HAL Portal UTC Université de Technologie de Compiègne
Not a member yet
11652 research outputs found
Sort by
Etude expérimentale et simulation CFD-DEM du transport pneumatique à échelle pilote en vue de l’étude de solides non conventionnels
No full textInternational audienc
Couplage expérimentation et simulation DEM pour la compréhension et la prédiction de l’écoulement de poudres
No full textInternational audienc
Perception multi-capteurs avec des cartes vectorielles pour la localisation des véhicules autonomes
No full textFor autonomous vehicles, it is crucial to find localization solutions that meet the requirements of navigation tasks. For safety reasons, the localization system must provide accurate and reliable pose estimates, with a high availability and a low latency. To this end, multisensor data fusion techniques are employed. They generally combine GNSS receivers with proprioceptive sensors that provide vehicle kinematics and dynamics. In this PhD thesis, we particularly focus on the use of additional exteroceptive sensors such as lidars and cameras which can provide measurements on features georeferenced in maps. These sensors help overcome GNSS limitations in complex environments like urban areas, enabling lane-level positioning. When using perception sensors, various methods can provide localization information. A common approach in robotics is implementing Simultaneous Localization and Mapping (SLAM) with GNSS constraints. This builds a local map online from raw sensor data, which can then be used for re-localization with the same sensors. Using accurate prior maps offers an interesting alternative enabling immediate localization upon entering a new area without the need to create a map anew. In this PhD thesis, we consider high-definition (HD) vector maps containing georeferenced road features represented as points or polylines. They encompass a wide range of physical elements essential for navigation such as traffic signs or road markings. The main goal of this thesis is to leverage all the potential offered by HD vector maps to improve localization, through a perception system whose performance has been optimized for this purpose. As a case study, our research focuses on detecting pole-like features, which are commonly found throughout road environments and georeferenced in HD maps. More specifically, we present camera and lidar perception approaches enabled by a map-based automatic annotation method. This method can annotate any kind of mapped poles. To enhance annotation accuracy and completeness, we integrate this primary method with additional automatic annotation sources. We train detectors to identify pole bases in camera images and from clusters of lidar points. This integration ensures that detected poles conform to the definitions used in the HD map. Finally, these detection approaches are integrated in a multi-sensor fusion system to assess their benefits for a localization system. Given the approaches explored, the thesis heavily relies on experimental data collected from vehicles equipped with lidar sensors and cameras. This work was carried out in synergy with the European project ERASMO, which aimed to develop a highly accurate and reliable localization system for autonomous vehicles.Pour les véhicules autonomes, il est crucial de trouver des solutions de localisation qui répondent aux exigences des tâches de navigation. Pour des raisons de sécurité, le système de localisation doit fournir des estimations de pose précises et fiables, avec une grande disponibilité et une faible latence. A cette fin, des techniques de fusion de données multicapteurs sont employées. Elles combinent généralement des récepteurs GNSS avec des capteurs proprioceptifs qui fournissent la cinématique et la dynamique du véhicule. Dans cette thèse de doctorat, nous nous concentrons particulièrement sur l’utilisation de capteurs extéroceptifs supplémentaires tels que les lidars et les caméras qui peuvent fournir des mesures sur des caractéristiques géoréférencées dans des cartes. Ces capteurs permettent de surmonter les limites du GNSS dans des environnements complexes tels que les zones urbaines, en fournissant un positionnement au niveau des voies. Diverses méthodes peuvent fournir des informations de localisation à partir de capteurs de perception. Une approche courante en robotique consiste à mettre en œuvre une méthode de localisation et cartographie simultanées avec des contraintes GNSS. Cette méthode construit une carte locale à partir des données brutes des capteurs, qui peut ensuite être utilisée pour la relocalisation avec ces mêmes capteurs. L’utilisation de cartes préalables précises offre une alternative intéressante permettant une localisation immédiate dès l’arrivée dans une nouvelle zone sans avoir à créer de carte. Dans cette thèse de doctorat, nous considérons des cartes vectorielles haute définition (HD) contenant des caractéristiques routières géoréférencées représentées sous forme de points ou de polylignes. Elles englobent un large éventail d’éléments physiques essentiels à la navigation, tels que les panneaux de signalisation ou les marquages routiers. L’objectif principal de la thèse est d’exploiter tout le potentiel offert par les cartes vectorielles HD pour améliorer la localisation, grâce à un système de perception dont les performances sont optimisées à cette fin. En guise d’étude de cas, notre recherche se concentre sur la détection d’éléments de types ”poteaux”, que l’on trouve couramment dans les environnements routiers et qui sont géoréférencés dans les cartes HD. Plus précisément, nous présentons des approches de perception par apprentissage automatique utilisant des caméras et des données de lidar. Afin d’éviter de labelliser manuellement les données, nous étudions des méthodes d’annotation automatique qui utilisent les cartes. Ces méthodes peuvent annoter automatiquement n’importe quel type de poteau cartographié. Pour améliorer la précision des annotations, nous combinons des sources d’annotation automatique supplémentaires. Nous présentons des détecteurs pour identifier les pieds des poteaux dans les images de caméras et les poteaux dans les nuages de points lidar. Cette approche garantit que les détecteurs de poteaux sont bien adaptés aux données géoréférencées dans les cartes HD. Enfin, ces approches de détection sont intégrées dans un système de fusion multi-capteurs afin d’évaluer leurs avantages pour un système de localisation. Compte tenu des approches explorées, la thèse s’appuie fortement sur des données expérimentales collectées à partir de véhicules équipés de capteurs lidar et de caméras. Ce travail a été fait en synergie avec le projet européen ERASMO qui visait à développer un système de localisation très précis et fiable pour les véhicules autonomes
La politisation de la moindre pièce: Tensions entre low-tech et high-tech dans les communautés du petit éolien auto-construit
No full textInternational audienceThis article looks at the relationship between technical and activist issues in the low-tech movement, based on the case of the community development of the Piggott wind turbine. It describes the ways in which those who promote self-built small wind turbines are led to mobilize tools that are far removed from their ideals, and to explore and manage the tensions that are generated. The survey of these communities highlights three stages in the making of this technology, guided by a singular relationship with material that I propose to call ‘politicizing down to the slightest part’. Acknowledging the political dimension of technological choices, the members of these communities problematize their entire material environment and its ramifications in a spectacular way. They do not, however, sink into immobilizing purism, and the successful distribution of this home-made turbine shows how their idealism can be combined with practical effectiveness.Este artículo examina la relación entre cuestiones técnicas y activistas en el ámbito de las bajas tecnologías, tomando como base el desarrollo comunitario del aerogenerador Piggott. Se describen las formas en que los impulsores de este pequeño aerogenerador autoconstruido se ven obligados a movilizar herramientas alejadas de su ideal, explorar las tensiones que se generan y aprender a gestionarlas. El estudio de estas comunidades arroja luz sobre tres momentos de la fabricación de esta tecnología, guiados por una relación singular con el material que propongo denominar la «politización de la más mínima parte». Al reconocer la dimensión política de las opciones tecnológicas, los miembros de estas comunidades problematizan ampliamente el conjunto de su entorno material y sus ramificaciones. Sin embargo, no caen en un purismo inmovilista y se apoyan en el éxito de esta máquina eólica casera para demostrar la manera en que su idealismo puede combinarse con la eficacia práctica.Cet article porte sur l’articulation entre enjeux techniques et militants dans la low-tech à travers le cas du développement communautaire de l’éolienne Piggott. Il raconte les façons dont les acteurs promouvant le petit éolien auto-construit sont amenés à mobiliser des outils éloignés de leur idéal, à explorer les tensions qui sont générées et à les gérer. L’enquête auprès de ces communautés éclaire trois moments de la fabrique de cette technologie, guidée par un rapport singulier à la matière que je propose de nommer la « politisation de la moindre pièce ». Reconnaissant la dimension politique des choix technologiques, les membres de ces communautés problématisent l’ensemble de leur environnement matériel et de ses ramifications, de façon vertigineuse. Ils ne sombrent pourtant pas dans un purisme immobilisant et montrent à travers le succès de la diffusion de cette machine à vent artisanale la manière dont se combine leur idéalisme à une efficacité pratique
From Spent Black and Green Tea to Potential Health Boosters: Optimization of Polyphenol Extraction and Assessment of Their Antioxidant and Antibacterial Activities
No full textInternational audienceTea, one of the most popular beverages worldwide, generates a substantial amount of spent leaves, often directly discarded although they may still contain valuable compounds. This study aims to optimize the extraction of polyphenols from spent black tea (SBT) and spent green tea (SGT) leaves while also exploring their antioxidant and antibacterial properties. Response surface methodology was utilized to determine the optimal experimental conditions for extracting polyphenols from SBT and SGT. The total phenolic content (TPC) was quantified using the Folin–Ciocalteu method, while antioxidant activity was evaluated through the DPPH assay. Antibacterial activity was assessed using the disk diffusion method. Additionally, high-performance liquid chromatography (HPLC) was employed to analyze the phytochemical profiles of the SBT and SGT extracts. Optimal extraction for SBT achieved 404 mg GAE/g DM TPC and 51.5% DPPH inhibition at 93.64 °C, 79.9 min, and 59.4% ethanol–water. For SGT, conditions of 93.63 °C, 81.7 min, and 53.2% ethanol–water yielded 452 mg GAE/g DM TPC and 78.3% DPPH inhibition. Both tea extracts exhibited antibacterial activity against Gram-positive bacteria, with SGT showing greater efficacy against S. aureus and slightly better inhibition of B. subtilis compared to SBT. No activity was observed against the Gram-negative bacteria E. coli and S. typhimurium. HPLC analysis revealed hydroxybenzoic acid as the main phenolic compound in SBT (360.7 mg/L), while rutin was predominant in SGT (42.73 mg/L). The optimized phenolic-rich extracts of SBT and SGT demonstrated promising antioxidant and antibacterial potential, making them strong candidates for use as natural health boosters in food products
Genes involved in auxin biosynthesis, transport and signalling underlie the extreme adventitious root phenotype of the tomato aer mutant
No full textInternational audienceAbstract The use of tomato rootstocks has helped to alleviate the soaring abiotic stresses provoked by the adverse effects of climate change. Lateral and adventitious roots can improve topsoil exploration and nutrient uptake, shoot biomass and resulting overall yield. It is essential to understand the genetic basis of root structure development and how lateral and adventitious roots are produced. Existing mutant lines with specific root phenotypes are an excellent resource to analyse and comprehend the molecular basis of root developmental traits. The tomato aerial roots ( aer ) mutant exhibits an extreme adventitious rooting phenotype on the primary stem. It is known that this phenotype is associated with restricted polar auxin transport from the juvenile to the more mature stem, but prior to this study, the genetic loci responsible for the aer phenotype were unknown. We used genomic approaches to define the polygenic nature of the aer phenotype and provide evidence that increased expression of specific auxin biosynthesis, transport and signalling genes in different loci causes the initiation of adventitious root primordia in tomato stems. Our results allow the selection of different levels of adventitious rooting using molecular markers, potentially contributing to rootstock breeding strategies in grafted vegetable crops, especially in tomato. In crops vegetatively propagated as cuttings, such as fruit trees and cane fruits, orthologous genes may be useful for the selection of cultivars more amenable to propagation
Leveraging large-scale Mycobacterium tuberculosis whole genome sequence data to characterise drug-resistant mutations using machine learning and statistical approaches
No full textInternational audienceTuberculosis disease (TB), caused by Mycobacterium tuberculosis ( Mtb ), is a major global public health problem, resulting in > 1 million deaths each year. Drug resistance (DR), including the multi-drug form (MDR-TB), is challenging control of the disease. Whilst many DR mutations in the Mtb genome are known, analysis of large datasets generated using whole genome sequencing (WGS) platforms can reveal new variants through the assessment of genotype-phenotype associations. Here, we apply tree-based ensemble methods to a dataset comprised of 35,777 Mtb WGS and phenotypic drug-susceptibility test data across first- and second-line drugs. We compare model performance across models trained using mutations in drug-specific regions and genome-wide variants, and find high predictive ability for both first-line (area under ROC curve (AUC); range 88.3–96.5) and second-line (AUC range 84.1–95.4) drugs. To aggregate information from low-frequency variants, we pool mutations by functional impact and observe large improvements in predictive accuracy (e.g., sensitivity: pyrazinamide + 25%; ethionamide + 10%). We further characterise loss-of-function mutations observed in resistant phenotypes, uncovering putative markers of resistance (e.g., ndh 293dupG, Rv3861 78delC). Finally, we profile the distribution of known DR-associated single nucleotide polymorphisms across discretised minimum inhibitory concentration (MIC) data generated from phenotypic testing ( n = 12,066), and identify mutations associated with highly resistant phenotypes (e.g., inhA − 779G > T and 62T > C). Overall, our work demonstrates that applying machine learning to large-scale WGS data is useful for providing insights into predicting Mtb binary drug resistance and MIC phenotypes, thereby potentially assisting diagnosis and treatment decision-making for infection control
Hyperspectral Imaging for Phenotyping Plant Drought Stress and Nitrogen Interactions Using Multivariate Modeling and Machine Learning Techniques in Wheat
No full textInternational audienceAccurate detection of drought stress in plants is essential for water use efficiency and agricultural output. Hyperspectral imaging (HSI) provides a non-invasive method in plant phenotyping, allowing the long-term monitoring of plant health due to sensitivity to subtle changes in leaf constituents. The broad spectral range of HSI enables the development of different vegetation indices (VIs) to analyze plant trait responses to multiple stresses, such as the combination of nutrient and drought stresses. However, known VIs may underperform when subjected to multiple stresses. This study presents new VIs in tandem with machine learning models to identify drought stress in wheat plants under varying nitrogen (N) levels. A pot wheat experiment was set up in the glasshouse with four treatments: well-watered high-N (WWHN), well-watered low-N (WWLN), drought-stress high-N (DSHN) and drought-stress low-N (DSLN). In addition to ensuring that plants were watered according to the experiment design, photosynthetic rate (Pn) and stomatal conductance (gs) (which are used to assess plant drought stress) were taken regularly, serving as the ground truth data for this study. The proposed VIs, together with known VIs, were used to train three classification models: support vector machines (SVM), random forest (RF), and deep neural networks (DNN) to classify plants based on their drought status. The proposed VIs achieved more than 0.94 accuracy across all models, and their performance further increased when combined with known VIs. The combined VIs were used to train three regression models to predict the stomatal conductance and photosynthetic rates of plants. The random forest regression model performed best, suggesting that it could be used as a stand-alone tool to forecast gs and Pn and track drought stress in wheat. This study shows that combining hyperspectral data with machine learning can effectively monitor and predict drought stress in crops, especially in varying nitrogen conditions
Existence and Uniqueness Results to a System of Hamilton–Jacobi Equations with Application to Dislocation Dynamics
No full textInternational audienceWe study the existence and uniqueness of a nonlinear system of eikonal equations in one space dimension for any BV initial data. We present two results. In the first one, we prove the existence of a discontinuous viscosity solution without any monotony conditions neither on the velocities nor on the initial data. In the second, we show the continuity of the constructed solution under continuous initial data, and continuous velocities verifying a certain monotony condition. We present an application to a system modeling the dynamics of dislocations densitie
Automatic part segmentation for full newborn skeleton-articulated geometries using geometric deep learning and 3D point cloud
No full textInternational audienceThe development of the maternal pelvis model including a detailedfoetal model with articulated joints is of great clinical relevance. The objectiveof the present study is to propose an automatic and fast segmentation workflowof the full newborn skeleton using geometric deep learning. Computedtomography scans of 124 newborn were retrieved and manually segmented.PointNet++, a geometric deep learning algorithm, was trained to performautomatic segmentation on the 3D point clouds of the reconstructedskeletons. This method was compared with the k-means clustering approach.The PointNet++ model provided highly accurate results, with an accuracy of95.7 ± 4.7% and an IoU of 91.7 ± 7.9%, while k-means clustering providedunsatisfactory results (Accuracy = 74.6 ± 3.7% and IoU = 59.8 ± 4.7%). Thisstudy provided a powerful and accurate automatic segmentation workflow forthe full newborn skeleton