HAL Portal UTC Université de Technologie de Compiègne
Not a member yet
11652 research outputs found
Sort by
An Efficient Hierarchical LSTM-based Framework for Intrusion Detection in Internet of Things (IoT) Systems
No full textInternational audienceThe rapid expansion of the Internet of Things (IoT) technology has enabled the interconnection of countless devices and systems worldwide. However, due to its lack of adequate cyber security measures, it has become a prime target for malicious hackers. To address this critical issue, a fully secure IoT network must be established, allowing for a safe and efficient deployment of this technology. This paper presents a sophisticated two-level deep learning-based intrusion detection model that uses long-short-term memory (LSTM), incorporating binary and multi-level classification techniques, time-based filtering and aggregation processes, all designed to optimize performance. The proposed model was evaluated using the UNSW-NB15 dataset, achieving highly satisfactory results, including reduced false alarm rates (false positives), and high accuracy and precision
Micro and Nanotechnology
No full textInternational audienceScale measurement is one of the frequently used approaches and tools for classifying micro, nano, and even macro technological application domains, despite the fundamentally significant variations in the chemical and physical mechanisms. Unit scales of macro, micro, and nano are used to understand and interpret better the size of objects in an associated and related manner at different level of scales in order to be able to understand the boundary-evading nature of materials fabricated at different levels for general and specialized fit-for-purpose applications. This book chapter focused on a relatively small portion of the larger and far-reaching field of micro/nanotechnology, which currently encompasses every aspect of science and engineering as well as anything we could possibly dream of or envision
Driving Towards Energy Efficiency: A Novel Torque Allocation Strategy for In-Wheel Electric Vehicles
No full textInternational audienceElectric vehicles (EVs) with four independent inwheel motors are classified as over-actuated systems, granting unprecedented possibilities to meet the total driving torque and yaw moment demands through an infinite number of feasible torque combinations. Ensuring an energy-efficient torque distribution among the motors is indispensable for mitigating energy consumption and extending the driving range. This is a pivotal factor in promoting eco-friendly and sustainable transportation solutions. This work focuses on the low-level control of a proposed multi-objective control architecture, encompassing longitudinal, lateral, stability, and maneuverability control. A novel configuration method for torque allocation is established, followed by developing and contrasting four multi-objective-based strategies. An energy-saving criterion is further developed, and the energy-efficient allocation strategies are carried out within the frameworks of online and offline optimization, based on the Sequential Quadratic Programming (SQP) algorithm. The proposed architecture is tested and validated in a joint simulation between Simulink/MatLab and SCANeR Studio vehicle dynamics simulator. The simulation outcomes demonstrate that implementing the suggested torque allocation can lead to enhancements in the energy efficiency, driving comfort, and stability of the electric vehicle
Normalised model-based processing diagrams for additive manufacturing of soft magnetic materials
No full textInternational audienceLaser beam melting (LBM) is an additive manufacturing technique that uses a laser beam to melt powdered materials to produce metal parts. LBM can be used to produce parts with desired magnetic properties, especially in the area of soft magnetic materials such as high silicon steel. The main goal of this work is create a concise chart integrating normalized energy density and provide a comprehensive overview of materials and parameters from multipleresearch articles
Rates of the strong uniform consistency with rates for conditional U-statistics estimators with general kernels on manifolds: Rates of the strong uniform consistency with rates for conditional U-statistics estimators with general kernels on manifolds
No full textInternational audience-statistics represent a fundamental class of statistics from modeling quantities of interest defined by multi-subject responses.-statistics generalize the empirical mean of a random variable to sums over every -tuple of distinct observations of . W. Stute [Ann. Probab. 19 (1991) 812--825] introduced a class of so-called conditional -statistics, which may be viewed as a generalization of the Nadaraya-Watson estimates of a regression function. Stute proved their strong pointwise consistency to :r^{(k)}(\varphi,\tilde{\mathbf{ t}}):=\mathbb{E}[\varphi(Y_{1},\ldots,Y_{k})|(X_{1},\ldots,X_{k})=\mathbf{\tilde t}], ~~\mbox{for}~~\tilde{\mathbf{ t}}=\left( \mathbf t_1, \ldots, \mathbf t_k \right)\in \mathbb{R}^{dk}. In the analysis of modern machine learning algorithms, sometimes we need to manipulate kernel estimation within the nonconventional setting with intricate kernels that might even be irregular and asymmetric. In this general setting, we obtain the strong uniform consistency result for the general kernel on Riemannian manifolds with Riemann integrable kernels for the conditional -processes. We treat both cases when the class of functions is bounded or unbounded, satisfying some moment conditions. These results are proved under some standard structural conditions on the classes of functions and some mild conditions on the model. Our findings are applied to the regression function, the set indexed conditional \textit{U}-statistics, the generalized \textit{U}-statistics, and the discrimination problem.The theoretical results established in this paper are (or will be) key tools for many further developments in manifold data analysis
Fast Current Control Based on Indirect Current Sensing of The Full-Bridge Converter Associated with An Electromagnetic Micro-Actuator
No full textInternational audienceThis paper deals with the fast current control of the full-bridge converter loaded with a digital electromagneticmicro-actuator. Input current sensing method is proposed for the initial full-wave control of output current. Inputcurrent analysis takes parasitic parameters into account and shows that input current approximately equalsoutput current during the full-wave state. The peak value of input current during steady state is twice the outputcurrent. A multi-scale closed-loop control is then implemented, which ensures stable fast current control. In themicro-actuator application, converters are tightly integrated with micro-actuator at output side. Input shuntsensor is easier to implement and has lower power dissipation compared with output shunt sensor. This solutionhas been validated with an experimental test bench. FPGA board is used to implement the proposed controlstrategy. Experimental results show that the start-up duration is greatly reduced, thereby realizing the fastcurrent response
Species-specific arsenic species and health risk assessment in seaweeds from tropic coasts of South China Sea
No full textInternational audienceArsenic (As) is a notorious toxic contamination in marine environments, while the toxicity and health risk of As is highly dependent on As species in seafoods. In this study, we hypothesized that the species-specific As bioaccumulation and species resulted in species-specific healthy risk of As in seaweeds. To test the hypothesis, we collected 10 common edible seaweeds from the coast of Hainan Island in South China Sea. Then we comparatively quantified concentration of total As and 5 major As species [AsB, DMA, MMA, As(III), and As(V)] in seaweeds. The results revealed that the concentrations of total As varied significantly among 10 seaweed species. Specially, the highest total As concentration were found in brown seaweeds, followed by red seaweeds, and green seaweeds. Furthermore, the percentage of 5 As species to total As differed significantly among 10 seaweeds. The percentage of AsB was highest in Caulerpa lentillifera (53%) and lowest in Sargassum oligocystum (13%), while that of As(V) was lowest in Caulerpa lentillifera (21%) and highest in Sargassum oligocystum (81%). The iAs [As(III) + As(V)] exhibited highest value in brown seaweeds and least value in green seaweeds. The potential human health risk assessment indicated that the consumption of brown seaweeds of Sargassum oligocystum and Sargassum polycystum could cause a considerable carcinogenic risk and non-carcinogenic risk to residents. Overall, our findings here largely validated our hypothesis that the species-specific As bioaccumulation and As species had great significance to healthy risk of As in seaweeds
Unlocking the polyfunctionality of cactus waste seed lignin in sustained catalysts: Optimizing the catalytic activity of a novel maleated lignin catalyst (MLC)
No full textInternational audienceLignin's polyfunctionality is evolving into alternative bio-based catalysts from under-used agro-industrial waste. To promote safe and eco-conscious disposal of these residues, maleated lignin catalyst (MLC) was prepared via simple and efficacious process, and used as a sustainable heterogeneous catalyst for the O-acetylation of SA (salicylic acid). Response surface methodology (RSM) built on Box-Behnken design of experiment (BBD) has reliably established the influence of reaction parameters on the catalytic properties of MLC. The characterization findings reveal stable yet efficient functional reactive sites on MLC, with an acid density of 4.92 mmol.g−1 and BET surface of 1.37 m2/g. With only 4% MLC, high activity was achieved (95.14%) with short reaction time and low temperature (30 min, 60 °C), and an Ea of 56.11 kj.mol−1. Upon reusability, the MLC was characterized by FTIR, XRD, TGA, SEM, BET, etc., and 57% of its catalytic activity was retained after 5 cycles, with high stability and no meaningful mass loss. The BBD model was statistically significant and exhibited good agreement between the predicted and actual yield. The adopted coupling of experimental and theoretical approaches for the first time in lignin-based catalysts can promote beneficial rational design and implementation of lignin catalysts in green chemical synthesis
Modélisation des effets de diffusion des batteries sodium-ion dans un circuit électrique équivalent
No full textInternational audienceUn modèle séparant les effets de diffusion en phases solides et liquide dans un circuit électrique équivalent est proposé. Ce modèle est testé sur les batteries sodium-ion de type NVPF/HC, dont la diffusion au sein des électrodes présente de fortes non-linéarités en état de charge. Grâce à ce modèle, les paramètres de diffusion présentent moins de non-linéarités et sont plus facile à déterminer. Le modèle proposé permet de bien simuler le comportement des batteries NVPF/HC pour des états de charge supérieurs à 30%
Parametric families of continuous belief functions based on generalized Gaussian random fuzzy numbers
No full textInternational audienceThe theory of epistemic random fuzzy sets is a general theory of uncertainty encompassing both possibility theory and the Dempster-Shafer theory of belief functions as special cases. Within this framework, Gaussian random fuzzy numbers have recently been introduced as a practical model of uncertainty about real variables. However, the limited flexibility of this model does not allow it to represent all kinds of beliefs encountered in applications. In this paper, it is extended in two ways. First, we study one-to-one transformations of random fuzzy sets and show that such transformations commute with combination. This property allows us to define parametric families of easily combinable random fuzzy numbers and vectors on different frames based on the Gaussian model. We then go one step further by studying mixtures of random fuzzy variables, which provide a very flexible model making it possible to construct belief functions on continuous frames with arbitrary complexity. To demonstrate the applicability and practical interest of these models, two applications are studied: the elicitation of expert beliefs about numerical quantities, and generalized Bayesian inference with weak prior information represented by random fuzzy numbers