Archive ouverte HAL-LAAS
Not a member yet
12189 research outputs found
Sort by
Constrained Reinforcement Learning for Unstable Point-Feet Bipedal Locomotion Applied to the Bolt Robot
No full textInternational audienceBipedal locomotion is a key challenge in robotics, particularly for robots like Bolt, which have a point-foot design. This study explores the control of such underactuated robots using constrained reinforcement learning, addressing their inherent instability, lack of arms, and limited foot actuation. We present a methodology that leverages Constraints-as-Terminations and domain randomization techniques to enable sim-to-real transfer. Through a series of qualitative and quantitative experiments, we evaluate our approach in terms of balance maintenance, velocity control, and responses to slip and push disturbances. Additionally, we analyze autonomy through metrics like the cost of transport and ground reaction force. Our method advances robust control strategies for point-foot bipedal robots, offering insights into broader locomotion
CABLE SOLAR : A Tethered Airborne Platform Dedicated To Solar Cell Characterization
No full textInternational audienceA Tethered Airborne Platform Dedicated To Solar Cell Characterization is presented that could be used ultimately as an Aerostat for Solar Power GenerationPV prototype, electronics for data acquisition and intregration on aerostat were tested with success. The first results are promising and could be associated witha simulation part (LAAS collaboration :https://hal.science/hal-04730979
Size Profiling of cfDNA with μLAS for Agnostic Biomarker Identification
No full textCell-free DNA (cfDNA) in human blood plasma is a valuable biomarker for various physiological and pathological conditions. Specific genetic and epigenetic alterations associated with diseases can be detected in the blood, offering a non-invasive method for monitoring at-risk patients and assessing therapy efficacy.While high-sensitivity PCR-based technologies have demonstrated proof-of-principle, the low concentration of cfDNA complicates analytical routines, especially when rapid results are needed. Additionally, these technologies often require knowledge of the disease's genetic background, increasing the risk of false-negative diagnostics.We have developed the µLAS technology for high-sensitivity DNA sizing in just 25 minutes. µLAS utilizes electrohydrodynamic separation and concentration of DNA in funnel-like microfluidic channels. Unlike competing assays, µLAS operates effectively in plasma with high salt concentrations, enabling the collection of cfDNA biomarkers with excellent selectivity, independent of prior genetic knowledge.In this talk, we will present the µLAS technology and discuss our findings from clinical applications
SIROCCO: A dry wind to warn you from Bluetooth attacks
No full textInternational audienceBluetooth Low Energy (BLE) has become ubiquitous in Internet of Things (IoT) devices, yet it remains vulnerable to sophisticated over-the-air attacks that can compromise data security. This paper introduces Sirocco, a novel host-based intrusion detection system that leverages the open source Zephyr RTOS to address critical security challenges in BLE communications. Sirocco provides a flexible and lightweight detection framework that integrates with the BLE protocol stack. It targets sophisticated and low-level attacks such as spoofing, signal injection, and key sharing vulnerabilities. By inserting lightweight metric collection functions directly into the BLE stack's interrupt service routines and employing a dedicated kernel thread for analysis, Sirocco minimizes performance overhead while maintaining realtime attack detection capabilities across different device roles. Experimental evaluation demonstrates the system's effectiveness, with the framework introducing minimal performance and power consumption overhead
T-matrix representation of optical scattering response: Suggestion for a data format
No full textInternational audienceThe transition matrix, frequently abbreviated as T-matrix, contains the complete information in a linear approximation of how a spatially localized object scatters an incident field. The T-matrix is used to study the scattering response of an isolated object and describes the optical response of complex photonic materials made from ensembles of individual objects. T-matrices of certain common structures, potentially, have been repeatedly calculated all over the world again and again. This is not necessary and constitutes a major challenge for various reasons. First, the resources spent on their computation represent an unsustainable financial and ecological burden. Second, with the onset of machine learning, data is the gold of our era, and it should be freely available to everybody to address novel scientific challenges. Finally, the possibility of reproducing simulations could tremendously improve if the considered T-matrices could be shared. To address these challenges, we found it important to agree on a common data format for T-matrices and to enable their collection from different sources and distribution. This document aims to develop the specifications for storing T-matrices and associated metadata. The specifications should allow maximum freedom to accommodate as many use cases as possible without introducing any ambiguity in the stored data. The common format will assist in setting up a public database of T-matrices
Formal approach for Evolutionary Microgrids Modeling and Simulation
No full textInternational audienc
Photonic Integrated Circuit with System on Chip for Sub-Picometric Displacement Sensor
No full textNational audiencePICSONDE aims to develop an advanced embedded sensing system using optical feedback interferometry (OFI) in a laser diode for predictive maintenance applications. The system integrates a photonicintegrated circuit (PIC) to capture the frequency modulated (FM) channel of the OFI signal, and a system-on-chip (SoC) for data acquisition and processing. The project targets a quantum-limitedperformance of 0.1 pm/√Hz noise power spectrum density (PSD) over a 100 kHz bandwidth, representing the state-of-the-art in OFI systems. To achieve this, PICSONDE must overcome four main technologicalchallenges: (1) acquiring OFI signals with high dynamic range and signal-to-noise ratio, (2) extracting information from non-linear OFI signals, (3) real-time assessment of optical feedback betweenlaser and target, and (4) detecting interferometric fringes in the presence of speckle
Measurement of Wheel Skidding on Racing Wheelchairs
No full textInternational audienceIn the context of wheelchair racing, research primarily focuses on studying wheelchair ergonomics and determining kinematic, kinetic, and rolling resistance variables. One factor identified as influencing athletes’ performance is wheel skidding on the ground, a parameter complementary to rolling resistance. The objective of this study, therefore, is to identify, within a laboratory setting, the parameters that influence the risk of skidding in racing wheelchairs by measuring skidding torque. The ultimate goal is to enhance athletes’ performance by optimizing the interaction between the athlete and their wheelchair, and the wheelchair and the environment. In this perspective, four parameters were examined: the type of tubular, the camber angle, the tire pressure, and the load applied to the wheel using a skidometer. This tool characterizes a tire’s grip on a surface by measuring torques. The aim is to develop a system for classifying tire grip on dry athletics track at ambient temperature. The findings revealed that only the effects of load and tubular type had a significant impact on the torque values obtained. The tire that minimized the risk of skidding, among all tested combinations, is the Vittoria Pista Speed 23–28″. Furthermore, as the mass applied to the wheel increases, so do the resulting torques. This implies that a heavier athlete would require a greater force to be applied to the hand rim for the tire to skid. However, it was also demonstrated that the risk of skidding in a racing wheelchair is unlikely, as the torques obtained were over a range of 90 to 190 Nm. These values far exceed those typically exerted by para-athletes, which are a maximum of 60 Nm. The long-term goal would be to adjust the mode of torque application on the wheel using the skidometer for a more realistic field approach
Exploring the Impact of Large Language Models on Safety-Critical Processes
No full textInternational audienceThis paper explores the integration of Large Language Models (LLMs) in safety-critical systems, focusing on risks and mitigation strategies. While LLMs offer potential benefits in information analysis and decision support, their nondeterministic nature requires careful human oversight. Using System-Theoretic Process Analysis (STPA), we identify key risks related to tool behavior and human-LLM interaction. We propose strategies such as validation practices and engineer training to manage these risks. Further empirical research is needed to assess the effectiveness of these approaches in real-world contexts, providing a foundation for the responsible use of LLMs in safety engineering
PowerSecBench: reveal microarchitectural power leakages using generic RISC-V microbenchmarks
No full textInternational audiencePower side-channel attacks exploit variations in power consumption to recover sensitive information. To design secure systems, we must consider all the leakage sources to ensure that attackers cannot exploit any weakness. However, they can greatly vary between different processor microarchitectures and their internal mechanisms. Particularly, with the growing interest in the open-source RISC-V ISA and its numerous implementations, generic tools and strategies become essential to finely evaluate the security offered by each system.In this paper we present PowerSecBench, a software environment to evaluate power leakages on RISC-V processors. Starting from a range of various in-order implementations, from single-cycle to superscalar designs, we show how PowerSecBench allows characterizing leakages in each of them considering both value-and transition-based leakages. It uses a set of configurable microbenchmarks to test the different microarchitectural features. By simply executing them on both simulated and real FPGA targets, it becomes possible to identify most of the hardware sources, as well as consider physical parameters. Finally, we show that PowerSecBench can also be used to study the impact and limitations of protection strategies, such as pure software fences