Portail HAL Ensta
Not a member yet
11080 research outputs found
Sort by
Reflectance Transformation Imaging for the quantitative characterization of experimental fracture surfaces of bonded assemblies
No full textInternational audienceControlling failure modes is an important issue for the development of bonded assemblies of multi-material structures in industry. The analysis of the failure modes of bonded assemblies, and in particular the evaluation of the adhesive/cohesive failure rate, constitute a significant issue. Generally, it is considered in industrial applications that obtaining a cohesive fracture makes it possible to generate better predictions regarding the rupture of the adhesive joints, even if recent work has shown that a good prediction of the mechanical behavior can also be obtained when adhesive or mixed failure modes occur. Currently, these evaluations are carried out by experts by means of a visual analysis, either directly on the fracture surfaces, or sometimes from high-resolution microscopy. The approach presented in this paper aims to set out a methodology based on the Reflectance Transformation Imaging technique for the objective characterization of experimental fracture surfaces of bonded assemblies. We show how this imaging technique can be used to quantify objective surface features, and how these local descriptors can be used to estimate and map out the failure mode on the fracture surfaces
Reliability Assessment of an Unscented Kalman Filter by Using Ellipsoidal Enclosure Techniques
No full textInternational audienceThe Unscented Kalman Filter (UKF) is widely used for the state, disturbance, and parameter estimation of nonlinear dynamic systems, for which both process and measurement uncertainties are represented in a probabilistic form. Although the UKF can often be shown to be more reliable for nonlinear processes than the linearization-based Extended Kalman Filter (EKF) due to the enhanced approximation capabilities of its underlying probability distribution, it is not a priori obvious whether its strategy for selecting sigma points is sufficiently accurate to handle nonlinearities in the system dynamics and output equations. Such inaccuracies may arise for sufficiently strong nonlinearities in combination with large state, disturbance, and parameter covariances. Then, computationally more demanding approaches such as particle filters or the representation of (multi-modal) probability densities with the help of (Gaussian) mixture representations are possible ways to resolve this issue. To detect cases in a systematic manner that are not reliably handled by a standard EKF or UKF, this paper proposes the computation of outer bounds for state domains that are compatible with a certain percentage of confidence under the assumption of normally distributed states with the help of a set-based ellipsoidal calculus. The practical applicability of this approach is demonstrated for the estimation of state variables and parameters for the nonlinear dynamics of an unmanned surface vessel (USV)
Watching artificial intelligence through the lens of cognitive science methodologies
No full textSee also https://developmentalsystems.org/watch_ai_through_cogsciUtilizing a large size of language models contributes to recent advances in machine learning literature. However, interpreting what representations these models acquire is challenging due to their computational complexity. How can we understand the models functionally? This blog post suggests importing cognitive science methodologies to interpret their functional mechanisms. We introduce a way to conduct cognitive science experiments for machine learning models and analyze the models’ internal representation using cognitive neuroscience methodologies
On the Meaning of Sustainable Development. Humanity and Culture in the Age of Gaia and the Singularity
No full textInternational audienceThis text, a contribution to the philosophy of education, examines the meaning of the term sustainable development from the point of view of ordinary language philosophy. It argues that there are at least two dominant ways in which sustainable development currently makes sense, and that these two senses of sustainable development are mutually incommensurable and in fact radically opposed with respect to how they interpret the meaning of reforming education in the name of developing sustainably. Rather than attempting to resolve this conflict I argue that educators must acknowledge that negotiating with this divergence of interpretations and this problem of making sense of sustainable development is an essential part of the challenge of thinking about education for tomorrow
RF eigenfingerprints, an Efficient RF Fingerprinting Method in IoT Context
No full textInternational audienceIn IoT networks, authentication of nodes is primordial and RF fingerprinting is one of the candidates as a non-cryptographic method. RF fingerprinting is a physical-layer security method consisting of authenticated wireless devices using their components’ impairments. In this paper, we propose the RF eigenfingerprints method, inspired by face recognition works called eigenfaces. Our method automatically learns important features using singular value decomposition (SVD), selects important ones using Ljung–Box test, and performs authentication based on a statistical model. We also propose simulation, real-world experiment, and FPGA implementation to highlight the performance of the method. Particularly, we propose a novel RF fingerprinting impairments model for simulation. The end of the paper is dedicated to a discussion about good properties of RF fingerprinting in IoT context, giving our method as an example. Indeed, RF eigenfingerprint has interesting properties such as good scalability, low complexity, and high explainability, making it a good candidate for implementation in IoT context
Achievement and Fragility of Long-term Equitability
No full text12 pages, 7 figuresInternational audienceEquipping current decision-making tools with notions of fairness, equitability, or other ethically motivated outcomes, is one of the top priorities in recent research efforts in machine learning, AI, and optimization. In this paper, we investigate how to allocate limited resources to {locally interacting} communities in a way to maximize a pertinent notion of equitability. In particular, we look at the dynamic setting where the allocation is repeated across multiple periods (e.g., yearly), the local communities evolve in the meantime (driven by the provided allocation), and the allocations are modulated by feedback coming from the communities themselves. We employ recent mathematical tools stemming from data-driven feedback online optimization, by which communities can learn their (possibly unknown) evolution, satisfaction, as well as they can share information with the deciding bodies. We design dynamic policies that converge to an allocation that maximize equitability in the long term. We further demonstrate our model and methodology with realistic examples of healthcare and education subsidies design in Sub-Saharian countries. One of the key empirical takeaways from our setting is that long-term equitability is fragile, in the sense that it can be easily lost when deciding bodies weigh in other factors (e.g., equality in allocation) in the allocation strategy. Moreover, a naive compromise, while not providing significant advantage to the communities, can promote inequality in social outcomes
Intelligent Reflecting Surfaces and Spectrum Sensing for Cognitive Radio Networks
No full textInternational audienceIn Cognitive Radio (CR) networks, Primary User (PU) and Secondary User (SU) coexist to efficiently share the spectrum. PU has the right to access its dedicated channel at any time, while SU, operating in an opportunistic mode. can access only when PU is absent. Thus, SU should continuously monitor the channel to avoid any interference with PU when transmitting. Several factors, such as fading and shadowing adversely impact the PU SNR at the SU receiver making the Spectrum Sensing (SS) process more challenging. Recently, Intelligent Reflecting Surface (IRS) has been proposed to control the propagation channel for wireless systems. Introducing IRS in CR networks impacts the SS performance because of altering the channel. In this paper, we investigate the effect of deploying IRS on the SS by considering two scenarios: in (S1) the IRS is configured to enhance the PU signal at SU, while in the second scenario (S2), the IRS is configured to assist the Primary Receiver (PR). First, we highlight several important challenges and research directions. Then, we derive the analytical average detection probability for both (S1) and (S2). Results show that deploying IRS can significantly enhance SS even when the IRS is deployed to assist the PR
The Morozov's principle applied to data assimilation problems
No full textInternational audienceThis paper is focused on the Morozov's principle applied to an abstract data assimilation framework, with particular attention to three simple examples: the data assimilation problem for the Laplace equation, the Cauchy problem for the Laplace equation and the data assimilation problem for the heat equation. Those ill-posed problems are regularized with the help of a mixed type formulation which is proved to be equivalent to a Tikhonov regularization applied to a well-chosen operator. The main issue is that such operator may not have a dense range, which makes it necessary to extend well-known results related to the Morozov's choice of the regularization parameter to that unusual situation. The solution which satisfies the Morozov's principle is computed with the help of the duality in optimization, possibly by forcing the solution to satisfy given a priori constraints. Some numerical results in two dimensions are proposed in the case of the data assimilation problem for the Laplace equation
Extension of the Adjustable Localization Operator Method to Anisotropic Elasto-Plastic Behavior for Low-Cycle Fatigue Life Prediction
No full textInternational audienceAbstract The adjustable localization operator (ALO) method allows for the local analytical calculation of complex structures under multiaxial confined plasticity cyclic loadings. This study proposes a theoretical framework for the extension of the method to cover anisotropic yield surface materials. The application is carried out on steel tubes that have been axially compressed beforehand to create a bulge that acts as a stress concentration factor during the fatigue loading process. The ALO method is compared to the reference finite element anisotropic analysis at different steps of the fatigue design chain. Results have shown that although a slight gap on the absolute strain values exists, the strain amplitude and thus the fatigue life are correctly predicted with a reduction of the calculation time by 100
Single-shot transverse coherence in seeded and unseeded free-electron lasers: A comparison
No full textInternational audienceThe advent of x-ray free-electron lasers (FELs) drastically enhanced the capabilities of several analytical techniques, for which the degree of transverse (spatial) coherence of the source is essential. FELs can be operated in self-amplified spontaneous emission (SASE) or seeded configurations, which rely on a qualitatively different initialization of the amplification process leading to light emission. The degree of transverse coherence of SASE and seeded FELs has been characterized in the past, both experimentally and theoretically. However, a direct experimental comparison between the two regimes in similar operating conditions is missing, as well as an accurate study of the sensitivity of transverse coherence to key working parameters. In this paper, we carry out such a comparison, focusing in particular on the evolution of coherence during the light amplification process