Portail HAL Ensta
Not a member yet
11080 research outputs found
Sort by
Les apports des techniques structurales à l’analyse des processus d’innovation : une introduction
No full textInternational audienc
Sensitivity of ship-borne GNSS troposphere retrieval to processing parameters
No full textInternational audienc
Postfilament supercontinuum on 100 m path in air
No full textInternational audiencePulses at 744 nm with 90 fs duration, 6 mJ energy, and a weakly divergent wavefront propagate for more than 100 m and generate a filament followed by an unprecedently long high intensity (≥1TW/cm2) light channel. Over a 20 m long sub-section of this channel, the pulse energy is transferred continuously to the infrared wing, forming spectral humps that extend up to 850 nm. From 3D+time carrier-resolved simulations of 100 m pulse propagation, we show that spectral humps indicate the formation of a train of femtosecond pulses appearing at a predictable position in the propagation path
Verification and Reachability Analysis of Fractional-Order Differential Equations Using Interval Analysis
No full textInternational audienceInterval approaches for the reachability analysis of initial value problems for sets of classical ordinary differential equations have been investigated and implemented by many researchers during the last decades. However, there exist numerous applications in computational science and engineering, where continuous-time system dynamics cannot be described adequately by integer-order differential equations. Especially in cases in which long-term memory effects are observed, fractional-order system representations are promising to describe the dynamics, on the one hand, with sufficient accuracy and, on the other hand, to limit the number of required state variables and parameters to a reasonable amount. Real-life applications for such fractional-order models can, among others, be found in the field of electrochemistry, where methods for impedance spectroscopy are typically used to identify fractional-order models for the charging/discharging behavior of batteries or for the dynamic relation between voltage and current in fuel cell systems if operated in a non-stationary state. This paper aims at presenting an iterative method for reachability analysis of fractional-order systems that is based on an interval arithmetic extension of Mittag-Leffler functions. An illustrating example, inspired by a low-order model of battery systems concludes this contribution
HEVC hardware vs software decoding: An objective energy consumption analysis and comparison
No full textInternational audienceWeb data are experiencing a proliferation of video content for mobile platforms. This is accompanied by new advances in heterogeneous general purpose processor (GPP) cores embedded in mobile devices which offer a great opportunity to enhance both performance and energy efficiency of software (SW) video decoding. On the other hand, hardware (HW) video accelerators are more energy-efficient but are not flexible and their time-to-market is significant. In this context, this paper proposes a characterization methodology to investigate the performance and power consumption of two video decoding approaches on mobile platforms. The first one uses a HW decoder intellectual property (HDIP) in addition to a GPP (for the control). The second one is SW-based and uses only a heterogeneous multi-core GPP. The objective is to study the behavior of both video decoding approaches by comparing them and to understand why and in which case it is worth relying on the GPP rather than the HDIP. We also derive the optimal GPP configuration (number of cores and their frequency) that minimizes the energy consumption for a given video bit-stream on a given platform. The proposed methodology was applied on the HEVC video codec standard. In some state-of-the-art work figures, the SW video decoding consumes up to 1000× more energy than HDIPs. Our results show that, for video resolutions of 1080p and lower and at the operating system perspective point of view, the HEVC SW decoding consumes on average less than 4× more energy (mJ/Frame) than the HW one. Then, the more we scale up the resolution, the more we get the advantage of using the HW video decoding. Furthermore, the HEVC HW and SW decoders consume effectively less than 30% and 50% of the global power consumption of the tested platforms, respectively
Settings of demersal longlines reveal acoustic cues that can inform toothed whales where and when to depredate
No full textInternational audienceFishing boats produce acoustic cues while hauling longlines. These acoustic signals are known to be used by odontocetesto detect the fishing activity and to depredate. However, very little is known about potential interactions before hauling.This article describes the acoustic signature of the setting activity. Using passive acoustic recorders attached to the buoys oflonglines, this work demonstrates an increase in the ambient sound of 6dB re 1 lPa2 Hz1 within 2–7 kHz during the settingactivity. This could also be used as an acoustic cue by depredating species, suggesting that predators can detect longlinesas soon as they are set
Development of Metal-Ceramic-Lubricant Composite Coatings Obtained by Cold Spray for Tribological Applications in the Automotive Industry
No full textInternational audienceOne of the main levers to reduce CO2 emissions in cars and trucks is mass and friction reduction, which is often achieved through the use of special coatings. The aim of the present work was to develop metal-ceramic-lubricant composite coatings with the best combination of wear, seizure, fatigue, and thermal resistance. Metal-based coatings incorporating hard particles and solid lubricants were cold sprayed onto steel substrates and the relationship between coating microstructure and tribology was studied. To meet the demanding tribological requirements of heavily loaded engines, the interfaces between the different components were optimized by selecting appropriate feedstock powders and assessing a wide range of process parameters. Alumina-reinforced bronze composite coatings were made from powders with different morphologies. Aggregated ceramic powders were found to be more beneficial in terms of wear than massive powders, and graphite was found to be effective for reducing seizure
Ellipsoidal Enclosure Techniques for a Verified Simulation of Initial Value Problems for Ordinary Differential Equations
No full textInternational audienceThe verified simulation of initial value problems (IVPs) for ordinary differential equations (ODEs) with uncertain parameters is an up-to-date research topic and a basic building block for predictor-corrector type state estimators. Such state estimators are based on a two-stage procedure: First, the continuous-time state equations are evaluated up to the discrete time instant at which new measured data become available. Second, the forecasted state enclosures need to be refined by accounting for the information provided by the available sensors. In this paper, we focus on the first stage by presenting a novel ellipsoidal enclosure technique for continuous-time processes. It is based on thick ellipsoids and temporal Taylor series for a verified integration of ODEs that in combination allow for determining inner and outer bounds for the domains of reachable states. Comparisons with other set-valued integration techniques conclude this paper
SocialAI 0.1: Towards a Benchmark to Stimulate Research on Socio-Cognitive Abilities in Deep Reinforcement Learning Agents
No full textAccepted at NAACL ViGIL Workshop 2021International audienceBuilding embodied autonomous agents capable of participating in social interactions with humans is one of the main challenges in AI. This problem motivated many research directions on embodied language use. Current approaches focus on language as a communication tool in very simplified and non diverse social situations: the "naturalness" of language is reduced to the concept of high vocabulary size and variability. In this paper, we argue that aiming towards human-level AI requires a broader set of key social skills: 1) language use in complex and variable social contexts; 2) beyond language, complex embodied communication in multimodal settings within constantly evolving social worlds. In this work we explain how concepts from cognitive sciences could help AI to draw a roadmap towards human-like intelligence, with a focus on its social dimensions. We then study the limits of a recent SOTA Deep RL approach when tested on a first grid-world environment from the upcoming SocialAI, a benchmark to assess the social skills of Deep RL agents. Videos and code are available at https://sites.google.com/view/socialai01
Two Stages Parallel LMS Structure: A Pipelined Hardware Architecture
No full textInternational audienceModern wireless communication systems have tighten the requirements of adaptive beamformers when implemented on Field Programmable Gate Array (FPGA). The set requirements imposed additional constraints such as designing a high throughput, low complexity system with fast convergence and low steady state error. Recently, a parallel multi-stage least mean square (pLMS) structure is proposed to mitigate the listed constraints. pLMS is a two stages least mean square (LMS) operating in parallel and connected by an error feedback. To form the total pLMS error, the second LMS stage (LMS2) error is delayed by one sample and fed-back to combine with that of the first LMS stage (LMS1). pLMS provides accelerated convergence while maintaining minimal steady state error and a computational complexity of order O(N), where N represent the number of antenna elements. However, pipelining the pLMS structure is still difficult due to the LMS coefficient update loop. Thus, in this paper, we propose the application of the delay and sum relaxed look ahead technique to design a high throughput pipelined hardware architecture for the pLMS. Hence, the delayed pLMS (DpLMS) is obtained. Simulation and synthesis result, highlight the superior performance of the DpLMS in presenting a high throughput architecture while preserving accelerated convergence, low steady state error and low computational complexity. DpLMS operates at a maximum frequency of 208.33 MHz and is obtained at the cost of a marginal increase in resource requirements, i.e. additional delay registers compared to the original pLMS design