unresolved trust issues and concerns over information overload. We discuss challenges such as trust in automation and adaptive take-over. These inform our proposals...
[tel-05037157] GNSS Signal Processing under Spoofing
Radio Frequency Interference (RFI), such as jamming and spoofing, represents a growing threat to GNSS receivers, especially in civil aviation, which requires a high level of accuracy, integrity, continuity, and availability. Although jamming and spoofing interference signal models have been extensively studied in the literature, and GNSS observables are commonly used in spoofing detection, the exact impact of spoofing on these observables remains relatively unexplored. Understanding the impact of jamming and spoofing on GNSS receivers is essential for enhancing their resilience and facilitating the development of more sophisticated detection methods. Developing a model of...
[hal-05037064] Humanity does not play DICE
The DICE model claims that technological progress can prevent global warming at low cost. However, many uncertainties about climate damages and backstop technologies are not taken into account in this model. Our analysis is based on a very large number of scenarios to account for the possible range of the climate damage function and the level of efficiency of future backstop technologies, as well as their financial costs. We show that if DICE's assumptions turn out to be less optimistic, it's a very risky bet to rely on technology alone to tackle climate change, with 82% of our scenarios leading to economic collapse. Then we introduce in the DICE model a social lever...
[hal-05035189] Transport aérien et sociétés : influences croisées
<div><p>À l’image de ses aéroports, des espaces où se matérialisent une multitude d’enjeux, économiques, sociaux, environnementaux, le transport aérien, dans son ensemble, est profondément enchevêtré dans le fonctionnement des sociétés humaines actuelles. Au-delà du moyen de transport, de l’outil au service de la mobilité, l’activité aérienne a infiltré toutes les composantes de la vie en société. Elle est par certains égards l’un de ses piliers organisateurs, et même d’un mode vie à part entière. À l’heure d’une réflexion sur le devenir de l’aviation au prisme du dérèglement climatique, il s’agit de mesurer et d’établir une vue globale sur ses implications sociétales...
[hal-05032254] Piezoelectric resonant ice protection systems - Part2/2 : benefits at aircraft level
Piezoelectric resonant de-icing systems are attracting great interest. This paper aims to assess the implementation of these systems at the aircraft level. The article begins with the model to compute the power requirement of a piezoelectric resonant de-icing system sized from the prototype detailed in Part 1/2 of this article. Then the mass, drag, and fuel consumption of this system and the subcomponents needed for its implementation are assessed. The features of a piezoelectric resonant de-icing system are finally computed for aircraft similar to Airbus A320 aircraft and aircraft of different categories (Boeing 787, ATR 72 and TBM 900) and compared with the existing...
[hal-05031713] Cocktail party training induces increased speech intelligibility and decreased cortical activity in bilateral inferior frontal gyri. A functional near-infrared study
The human brain networks responsible for selectively listening toavoice amid other talkers remain tobe clarified. The present study aimed toinvestigate relationships between cortical activity and performance inaspeech-in-spee ch task, before (Experiment I)and after training-induced improvements (Experiment II). InExperiment I,74 participants performed a speech-in-speech task while their cortical activity was measured using afunctional near infrared spectroscopy(fNIRS) device. One target talker and one masker talker were simultaneously presented atthree different target-to-masker ratios (TMRs): adverse, intermediate and favorable. Behavioral results show that performance...
[hal-05030473] The Impact of Age on Cognitive Performance at Work: A Focus on Pilots and Air Traffic Controllers
The aging of the working population is the natural consequence of the demographic evolution of western countries. Managing this change is a real challenge for society and it raises the issue of the retirement age extension. Beyond economic considerations, it is imperative to explore strategies for preserving performance and safety at work with older individuals. This is naturally true for physical jobs, but also for critical and cognitively demanding fields of activity, involving a rapid pace of decision-making such as piloting (van Drongelen et al., 2017) and air traffic control (ATC) (Heslegrave, 1998). In 2023, the CEO of the Regional Airline Association stated that...