HAL-CentraleSupelec
Not a member yet
77624 research outputs found
Sort by
Local Fan of Digital Planes for Parameter-Free Normal Vector Estimation on Digital Surfaces
No full textDetermining and analyzing geometric features is often an essential step in understanding and classifying three-dimensional objects. Among geometric features, the normal vector field provides a local and global approach to the surface structure of objects. However, for digital surfaces, obtaining this field often requires parameter adaptation or comes up against configurations that do not exist in digital planes. In this sense, the normal vector field is different from the vector field defined by pointwise tangent planes. In this article, we propose an approach based on a fan of digital planes around a point. This allows us to robustly capture all local configurations and to adapt to the local flatness of digital surfaces. Experimental evaluations using multigrid approaches show that it is both faster and more robust than state-of-the-art methods in the field, while maintaining comparable accuracy.</div
Design and fabrication of a novel millimeter-wave bandstop filter using transversal signal interference technique
No full textInternational audienceThis paper presents the design, simulation, and fabrication of a miniaturized millimeter-wave band-stop filter (BSF) on microstrip technology. The proposed arrangement is based on the concept of transversal interference of signals, and is realized as a pair of parallel transmission lines with the appropriate characteristic impedance and electrical length. Such scheme provides sharpening signal discrimination without resonator structures additional to the filter. The filter is fabricated on a ROGERS RT/Duroid 5880 substrate, selected for its favorable properties at high frequencies, including a thickness of 0.13 mm, a relative dielectric constant of 2.2, and a loss tangent of 0.0009. The optimized design was validated through electromagnetic simulations by two types of electromagnetic solvers, and then fabricated and measured by coplanar waveguide (CPW) probe to confirm its practical performance
Design, Equivalent Circuit Analysis, and Verification of A 400-Cell Broadband Dual-Polarized Reconfigurable Transmit-Array at Ka-Band
No full textInternational audienceA broadband 1-bit reconfigurable transmit-array (TA) featuring independent dual-polarized beam-scanning capability is reported at Ka-band. Bandwidth enhanced transmissive elements based on the current reversed mechanism are developed, which are optimized based on an equivalent circuit model. The elements are further miniaturized and aligned orthogonally in an end-to-middle scheme for dual-polarized operation. A 20 × 20-cell dual-polarized reconfigurable TA is fabricated and measured. The prototype exhibits a peak gain of 24.8 dBi, a 3-dB gain bandwidth of 18%, as well as an aperture efficiency of 26%. Moreover, it supports sum beams scanning up to 60° and difference beams steering up to 50°. The demonstrated dual-polarized reconfigurable TA can be a promising candidate for millimeter-wave communication and radar systems
Where Graphs Meet Fuzzy Logic — A DBMS-Centered Engine for Polyphonic Music Matching in Score Databases (Best demo Award)
No full textInternational audienceA Digital Score Library (DSL) is a digital system for storing, managing, and disseminating musical scores. Although such systems have traditionally relied on sheet music for engraving, modern digital representations now enable content-aware services, such as searching for specific musical patterns within the library's scores. Though monophonic pattern retrieval is well-studied, polyphonic pattern retrieval, which is the subject of this demonstration, remains a complex, open challenge. This demonstration introduces Skrid, an online DSL using graph-based storage for musical content, and Malo, its flexible querying module that enables the approximate search of polyphonic melodic patterns, ranks the results by relevance, and provides a detailed explanation of the answers
Beyond Silicon Area: Co-Product Allocation for a Binning-Aware Carbon Footprint of Processors
No full textInternational audienceCentral Processing Unit (CPU) manufacturing is a major contributor to the environmental footprint of digital technology. State-of-the-art Life Cycle Assessment (LCA) and embodied Greenhouse Gasses (GHG) emission models evaluate carbon impacts of CPU manufacturing based on chip area and lithography resolution, which results in a uniform carbon footprint for an entire range of products derived from the same design. However, this simplification masks the heterogeneity arising from industrial binning, where chips from the same lot are differentiated according to their performance. To resolve this inconsistency, we propose a new analytical framework that considers the different variants of a processor as co-products. We develop and compare several allocation methods for CPU manufacturing impact, including an economic allocation (price-based) and a physical allocation (performance-based). Moreover, we model the influence of production yield by contrasting a uniform distribution hypothesis with a more realistic Gaussian yield distribution. Our results reveal a significant redistribution of the environmental burden when applying our methods: the carbon impact of high-end processors is substantially higher than current estimates, while that of entry-level models is reduced by up to 87%. This methodological framework provides a fairer and more accurate picture, essential for guiding eco-design strategies and hardware acquisition policies for enhancing electronics sustainability.</div
eGoRG: GPU-accelerated depth estimation for immersive video applications based on graph cuts
No full textInternational audienceImmersive video is gaining relevance across various fields, but its integration into real applications remains limited due to the technical challenges of depth estimation. Generating accurate depth maps is essential for 3D rendering, yet high-quality algorithms can require hundreds of seconds to produce a single frame. While real-time depth estimation solutions exist — particularly monocular deep learning-based methods and active sensors such as time-of-flight or plenoptic cameras — their depth accuracy and multiview consistency are often insufficient for depth image-based rendering (DIBR) and immersive video applications. This highlights the persistent challenge of jointly achieving real-time performance and high-quality, correlated depth across views. This paper introduces eGoRG, a GPU-accelerated depth estimation algorithm based on MPEG DERS, which employs graph cuts to achieve high-quality results. eGoRG contributes a novel GPU-based graph cuts stage, integrating block-based push-relabel acceleration and a simplified alpha expansion method. These optimizations deliver quality comparable to leading graph-cut approaches while greatly improving speed. Evaluation on an MPEG multiview dataset and a static NeRF dataset demonstrates the algorithm’s effectiveness across different scenarios
Group-Robot Interaction in the Wild: An Exploratory Field Study in Semi-public Spaces
No full textInternational audienceHow does interaction with robots differ between spontaneously formed groups and individuals?Despite increasing robot deployment in public spaces, this question remains understudied in real-world settings. We conducted a field study deploying a service stationary robot in semi-public office spaces, tracking 221 individuals (42 alone, 179 in groups) across 95 interaction opportunities. Cookies were placed on accessible trays, creating a low-barrier functional interaction opportunity (taking a cookie) while allowing observation of spontaneous social behaviors. Groups demonstrated significantly higher engagement: functional interactions and social gestures. Within groups, leader presence amplified social engagement threefold. These findings are consistent with descriptive norm theory: group presence and leader behavior were associated with increased social engagement, though context-specific factors may moderate these effects. Results highlight the potential value of group detection for robots in multi-user environments, and demonstrate the feasibility of integrating psychological theory with automated tracking to study spontaneous human-robot encounters in the wild
Smt.ml: A Multi-Backend Frontend for SMT Solvers in OCaml
No full textInternational audienceSMT solvers are essential for applications in artificial intelligence, software verification, and optimisation. However, no single solver excels across all formula types, and different applications may require the use of different solvers. While the SMT-LIB language enables multi-solver support, it also incurs heavy I/O overhead. To address this, we introduce Smt.ml, an SMT-solver frontend for OCaml that simplifies integration with various solvers through a consistent interface. Its parametric encoding facilitates the easy addition of new solver backends, while optimisations like formula simplification, result caching, and detailed error feedback enhance performance and usability. Furthermore, Smt.ml is the only SMT frontend that includes a simplification-management engine for streamlining the integration of new formula simplifications and the verification of their correctness. Our evaluation demonstrates that Smt.ml’s results are consistent with those of its backend solvers and that its optimisations are highly effective on formulas generated from the symbolic execution of an extensive program-analysis benchmark
VLSF Decoding with Reliability Guarantees over Correlated Noncoherent Fading Channels
No full textThis report studies reliability-guaranteed decoding for variable-length stop-feedback (VLSF) codes over correlated noncoherent fading channels. The decoding rule is based on the evolution of the information density associated with a channel input-output realization. Due to channel memory, exact evaluation of this information density is intractable. To enable a constructive decoding, computable finite-blocklength lower and upper bounds on the information density that hold uniformly over time along each input-output sequence are derived. The lower bound enables a stopping-time analysis for VLSF decoding and has an operational meaning, while the upper bound quantifies the relaxation gap. Moreover, the relaxation gap between the bounds is explicitly characterized. For Gaussian signaling, the stopping-time distribution and the impact of fading correlation on decoding performance are numerically studied
