HAL-INSA Toulouse
Not a member yet
34325 research outputs found
Sort by
Averaged Tracking Controllability of Parameter-Dependent Linear Systems and Volterra Integral Equations
No full textIn this paper, we explore linear parameter-dependent control systems, analyzing conditions under which an output variable, averaged over the parameters, can effectively track a prescribed trajectory within a finite time horizon. We characterize average tracking controllability through criteria that may be subtle to verify due to their infinitedimensional nature. However, in certain special cases, such as the single-input and single-output scenarios, we have applied an approach that we believe is novel, coming from the Volterra theory of integral equations, to derive criteria that lend themselves to straightforward verification. More interestingly, based on the system's architecture, this approach offers us, almost free of charge, the type of trajectories that can be reached exactly or approximately
Contribution of Raman analysis on tribological study of PEEK reinforced with micro or nano SiC particles
No full textInternational audienceThis study examines the tribological behavior of polyetheretherketone (PEEK) composites reinforced with silicon carbide (SiC) particles of varying sizes and concentrations. The composites, filled with both nano- and micro-sized SiC particles at different loadings (2.5%, 5%, 7.5%, and 10%), were evaluated for their wear resistance and friction properties using a ball-on-plate tribometer. Raman spectroscopy was employed to track crystallinity changes in the PEEK matrix within the wear tracks, revealing significant amorphization in high-stress regions, especially with higher SiC content.The findings show that nano-sized SiC particles provide superior reduced friction compared to their micro-sized counterparts, with optimal performance at 5–7.5% filler content. Nano-sized SiC particles enhance tribological performance by reducing friction while maintaining structural integrity in high-performance applications. However, higher SiC content led to increased wear in both micro- or nano-reinforced composites due to the instability of the wear debris. The study emphasizes the importance of particle size and filler content in improving tribological properties and suggests that further investigation into the mechanisms of wear and molecular disorientation could enhance the performance of PEEK-SiC composites
Grayscale evaporation of organic thin-film microstructures
No full textInternational audienceWe demonstrate that the deposition of organic thin films whose thickness is spatially shaped with sub-100-µm lateral resolution can be achieved by evaporation through a reusable flux-controlling micrometric-hole sieve that is separated from the to-be-coated substrate. This approach offers a practical way to grayscale-pattern organic materials without exposing them to detrimental ambient air or chemical solutions. As such, it facilitates the fabrication of long-lasting organic microstructured devices such as the reported spatially-varying absorptive color filters based here on the organic semiconductor (4,4'-Bis[4-(di-p-tolylamino)-styryl]biphenyl (DPAVBi)). Additionally, the ability to readily make multiple thickness-coated samples offers a material-sparing yet convenient way to measure the properties of the evaporated material, a feature exploited hereafter to extract the optical dielectric constants of DPAVBi over the visible and near-infrared range. This work and its extensions to inorganic compounds and freeform shaping pave the way for a new generation of devices that should find applications in light-matter coupling studies, hyperspectral imaging, structural coloring, or optical systems exploiting freeform optics
Competition between ordered morphologies of functionalized silver nanoparticles elucidated by a joint experimental and multiscale theoretical study
No full textInternational audienc
Gradient correlation is a key ingredient to accelerate SGD with momentum
No full textInternational audienceEmpirically, it has been observed that adding momentum to Stochastic Gradient Descent (SGD) accelerates the convergence of the algorithm. However, the literature has been rather pessimistic, even in the case of convex functions, about the possibility of theoretically proving this observation. We investigate the possibility of obtaining accelerated convergence of the Stochastic Nesterov Accelerated Gradient (SNAG), a momentum-based version of SGD, when minimizing a sum of functions in a convex setting. We demonstrate that the average correlation between gradients allows to verify the strong growth condition, which is the key ingredient to obtain acceleration with SNAG. Numerical experiments, both in linear regression and deep neural network optimization, confirm in practice our theoretical results
PHYSICS-INFORMED MACHINE LEARNING FOR PARAMETER IDENTIFICATION ON SHALLOW-WATER EQUATIONS
No full textInternational audienceThis work aims to develop a Physics-Informed Machine Learning method for efficiently solving inverse problems, with a focus on parameter identification for flood models, a critical task for improving forecast accuracy.More precisely, we propose a method to infer spatially-distributed physical parameters (e.g., friction or infiltration coefficients) using PINNs. The approach incorporates a physical residual term in the loss function, corresponding to the residuals of the 2D Shallow Water Equations with a hydrological forcing term (rainfall and infiltration). Additionally, a data discrepancy term is included in the loss function, based on observations of free surface height and mass flow rate at various locations within the computational domain. These observations are generated from a reference solution obtained using the Dassflow1 software for twin experiments, before considering real data. The PINN parameters and the physical parameters are then optimized by minimizing the weighted sum of the physical residual and data discrepancy terms. The training involves a pre-training step where only the data discrepancy term is minimized, followed by an alternating minimization strategy between the PINN parameters and the physical parameters. To address high-frequency components in the solution and optimization challenges, a Fourier Features embedding and an adaptive weighting scheme for the loss terms are used.To illustrate the performance of this method, several test cases will be discussed, some based on real-world data. Overall, the proposed method appears efficient and robust. Moreover, it is simple to implement (non-intrusive) compared to more traditional Variational Data Assimilation approaches, making it a viable strategy to further enhance for rapid flood forecasts
Dynamics and energy dissipation of collisional blast waves in a perpendicular magnetic field
No full textInternational audienceWe experimentally investigate the evolution and dynamics of laser-produced collisional blast waves (BW) under the influence of a perpendicular magnetic field up to 20 T. We show that an external magnetic field causes the BW to diverge from the Taylor–Sedov solution while also impacting its structural morphology. We notably explore the significance of various magnetohydrodynamic (MHD) processes occurring on scales similar to the width of the BW front by comparing their characteristic lengths to it and demonstrate that the downstream plasma's transition from being super- to sub-magnetosonic plays a pivotal role in the overall structure. Our results show that multiple MHD effects can contribute to shaping a magnetized BW, illustrating the complexity of the underlying physics
Visually Guided Model Predictive Robot Control via 6D Object Pose Localization and Tracking
No full textThe objective of this work is to enable manipulation tasks with respect to the 6D pose of a dynamically moving object using a camera mounted on a robot. Examples include maintaining a constant relative 6D pose of the robot arm with respect to the object, grasping the dynamically moving object, or co-manipulating the object together with a human. Fast and accurate 6D pose estimation is crucial to achieve smooth and stable robot control in such situations. The contributions of this work are three fold. First, we propose a new visual perception module that asynchronously combines accurate learning-based 6D object pose localizer and a high-rate model-based 6D pose tracker. The outcome is a low-latency accurate and temporally consistent 6D object pose estimation from the input video stream at up to 120 Hz. Second, we develop a visually guided robot arm controller that combines the new visual perception module with a torque-based model predictive control algorithm. Asynchronous combination of the visual and robot proprioception signals at their corresponding frequencies results in stable and robust 6D object pose guided robot arm control. Third, we experimentally validate the proposed approach on a challenging 6D pose estimation benchmark and demonstrate 6D object pose-guided control with dynamically moving objects on a real 7 DoF Franka Emika Panda robot
Journée Métal France 2025 - Protection & Conservation du Patrimoine MétalliqueLe Métal en danger et les dangers du Métal
No full textInternational audienceThe main aim of protecting our metal heritage is to ensure that it can be passed on to future generations. But beyond the aspects of preventive or curative protection, the durability of the metal, of its cultural, social and artistic vector, is also a priority. This applies just as much to recurring phenomena caused by deterioration or pollution, as it does to 'dramatic' events, from the scale of a workshop to that of a building. This also applies to the rarer cases where the dangers of metal as a source of pollution can also be considered a posteriori as 'positive' elements of human activity and serve as a historical marker.For this METAL France day, the papers will focus on the question of identifying risks (environment and metal concerned) and preventing them. Topics will include the deterioration of metal heritage and risk prevention in "extreme" cases (shipwrecks, war, etc.) or "simple" incidents or blunders (fire, flooding). It will also cover emergency response to damage, post-damage removal, development of sensors and preventive solutions. The dangers of metal for heritage also covers cases where metal presents a danger to the environment (toxic and radioactive emissions (asbestos, mercury, beryllium, lead, etc.), pollution (peat bogs, aquatic environments, etc.), while sometimes being a remarkable trace of human activity.Finally, this truly interdisciplinary day will look at medium-term conservation in 'usual' contexts (museums, private collections, local authorities, etc.) as well as 'hostile' contexts (e.g. aggressive environments, polluted sites, armed conflicts, complex administrative procedures, etc.).These aspects will be addressed through case studies on sites and in laboratories, feedback from field experience and dedicated methodological approaches.La protection du patrimoine métallique a pour principale objectif d’en permettre la transmission. Mais, au-delà des aspects de protection préventive ou curative, la pérennité du métal, de son vecteur culturel, social et artistique, est aussi à privilégier. Ceci concerne tout autant les phénomènes récurrents induits par la dégradation ou la pollution, que les évènements « dramatiques » de l’échelle de l’atelier à celui d’un bâtiment. Cela concerne des cas, plus rares, où les dangers du métal comme source de pollution, peuvent aussi être considérés a posteriori comme des éléments « positifs » d’activités humaines et servir de marqueur historique.Pour cette journée METAL France, les communications seront centrées sur la question de l’identification des risques (environnement et métal concerné) et de leur prévention. Sera ainsi abordée la dégradation du patrimoine métallique ainsi que la prévention des risques dans les cas « extrêmes » (naufrages, guerre...) ou de « simples » incidents ou maladresses (incendie, inondation). Elle traitera également des interventions d’urgence après sinistre, retraitements après sinistre, mise au point de capteurs et solutions préventives. Les dangers du métal pour le patrimoine, concerne aussi les cas où le métal présente un danger pour l’environnement (émanations toxiques et radioactives (amiante, mercure, béryllium, plomb…), pollution (tourbières, milieu aquatique, ...), tout ou étant parfois un remarquable traceurs d’activités anthropiques.Enfin, cette journée véritablement interdisciplinaire abordera la conservation à moyen terme en contexte « usuel » (musée, collections privées, collectivités…) mais aussi « hostile » (ex. milieu agressif, site pollué, conflits armés, complexité administrative, …).Ces aspects seront abordés par des cas d’études sur sites et en laboratoire, des retours d’expérience de terrain et aussi via des approches méthodologiques dédiées