NOTE! This site uses cookies and similar technologies.

If you not change browser settings, you agree to it. Learn more

I understand

Learn more about cookies at : http://www.aboutcookies.org/Default.aspx?page=1

16 décembre 2020 | Le projet INSPEX reçoit le prix « Étoiles de l’Europe 2020»

CEA002533 BDVisant à réaliser un dispositif portable de perception de l’environnement pour les personnes déficientes visuelles, le projet européen INSPEX reçoit le prix spécial du jury « Étoiles de l’Europe 2020 ». Couronnant les efforts de ses 9 partenaires, ce prix a été remis à Suzanne Lesecq, directrice de recherche au CEA List et coordinatrice du projet, lors d’un événement placé sous le haut patronage de Madame Frédérique Vidal, ministre de l’Enseignement supérieur, de la Recherche et de l’Innovation.

Voir la vidéo de la remise du prix avec la présentation du projet à la fin de l'article

Si la canne blanche est adoptée au quotidien par les personnes déficientes visuelles, elle ne leur permet pas de détecter tous les obstacles qui jalonnent l'espace public, en particulier au niveau du bassin et du haut du corps. Offrir une véritable « bulle de sécurité virtuelle » qui fonctionne sous différentes conditions environnementales (y compris brouillard, pluie, lumière rasante, plein soleil) est essentiel pour offrir aux personnes déficientes visuelles une réelle autonomie de déplacement, en toute sécurité.

Le projet INSPEX, achevé en 2019, avait pour objectif le développement d’un dispositif portable de perception de l’environnement par fusion de données (notamment mesures de distance) en utilisant la technologie SigmaFusion™ développée au CEA. Placé sur une canne blanche et couplé à une interface audio développée par l’entreprise GoSense, qui spatialise en 3D la présence d’un obstacle potentiellement dangereux. Le prototype développé par le projet permet de créer une « bulle de sécurité » autour de son utilisateur.

« Après une phase d’initiation très simple et quelques heures d’utilisation d’Inspex, la personne malvoyante parvient à se déplacer de façon fluide et à contourner les obstacles sans les toucher. Tout cela grâce à la localisation des obstacles dans l’environnement et à réalité augmentée sonore », explique Suzanne Lesecq.

img 1A partir des mesures de distances fournies par les différents capteurs intégrés dans le dispositif, un modèle de l’environnement est construit. Son analyse conduit à définir une « bulle de sécurité » autour de la personne : tout obstacle entrant dans cette bulle est alors rendu audible et localisé en trois dimensions, offrant une réelle interface audio 3D via des écouteurs extra-auriculaires.

 

 

Outre son rôle de coordinateur du projet, le CEA a optimisé une technologie de radar UWB (Ultra Wide Band) du CEA-Leti répondant aux contraintes de poids, de consommation et d’encombrement du dispositif. Ce radar permet de détecter des obstacles mobiles et d’estimer leur vitesse relative, et donc d’anticiper un potentiel impact.

L'institut Carnot CEA List a également développé des algorithmes embarqués de perception de l’environnement, avec le CEA-Leti. Ceux-ci tiennent compte du déplacement du dispositif dans le modèle numérique de l’environnement, et analysent ce modèle pour en extraire les informations pertinentes, c’est-à-dire les obstacles potentiellement dangereux sur la trajectoire de déplacement de la personne et sur toute sa hauteur.

img 2   img 3   img 4
    Prototype de tests   Prototype Final

 

img 5

Mené sur une période comprise entre janvier 2017 et décembre 2019, le projet Inspex réunit neuf partenaires de six pays européens, sous la coordination du CEA. Bénéficiant d'une subvention de 4 millions d'euros de l'Union européenne et de 500 000 € du Gouvernement suisse, ce projet permet de concevoir une canne blanche intelligente pour personnes malvoyantes transmettant un retour audio spatial 3D sur la localisation des obstacles. https://www.inspex-ssi.eu/

 

 

 

December 15, 2020 | CEA Combines 3D Integration Technologies & Many-Core Architectures to Enable High-Performance Processors That Will Power Exascale Computing

IEDM 2020 Paper Details

 

IEDM 1
INTACT Active Interposer with 6 Chiplets (c) CEA

SAN FRANCISCO – Dec. 15, 2020 – CEA-List and CEA-Leti, research institutes at CEA, presented their technologies for achieving exascale-level, high- performance computing (HPC). Highlighting CEA-List’s advanced demonstrators and CEA-Leti’s state-of-the-art, 3D-technology toolbox that together enable higher bandwidth and heterogeneity for processors, the researchers explain how these empowered properties are critical for hardware innovations that help enable exascale computing.

- Exascale computing refers to computing systems capable of calculating at least 10¹⁸, or one-billion-billion, floating-point operations per second, which would be twice as fast as the fastest computer available today. Efforts to develop exascale computing are driven by highly data-intensive scientific and industrial applications, such as climate research, drug discovery and material design. This level of performance in HPC and Big Data will be achieved with heterogeneous computing nodes composed of generic processor chiplets hosting accelerator chiplets for improved operational intensity.

“Profound evolutions brought by high performance computing (HPC) applications are based on continuous and exponential increases in computing performances over the past decades,” explained Denis Dutoit, a CEA-List scientist and lead author of the IEDM paper, How 3D integration technologies enable advanced compute node for Exascale-level High Performance Computing?. “Supercomputers will soon achieve exascale-level computing performances mainly thanks to the introduction of innovative hardware technologies around the processors.”

IEDM 2
ExaNoDe Heterogeneous Multi-Chip Module (c) CEA
  • The CEA technologies are powering demonstrators in the ExaNoDe and INTACT projects, which have developed integrated prototypes with technology building blocks to support the EU’s drive towards exascale computing. The two institutes combined expertise in silicon and 3D sequential integration with many-core architectures, which are differentiated by their high level of scalability and power efficiency. They have demonstrated the benefit of new integration methods and processes following two main paths: finer 3D interconnect pitches, leading to improved bandwidth between compute chiplets, and assembly technologies that allow increasing heterogeneity in packaging, which improves peak performance.

  • In addition, the importance of a 3D-integration solution to developing HPC processors is confirmed by the European Processor Initiative (EPI), with which CEA is deeply involved. Its aim is to design and implement a roadmap for a new family of low-power European processors for extreme scale computing, high-performance Big Data and a range of emerging applications.

“These R&D successes open a path towards heterogeneous processors that will enable exascale-level supercomputers,” said Denis Dutoit. “We demonstrated that co-optimization of advanced architectures with 3D integration technologies achieves the level of computing performance and bandwidth required for HPC.”

  • Because chiplets stacked on active interposer allow modularity and reusability at low development costs, CEA-List also is investigating using this new methodology for HPC architectures in the embedded world, for compute-intensive accelerators. For edge applications requiring a high level of computation and memory, such as artificial intelligence (AI), chiplet-based partitioning will enable the creation of a broad range of solutions to meet the needs for embedded AI. Potential uses include autonomous driving, transport applications and industry 4.0.

  • Current CEA-Leti research work addresses die-to-wafer direct hybrid-bonding technology, which offers denser 3D interconnects with better electrical, mechanical and thermal parameters, and allows ultrahigh-bandwidth capabilities in heterogeneous systems. CEA-Leti also is working on high-density through silicon vias (TSV) (pitch 1 to 4 µms) to create together with die-to-wafer hybrid bonding a complete dense 3D stack. For the longer term, CEA-Leti is also investigating innovative photonic-interposer technology as a 3D-based photonic chiplet approach to enable interconnection of tens of computing chiplets with the resulting chip-to-chip communication bandwidth, latency and energy.

  • Over the next decade, co-optimization of advanced integration technologies with disruptive architectures is expected to establish the key foundations for HPC components.

  • This work was funded by the French National Programme d’Investissements d’Avenir (Investments in the Future), IRT Nanoelec, under Grant ANR-10-AIRT-05.

  • This work also was supported by the ExaNoDe project, funded by the European Union’s H2020 program under grant agreement No. 671578.

References: This email address is being protected from spambots. You need JavaScript enabled to view it.

 

 

December 15, 2020 | Keeping data more secure on the internet

actu snowpackIndividuals, businesses, and government are all increasingly concerned about data privacy and security when sending personal or sensitive information over the internet. CEA List has developed a novel data anonymization solution in response to growing demand for more robust data protection. This patented breakthrough technology will soon be commercialized by a startup called Snowpack.

 

 

 

More effective data protection could soon be a reality

Currently, data are exchanged over the internet in the form of packets, which include the data of interest, plus metadata such as IP addresses needed to deliver the data. The data can be protected using encryption. The metadata, however, remain visible, creating a vulnerability that hackers can exploit to very easily determine the identity and location of the sender and the recipient and, in some cases, access the encrypted data.

Data anonymization techniques do exist. These can entail using a trusted third-party to mask the sender’s, the recipient’s, and data subjects’ true identities. However, this information can still be seen by the third party. Over the past two years, successful cyberattacks exploiting this vulnerability have shown just how fragile today’s protection is.

A breakthrough concept

CEA List came up with a particularly innovative solution to enhance data anonymization and better protect data exchanges online. The data packets are divided into fragments made up of random but complementary data; these fragments are then anonymized and transmitted over separate pathways. This makes it impossible for hackers to get to sensitive data simply by observing the network used to transmit the data.

actu snowpack 2

In practical terms, it is also impossible for hackers to effectively monitor all of the fragments traveling over a network for a given message. And, even if hackers could find all of these apparently unrelated fragments, they would not be able to put them back together in a coherent manner, rendering them effectively useless.

A powerful plug-and-play technology

CEA List’s novel approach offers protection against mass network monitoring software, drastically reduces the area visible to hackers, and eliminates the need to use trusted third parties. Prototypes of the solution have been tested with success, and two patents have been filed to protect it.

Startup Snowpack, a spinoff of CEA List, was founded to develop and commercialize this technology. The company has set the ambitious goal of becoming the leading online data anonymization and security platform. The solution will be easy to integrate into software products and services, where it will seamlessly replace today’s solutions.

Contact: This email address is being protected from spambots. You need JavaScript enabled to view it.




December 1, 2020 | Transcranial ultrasound could help diagnose stroke

transcranienne 250
AdobeStock

Ultrasound could soon become a fast, accurate technique for diagnosing strokes. CEA List, a CEA Tech institute, is developing digital tools for this purpose on machines at the BioMaps lab at the CEA Joliot imaging platform.

To effectively treat strokes, a fast, accurate diagnosis is vital. Currently, the techniques used to diagnose different kinds of strokes all require major medical imaging equipment like MRIs and X-ray scanners that are not generally available in emergency rooms and that can be difficult to access rapidly. If ultrasound could be used to diagnose stroke, around 140,000 people who suffer strokes each year in France alone could benefit from faster and more effective treatment.

Traditionally, a distinction is made between two kinds of strokes, each with its own treatment protocol. Currently, MRI and X-ray tomography are the two main techniques used to diagnose strokes. Ultrasound is much easier to access, cheaper, and more compact than these techniques, making it suitable for use on board ambulances, for instance. The problem is that the skull prevents the ultrasound waves from propagating correctly, which has a substantial negative impact on the quality of the image.

Researchers at CEA List, a member of the Carnot Network, are using models they developed to simulate the propagation of ultrasound waves in complex solid materials for non-destructive testing applications to simulate the disturbances to ultrasound wave propagation caused by the presence of the skull between the ultrasound probe and tissue observed. If the morphology of the skull wall is known, the phase laws to apply to the multielement probe can be calculated, offsetting the disturbances to the ultrasound wave front.

The researchers were able to refocus the ultrasound waves on the soft cranial tissue and significantly improve the quality of the ultrasound image. The next step will be to demonstrate the advantages of this kind of adaptive imaging method (developed for NDT) and use it to develop a stand-alone imaging system.

Read article at http://www.cea-tech.fr/

 

 

November 9, 2020 | Improved early fault detection in electrical cables

détection de défauts 250
©ADOBESTOCK

CEA List, a CEA Tech institute, partnered with Nicomatic to develop a non-destructive testing (NDT) method for detecting the early signs of faults in electrical cables. They used a new, more sensitive and effective type of reflectometry.

Electrical cables are everywhere in our environment, carrying both energy and data for a variety of uses. The early detection of faults in these cables is vital—and today's commonly-used reflectometry-based systems[1] are plagued by a phenomenon known as self-blinding, which prevents faults from being detected early on. The Carnot CEA List recently came up with a solution to this problem in research conducted under the institute's joint lab with Nicomatic.

In electrical cables—one-dimensional systems—the sensor is located in a direct line of sight with the electrical signal that is sent to complete the test. The test signal does not contain any information, but still crowds out the signal returned via the channel being tested—and it is this far-less-perceptible signal that contains the information that is useful for the test. Nicomatic turned to CEA List experts to solve this problem, which they did by generating a signal opposite to the excitation signal and adding this signal to the channel's response. This ensures that the return signal only contains useful information.

The initial test results indicated that the new technique could effectively improve the detection of faults with very weak signatures. It could potentially be used in a wide range of applications, from self-diagnosing cable and cable network systems in the space (reusable space launchers), aeronautics, automotive (autonomous vehicles), rail, and energy transmission and distribution industries.

[1] Reflectometry is a technique where an electrical signal is propagated inside the system being tested. When the signal encounters a fault inside the system, part of the signal's energy is sent back to the point from which it originated. This reflected signal is analyzed for information on the system being tested.

Read article at http://www.cea-tech.fr/