HAL-CentraleSupelec
Not a member yet
77624 research outputs found
Sort by
IoT over Non-Terrestrial Networks: A Bibliometric Analysis
No full textInternational audienceThe integration of the Internet of Things (IoT)with Non-Terrestrial Networks (NTN) has attracted increasingresearch attention as a promising solution for achieving globalconnectivity. In recent years, satellite based IoT technologiesleveraging Low Earth Orbit (LEO), Medium Earth Orbit(MEO), and Geostationary Earth Orbit (GEO) constellationshave emerged as key enablers for large scale, low power, andwide area communication with global coverage. To systematicallyexamine the evolution and research trends of IoT–NTN research,this paper presents a comprehensive bibliometric study thatidentifies dominant topics and highlights emerging trends andresearch directions. The analysis is based on 4,878 publications,primarily conference papers and journal articles, publishedbetween 2015 and 2025. The IEEE emerges as the leadingpublisher in this research area and IEEE Internet of ThingsJournal represents the top publication venue in terms of thenumber of publications. Among the publications, “Orbits” is themost popular topic. A keyword co-occurrence analysis identifiesfive main clusters. The leading keywords in each cluster areInternet of Things (Cluster 1), satellite communication (Cluster2), antennas (Cluster 3), 5G (Cluster 4), and GPS (Cluster 5).Among these, the first two clusters are the most dominant andtogether include 30 keywords. In addition this study investigatesthe growing research focus on LPWAN-based IoT technologies,including NB-IoT, LoRaWAN, LoRa, and LR-FHSS, highlightingtheir role in enabling global, low-power, and scalable connectivityfor IoT devices
Design Considerations for Visualization Transitions of 3D Spatial Data in Hybrid AR‐Desktop Environments
No full textInternational audienceWe present design considerations for animated transitions of the appearance of 3D spatial datasets in a hybrid Augmented Reality‐desktop context. Such hybrid interfaces combine both traditional and immersive displays to facilitate the exploration of 2D and 3D data representations in the environment in which they are best displayed. One key aspect is to introduce suitable transitional animations that change between the different dimensionalities to illustrate the connection between the different representations and to reduce the potential cognitive load on the user. The specific transitions to be used depend on data type, needs of the application domain, and other factors. We summarize these as design considerations to simplify the decision‐making process and provide inspiration for future designs. We apply our concept to three case studies: sparse 3D point data, MRI scan data, and molecular data. Finally, we give some practical guidance for the prescriptive use of our design considerations
Bi-Level Optimization for Contact and Motion Planning in Rope-Assisted Legged Robots
No full textThis paper presents a planning pipeline framework for locomotion in rope-assisted robots climbing vertical surfaces. The proposed framework is formulated as a bi-level optimization scheme that addresses a mixed-integer problem: selecting feasible terrain regions for landing while simultaneously optimizing the control inputs, namely rope tensions and leg forces, and landing location. The outer level of the optimization is solved using the Cross-Entropy Method, while the inner level relies on gradient-based nonlinear optimization to compute dynamically feasible motions. The approach is validated on a novel climbing robot platform, ALPINE, across a variety of challenging terrain configurations
Colouring the interference digraph of a set of requests in a bidirected tree
No full textIn this paper, we investigate the impact of the broadcast effect arising in filterless optical networks on the computational complexity of the wavelength assignment problem. We model conflicts using an appropriate interference digraph, whose proper colourings correspond to feasible wavelength assignments. Minimizing the number of required wavelengths therefore amounts to determining the chromatic number of this interference digraph. Within this framework, we first present a polynomial-time 2-approximation algorithm for minimizing the number of wavelengths. We then show that the problem is fixed-parameter tractable when parameterized by the number of available wavelengths. We also derive polynomial-time algorithms for computing the independence and clique numbers of this interference digraph
Energy-Efficient Orchestration and Placement of 5G Network Slices
No full textInternational audienceThe advent of fifth-generation (5G) mobile networks is accompanied by a massive and hetero-geneous demand for services with stringent requirements. To deliver customizable guarantees,5G adopts the network slicing paradigm, virtualizing isolated sub-networks over a shared phys-ical infrastructure [1]. While slicing increases flexibility, it also raises concerns about energyconsumption in virtualized infrastructures [2]. This work focuses on energy-aware orchestrationand placement of multiple slices under Service Level Agreement (SLA) constraint
When Numbers Tell Half the Story: Human-Metric Alignment in Topic Model Evaluation
No full textInternational audienceTopic models uncover latent thematic structures in text corpora, yet evaluating their quality remains challenging, particularly in specialized domains. Existing methods often rely on automated metrics like topic coherence and diversity, which may not fully align with human judgment. Human evaluation tasks, such as word intrusion, provide valuable insights but are costly and primarily validated on general-domain corpora. This paper introduces Topic Word Mixing (TWM), a novel human evaluation task assessing inter-topic distinctness by testing whether annotators can distinguish between word sets from single or mixed topics. TWM complements word intrusion's focus on intra-topic coherence and provides a human-grounded counterpart to diversity metrics. We evaluate six topic models-both statistical and embedding-based (LDA, NMF, Top2Vec, BERTopic, CFMF, CFMF-emb)-comparing automated metrics with human evaluation methods based on nearly 4,000 annotations from a domain-specific corpus of philosophy of science publications. Our findings reveal that word intrusion and coherence metrics do not always align, particularly in specialized domains, and that TWM captures human-perceived distinctness while appearing to align with diversity metrics. We release the annotated dataset and task generation code. This work highlights the need for evaluation frameworks bridging automated and human assessments, particularly for domain-specific corpora
PC-Posits: Enhanced Soft Error Resilience of Posit Arithmetic Through Analytical Modeling
No full textThe Posit number system was introduced to overcome the limitations of traditional real-number representations. It provides a representation that achieves higher near-unity precision and a wider dynamic range compared to standard numerical formats. Beyond numerical accuracy, prior studies have shown that certain Posit configurations exhibit notable resilience to Soft Errors (SEs). However, this resilience varies across different Posit configurations, and the underlying causes remain poorly understood. To address these limitations, we propose the first analytical framework for characterizing SE resilience in Posits. Our framework introduces the Expected Catastrophic Error and Exception Probability metrics to quantify severe deviations and the likelihood of anomalies under SEs, respectively. Together, our metrics explain the varying SE resilience across Posit configurations. We observe that Posit inherently self compensates against SEs, but this phenomenon weakens when the regime grows too large, leaving insufficient bits to fully encode the exponent. Building on these insights, we propose SE resilient Proactively-Clipped Posits (PC-Posits) that enforces "proactive clipping" while preserving a dynamic range and precision adequate for real-world applications. To evaluate PC-Posits, we perform memory fault injections across heterogeneous benchmarks. PC-Posits deliver up to 29.1 percent point (pp) higher SE resilience than standard Posits. Moreover, the proactive clipping enables simpler hardware and reduces energy consumption by up to 47.2% compared to standard Posits. Finally, PC-Posits offer up to 33.9 pp higher SE resilience compared to IEEE Floating-point (FP) formats
Quantification de l'incertitude dans le calcul du débit d'absorption spécifique induit dans le corps humain par des dispositifs 2G à 5G
No full textThis thesis focuses on uncertainty quantification for specific absorption rate calculations, a key metric for evaluating human exposure to electromagnetic fields from wireless communication devices. It addresses a major source of uncertainty : the variability of dielectric properties of human body tissues, whose impact is critical to the reliability of electromagnetic dosimetry simulations and regulatory compliance assessments. Four nonintrusive uncertainty quantification methods are investigated and compared : the Monte Carlo method (reference), non-intrusive polynomial chaos, GUM-based combined uncertainty method, and Bayesian Neural Networks. The methods are validated on exposure models of increasing complexity. The results highlight the necessity of integrating uncertainty quantification into regulatory testing procedures to ensure user safety.Cette thèse porte sur la quantification de l'incertitude dans le calcul du débit d'absorption spécifique, grandeur clé pour évaluer l'exposition humaine aux champs électromagnétiques des dispositifs de communication sans fil. Elle se concentre sur une source majeure d'incertitude : la variabilité des propriétés diélectriques des tissus biologiques, dont l'impact sur la fiabilité des simulations de dosimétrie électromagnétique est déterminant pour la conformité aux normes de sûreté. Quatre méthodes non-intrusives de la quantification de l'incertitude sont étudiées et comparées : la méthode de Monte Carlo (référence), les polynômes de chaos non-intrusifs, la méthode d'incertitude combinée basée sur le GUM, et les réseaux de neurones bayésiens. Elles sont validées sur des modèles d'exposition de complexité croissante. Les résultats soulignent l'importance d'intégrer la quantification d'incertitude dans les procédures de test réglementaires pour garantir la sécurité des utilisateurs
2D Direction-of-Arrival Estimation With a Fast-Scanning Single-Port Leaky Wave Antenna
No full textInternational audienceThe article presents a two-dimensional (2D), fourquadrant, direction-of-arrival (DoA) estimation using a leaky wave antenna (LWA). The proposed antenna constructed by cascading two LWA radiating sections arranged in an L-shape, sweeps two orthogonal planes to scan a field-of-view (FoV) over θ ∈ [0, 60°], ϕ ∈ [0, 360°], with scanning rates of 73.6°/% and 77°/%. The scanning is intended for DoA estimation over a 400 MHz channel in 5G+ mmWave systems. The working principle of the proposed antenna is discussed and validated with numerical results. DoA operations with a multiple signal classification (MUSIC) algorithm is also performed with the antenna responses to detect single and multiple sources on a 2D FoV
Geometry-Based Analytical Modeling for Effective Permittivity in Meshed Microstrip Structures
No full textInternational audienceThe design of transparent and lightweight microwavecomponents increasingly relies on meshed conductors. Sincestandard solid-line models cannot fully capture their electro-magnetic performance, designers frequently turn to full-wavesimulations for accurate analysis. These simulations, however,are computationally expensive and time-consuming, makingthem impractical for fast prototyping and iterative design. Inthis work, we propose a modified formulation of the effectivepermittivity model for meshed microstrip transmission lines.The approach uses a modified Hammerstad equation, whichis corrected by introducing a transparency-dependent parameterthat links the physical geometry of the mesh to the effectivepermittivity through an empirical correction applied to thepermittivity expression as a function of the fill factor, derived frommultiple simulation data points. The extraction procedure basedon the simulated S-parameters and chain-matrix formulation wasemployed to validate the proposed model with an average errorof 0.36%. This formulation effectively captures the dependenceof the model parameters on mesh geometry, providing anaccurate representation of the effective permittivity behavioracross different transparency levels within the frequency rangeof 2–4 GHz