Portail HAL Ensta
Not a member yet
11080 research outputs found
Sort by
Advancing composite materials: Exploring thermomechanical properties of Aerosil/polycarbonate composites via additive manufacturing
No full textInternational audienc
Overview of Motion Planning Techniques and Their Suitability for an Off-Road Navigation Use-Case
No full textInternational audienc
Modeling swarm mission with COTS characterization: a series of return on experience
No full textInternational audienceAbstract System design in defense systems is a competitive field, in which economical viability relies on a sequence of architectural decisions, aiming at quality, resource and time (Q,R,T) compromises. We observe that low‐cost unmanned ground vehicles (UGV) and drones appear as new threats on current battlefield. To face these new threats, Direction Générale de l'Armement (DGA) have organized challenges around robotization of battlefield, to design future employment doctrines and help technologies to reach maturity in a reasonable time. This article exposes a NATO Architecture Framework (NAF) 3.1‐based workflow that includes return of experience form the field over yearly iterations of such challenges. The capabilities depicted are requirements to match, constituent systems are based on Components‐Off‐The‐Shelf (COTS) answering to both edition of the challenge. This article details how manually re‐injecting feedback from field back to the system model failed to ensure the next iterations of the challenge. Our works propose conclusions on formulation of the “engineering leakage problem” and how resolution of this problem is NP‐Hard and should be addressed using optimization
Effects of duty cycle on passive acoustic monitoring metrics: The case of blue whale songs
No full textInternational audienceLong-term fixed passive acoustic monitoring of cetacean populations is a logistical and technological challenge, often limited by the battery capacity of the autonomous recorders. Depending on the research scope and target species, temporal subsampling of the data may become necessary to extend the deployment period. This study explores the effects of different duty cycles on metrics that describe patterns of seasonal presence, call type richness, and daily call rate of three blue whale acoustics populations in the Southern Indian Ocean. Detections of blue whale calls from continuous acoustic data were subsampled with three different duty cycles of 50%, 33%, and 25% within listening periods ranging from 1 min to 6 h. Results show that reducing the percentage of recording time reduces the accuracy of the observed seasonal patterns as well as the estimation of daily call rate and call type richness. For a specific duty cycle, short listening periods (5–30 min) are preferred to longer listening periods (1–6 h). The effects of subsampling are greater the lower the species' vocal activity or the shorter their periods of presence. These results emphasize the importance of selecting a subsampling scheme adapted to the target species
Validated Uncertainty Propagation for Estimation and Measure Association, Application to Satellite Tracking
No full textInternational audienceIn this paper, we present a new uncertainty propagation algorithm based on interval arithmetic, and its applications for space surveillance. Using validated simulation, the goal is to explore the benefits of a set-based approach of estimation and data association for satellite tracking. The presented algorithm capitalises on the position measures of a satellite to improve its estimation and reduce uncertainties on its trajectory. Our approach also contributes to data association by computing the precision required for a measure to belong to a given track with confidence-levels. This paper illustrates the contributions of this new algorithm with several scenarios of orbit determination and satellite tracking and their numerical simulations
Latent Learning Progress Drives Autonomous Goal Selection in Human Reinforcement Learning
No full textInternational audienceHumans are autotelic agents who learn by setting and pursuing their own goals. However, the precise mechanisms guiding human goal selection remain unclear. Learning progress, typically measured as the observed change in performance, can provide a valuable signal for goal selection in both humans and artificial agents. We hypothesize that human choices of goals may also be driven by latent learning progress, which humans can estimate through knowledge of their actions and the environment -even without experiencing immediate changes in performance. To test this hypothesis, we designed a hierarchical reinforcement learning task in which human participants (N = 175) repeatedly chose their own goals and learned goal-conditioned policies. Our behavioral and computational modeling results confirm the influence of latent learning progress on goal selection and uncover inter-individual differences, partially mediated by recognition of the environment's hierarchical structure. By investigating the role of latent learning progress in human goal selection, we pave the way for more effective and personalized learning experiences as well as the advancement of more human-like autotelic machines.</div
Parcours d'étudiants en médecine intensive réanimation. Une approche sociologique au prisme du genre
No full textInternational audienceLa médecine intensive-réanimation (MIR) est une spécialité médicale où il existe historiquement une surreprésentation masculine, en particulier dans les postes hospitalo-universitaires et de chefs de service. Alors que parmi les médecins réanimateurs les plus jeunes, le nombre de femmes tend à rejoindre celui des hommes, une étude récente rapporte les difficultés d’épanouissement au travail parmi les réanimatrices avec un vécu difficile en particulier de la grossesse et de la maternité (Hauw-Berlemont et al, 2021). Cette même étude (Hauw-Berlemont et al, 2021) évoque les croyances relatives au genre des individus. Ces quelques recherches soulignent l’intérêt de porter attention aux dynamiques professionnelles de la MIR dans le contexte de sa féminisation. Différents travaux ont montré que c’est pendant les études que les médecins acquièrent les codes et connaissances propres à leur futur environnement professionnel et qu’ils et elles apprennent à reproduire les schémas comportementaux hérités du passé (Bercot, 2015). Pour autant, sous l’effet conjugué du renouvellement des générations et de la féminisation, ce contexte d’études peut évoluer et jouer sur ces phénomènes. Hardy-Dubernet (2005) souligne ainsi que les nouvelles générations sont moins enclines à sacrifier la vie familiale au profit de la vie professionnelle mais pour autant le dilemme entre être médecin et être une femme n’a pas disparu. Dans ce contexte, notre recherche s’attache à comprendre les parcours et les conditions de formation des internes en MIR afin de saisir tout à la fois les représentations attachées à la spécialité ainsi que les vécus de la formation. Alors que le sexisme est particulièrement prégnant dans des spécialités à forte présence masculine, qu’en est-il de la MIR qui connait une évolution démographique sensible ? Comment hommes et femmes ayant fait le choix de la MIR se représentent la spécialité et anticipent leur avenir professionnel notamment sur la question de l’articulation des temps de vie ? Comment ces étudiant.es vivent les collectifs de travail durant leurs études et comment cela peut façonner leur identité professionnelle
Ultrafast magnetization dynamics in NiCo thin films presenting weak stripes magnetic domains
No full textInternational audienceControlling magnetization without using magnetic fields is a technology-driven strong motivation in the quest for new electronic devices allowing for fast control with low energy consumption. A lot of results exist for ultrafast demagnetization in pure material (as pure Ni for example) but it is essential to understand this phenomenon in alloys or hetero structures since these systems present the highest potential for applications. We describe how we are using the chemical selectivity and the optical properties of the high harmonics to study the demagnetization induced by femtosecond laser pulses in thin nickel cobalt films presenting weak stripes magnetic domains
Collision avoidance in maritime traffic under COLREGs constraints: a reinforcement learning approach
No full textInternational audienceSpace exploration is often too hazardous for humans to perform, relying on autonomous rovers and probes to safely operate. Sea navigation presents a similar challenge within a non-modelled open environment, with autonomous surface vehicles (ASVs) being developed to operate without direct human control. Before being integrated into actual traffic, ASVs must follow the International Regulations for the Prevention of Collisions at Sea (COLREGs) However, automating these rules is challenging since they have multiple interpretations depending on the situation. A COLREG-compliant framework is proposed where the ASV predicts the actions of nearby vessels. Most of the time, it will abide by the rules, except in extreme situations. Then, the ship must find a safe strategy to avoid collisions and should be able to depart from COLREGs if necessary. This article shows how Deep Reinforcement Learning can be used to achieve this task, specifically the Proximal Policy Optimization. This method has been chosen because it is efficient at creating a policy solving the dual problem of path planning and collision avoidance while following rules. This framework has potential applications beyond maritime applications, such as spacecraft collision avoidance in congested orbits, highlighting the broader implications of this work in scenarios requiring protocol adherence in hazardous environments.</div