Archive ouverte HAL-LAAS
Not a member yet
12189 research outputs found
Sort by
Optical properties and photonic applications of molecular spin-crossover materials
No full textInternational audienceWe review the current knowledge on the optical properties of molecular spin-crossover (SCO) materials, which exhibit reversible switching between low-spin and high-spin electronic configurations, as well as their use for practical applications in the fields of optics and photonics. We focus on the physical origins and magnitude of the complex refractive index change (both real and imaginary parts) in SCO materials, including a brief excursion to less-studied lower frequency ranges (THz–Hz). Until now, the optical property changes have primarily been used as a means of detection to characterize the SCO phenomenon itself (particularly at the nanometric scale) or to develop optical sensors for monitoring variations in an external stimulus. However, more recently, optical signal processing and modulation have also been explored using SCO nanomaterials as active elements in tunable photonic devices. Herein, we review the current state of the art of such functional devices, showing that molecular SCO compounds are emerging as a promising class of phase-change materials with high potential for specific active photonic applications, particularly in the visible spectral range. Key results on light-induced spin transitions and the photoswitching dynamics of SCO materials are also briefly reviewed in the context of photonic devices. Finally, the results of recent studies reporting strong light-matter coupling phenomena between SCO molecules and confined electromagnetic fields in resonant optical cavities are also discussed, with, in sight, the fascinating perspectives of manipulating the molecular properties using light
Backscattering-Based Security in Wireless Power Transfer Applied to Battery-Free BLE Sensors
No full textInternational audienceThe integration of security and energy efficiency in Internet of Things systems remains a critical challenge, particularly for battery-free and resource-constrained devices. This paper explores the scalability and protocol-agnostic nature of a backscattering-based security mechanism by integrating it into Bluetooth Low Energy battery-free Wireless Sensor Network. The proposed approach leverages the Wireless Power Transfer link, traditionally used for energy harvesting, to generate additional identification signals without increasing energy consumption or computational demands. Experimental validation demonstrates the solution's functionality using compact, low-gain antenna, ensuring compatibility with size-constrained applications such as Structural Health Monitoring and smart transport. Furthermore, this work addresses the challenges associated with backscattering dynamic range and multi-node Wireless Sensor Network scenarios, discussing potential collisions between identification signals and proposing future improvements to enhance generalizability and scalability. The findings underscore the potential of the backscattering-based security mechanism for creating secure, sustainable, and scalable IoT deployments across diverse protocols and applications
Point Cloud-Based Visual Planning and Servoing for Autonomous Vine Pruning
No full textThis paper addresses the challenge of vine pruning, a crucial and laborious task in agriculture, using robotic technologies and vision-based feedback control. The complex structure of vine makes visual servoing difficult due to challenges in visual feature extraction and 3D pose estimation. A novel approach combining visual servoing and vision-based planning is proposed for vine pruning. The online planner, inspired by nonlinear model predictive control (NMPC), dynamically computes relevant waypoints and generates corresponding point clouds. The visual servoing system integrates Iterative Closest Point (ICP) for point-cloud alignment and positionbased visual servoing (PBVS) to navigate the waypoints. A specialized pose estimation is also integrated to enhance both the settling time and accuracy of the process. Live experiments have successfully been conducted under controlled perception conditions with a Franka Emika manipulator and stereo camera, using three real vines with multiple cut points and scenarios involving motion of both the vine and robot's base
Partitionnement en cliques à coût minimum de graphes orientés avec contraintes de flot
No full textNational audienc
How an international engineers exchange brings to a successful process/project
No full textIn the world of research, international collaborations between researchers are common, well-structured, and benefit from a lot of funding. Exchanges between technical staffs in general are rarer, it is difficult to find dedicated funding. Through the opportunity offered by MEXT X-nics UTokyo Agile-X project two engineers of LAAS-CNRS have been hosted in Takeda Super Clean room in 2023; and two engineers and the Professor Mita has been welcomed in LAAS clean room during 3 weeks in two times in 2024.Each institute have their own knowledges and specialties; that have been share, train and learn. The two first exchanges permit to learn each other habits and work procedures. Between these, the distance exchanging also introduce some misunderstandings that create some fails. But learning about our errors the third exchange was crowned with success in only 2 days of technology. We did a large gap due to these win/win technological exchanges. We will show that these exchanges are an opportunity for real scientific and technological collaborations; and that as such they deserve to be developed. We will present very good results about HARNS technology 1 ; as well as all the benefits brought by these exchanges on the technological, scientific and human points of view (Fig. 1). They will benefit the two laboratories and will accelerate new scientific results. These international exchanges allow us to confront and enrich different and complementary visions of how to implement technologies. Which is all the more enriching and motivating for the evolution of ways of working and exchanging. We hope that our presentation will create new opportunities, and show the real interest for our supervisory bodies to organize this type of opportunities beneficial to the whole scientific and technological community
BFS-Based Canonical Codes for Generating Graphs with Constraint Programming
No full textInternational audienceWe consider the problem of generating all graphs that satisfy some given additional constraints (on vertex degrees, or cycle lengths, for example). Most previous works have proposed to generate canonical codes associated with adjacency matrices. In this paper, we consider canonical codes based on Breadth First Search (BFS), and we show how to generate them with Constraint Programming (CP): we introduce a set of basic constraints that must be satisfied by all canonical codes, thus breaking many symmetries, and we introduce a global constraint to break other symmetries. We illustrate the interest of our approach on connected claw-free cubic graphs, and show that it outperforms state-of-the-art CP and SAT Modulo Theory (SMT) approaches
Layer-wise dynamic event-triggered neural network control for discrete-time nonlinear systems
No full textRecent efforts in the control community focused on developing methods to guarantee closed-loop stability of systems controlled by multilayer perceptron (MLP)-based policies. However, little attention has been paid to the computational demand of such controllers due to frequent network evaluations. In this paper, we address this challenge by proposing a dynamic Event-Triggering Mechanism (ETM) to reduce the computational burden. Specifically, we focus on the stabilization of discrete-time Lur’e systems with input saturation. The proposed strategy reduces the evaluation frequency of the layers of the neural controller while preserving stability guarantees. The ETM is constructed using Linear Matrix Inequality (LMI)-based conditions, which leverage the known properties of activation functions and employ Finsler’s lemma to reduce conservativeness. Numerical results demonstrate the effectiveness of the proposed method, achieving significant computational savings compared to state-of-the-art solutions
Queries, Representation & Detection: The Next 100 Model Fingerprinting Schemes
No full textInternational audienceThe deployment of machine learning models in operational contexts represents a significant investment for any organisation. Consequently, the risk of these models being misappropriated by competitors needs to be addressed. In recent years, numerous proposals have been put forth to detect instances of model stealing. However, these proposals operate under implicit and disparate data and model access assumptions; as a consequence, it remains unclear how they can be effectively compared to one another. Our evaluation shows that a simple baseline that we introduce performs on par with existing state-of-the-art fingerprints, which, on the other hand, are much more complex. To uncover the reasons behind this intriguing result, this paper introduces a systematic approach to both the creation of model fingerprinting schemes and their evaluation benchmarks. By dividing model fingerprinting into three core components – Query, Representation and Detection (QuRD) – we are able to identify ~100 previously unexplored QuRD combinations and gain insights into their performance. Finally, we introduce a set of metrics to compare and guide the creation of more representative model stealing detection benchmarks. Our approach reveals the need for more challenging benchmarks and a sound comparison with baselines. To foster the creation of new fingerprinting schemes and benchmarks, we open-source our fingerprinting toolbox
Safety-Counter-Player: Utilizing potentially unsafe capabilities in safety-critical systems
No full textInternational audienceIn safety-critical systems, integrating machine learning components (MLCs) presents significant challenges in balancing safety with functional performance. Engineers strive to harness machine learning to enhance both system functionality and safety. However, they face obstacles in ensuring sound safety assurance for these machine learning components. This paper proposes a novel architecture that distinguishes between two roles: the safety-player, which is responsible for making critical safety interventions, and the counter-player, which focuses on optimizing functional performance. By permitting the safety-player to intervene only when absolutely necessary, the counter-player is allowed greater freedom in its operations. This separation not only improves performance but also maintains safety, fostering a more effective interaction between safety, comfort, and overall system utility
Business Drivers for Trustworthy DevOps
No full textInternational audienceThere is a huge profit potential for the automotive industry to adopt continuous deployment. By tradition, the automotive industry is dominated by long platform development projects. Such a project typically spans over several years and ends with the start of production (SOP). This way of working is very costly, and it also gives lower business value to the market, compared to an ideal situation with much shorter development cycles. The reason for higher costs and for lower income is inherent in the length of the development projects. Major contributors to this phenomenon are increasing market and development risks. Furthermore, the longer projects, the longer the time-to-market, i.e. longer time without income from what has been developed. Longer projects also generate higher capital costs. By formulating these risks explicitly and putting realistic monetary values on them, it is possible to evaluate products strategies aiming to minimize these risks. In an automotive context, we claim that this points to a need for modular, updatable product architectures