HAL-Ecole des Ponts ParisTech
Not a member yet
42743 research outputs found
Sort by
The untapped health and climate potential of cycling in France: a national assessment from individual travel data
No full textInternational audienceBackground: Promoting active modes of transportation such as cycling may generate important public health, economic, and climate mitigation benefits. We aim to assess the mortality and morbidity impacts of cycling in a country with relatively low levels of cycling, France, along with associated monetary benefits. We further assess the potential additional benefits of shifting a portion of short trips from cars to bikes, including projected greenhouse gas emissions savings.Methods: Using individual data from a nationally representative mobility survey, we described the French 2019 cycling levels by age and sex. We conducted a burden of disease analysis to assess the incidence of five chronic diseases (breast cancer, colon cancer, cardiovascular diseases, dementia, and type-2 diabetes) and the number of deaths prevented by cycling, based on national incidence and mortality data and dose–response relationships from meta-analyses. We assessed the corresponding direct medical cost savings and the intangible costs prevented based on the value of a statistical life year. Lastly, based on individual simulations, we assessed the likely additional benefits of shifting 25% of short (<5 km) car trips to cycling.Findings: The French adult (20–89 years) population was estimated to cycle on average 1 min 17 sec pers−1 day−1 in 2019, with important heterogeneity across sex and age. This yielded benefits of 1,919 (uncertainty interval, UI: 1,101–2,736) premature deaths and 5,963 (UI: 3,178–8,749) chronic disease cases prevented, with males reaping nearly 75% of these benefits. Direct medical costs prevented were estimated at €191 million (UI: 98–285) annually, while the corresponding intangible costs were nearly 25 times higher (€4.8 billion, UI: 3.0–6.5). We estimated that on average, €1.02 (UI: 0.59–1.62) of intangible costs were prevented for every km cycled. Shifting 25% of short car trips to cycling would yield approximatively a 2-fold increase in deaths prevented, while also generating important CO2 emissions reductions (0.257 MtCO2e, UI: 0.231–0.288).Interpretation: In a country with a low- to moderate-cycling culture, cycling already generates important public health and health-related economic benefits. Further development of active transportation would increase these benefits while also contributing to climate change mitigation targets
BEVContrast: Self-Supervision in BEV Space for Automotive Lidar Point Clouds
No full textInternational audienceWe present a surprisingly simple and efficient method for self-supervision of 3D backbone on automotive Lidar point clouds. We design a contrastive loss between features of Lidar scans captured in the same scene. Several such approaches have been proposed in the literature from PointConstrast, which uses a contrast at the level of points, to the state-of-the-art TARL, which uses a contrast at the level of segments, roughly corresponding to objects. While the former enjoys a great simplicity of implementation, it is surpassed by the latter, which however requires a costly pre-processing. In BEVContrast, we define our contrast at the level of 2D cells in the Bird's Eye View plane. Resulting cell-level representations offer a good trade-off between the point-level representations exploited in PointContrast and segment-level representations exploited in TARL: we retain the simplicity of PointContrast (cell representations are cheap to compute) while surpassing the performance of TARL in downstream semantic segmentation
Geometry-dependent matching pursuit: a transition phase for convergence on linear regression and LASSO
No full textGreedy first-order methods, such as coordinate descent with Gauss-Southwell rule or matching pursuit, have become popular in optimization due to their natural tendency to propose sparse solutions and their refined convergence guarantees. In this work, we propose a principled approach to generating (regularized) matching pursuit algorithms adapted to the geometry of the problem at hand, as well as their convergence guarantees. Building on these results, we derive approximate convergence guarantees and describe a transition phenomenon in the convergence of (regularized) matching pursuit from underparametrized to overparametrized models
Planification dynamique intelligente de l'évacuation et gestion en ligne à l'aide d'un système de communication véhiculaire
No full textInternational audienceDuring disasters, swiftly and efficiently evacuating populations in hazardous situations is crucial to minimize losses. This study proposes a novel framework to address dynamic population evacuation (DPE) problems, which includes planning and online evacuation management phases facilitated by vehicular communication. In the planning phase, a shelter allocation problem (SAP) is solved dynamically for destination choice and a dynamic traffic assignment (DTA) for path choice toward the chosen destinations to obtain an initial evacuation plan. The initial plan is then enhanced by employing a vehicular ad hoc network (VANET) within the vehicular edge computing (VEC) architecture. This enhancement enables communication among evacuees, allowing them to revise their vehicle's route choice and planned destination. These revisions take into account the changing risk and traffic conditions. The proposed online DPE framework is applied to the real evacuation scenario of Mill Valley City, CA. The proposed model is evaluated with different VANET architectures, including vehicular cloud computing (VCC) and VEC. The results show that the VEC framework outperforms other configurations and improves the evacuation process compared to the scenario with an initial plan by more than 30% in network clearance time. Additionally, a performance analysis is carried out for evacuation scenarios with different penetration rates of connected vehicles in VANET.Lors de catastrophes, il est crucial d'évacuer rapidement et efficacement les populations en situation de danger afin de minimiser les pertes. Cette étude propose un nouveau cadre pour traiter les problèmes d'évacuation dynamique de la population (DPE), qui comprend des phases de planification et de gestion de l'évacuation en ligne facilitées par la communication véhiculaire. Dans la phase de planification, un problème d'attribution d'abris (SAP) est résolu dynamiquement pour le choix de la destination et une affectation dynamique du trafic (DTA) pour le choix du chemin vers les destinations choisies afin d'obtenir un plan d'évacuation initial. Le plan initial est ensuite amélioré par l'utilisation d'un réseau ad hoc véhiculaire (VANET) au sein de l'architecture informatique périphérique véhiculaire (VEC). Cette amélioration permet la communication entre les personnes évacuées, ce qui leur permet de réviser le choix de l'itinéraire de leur véhicule et la destination prévue. Ces révisions tiennent compte de l'évolution des risques et des conditions de circulation. Le cadre DPE en ligne proposé est appliqué au scénario d'évacuation réel de Mill Valley City, CA. Le modèle proposé est évalué avec différentes architectures VANET, y compris l'informatique en nuage véhiculaire (VCC) et VEC. Les résultats montrent que le cadre VEC est plus performant que les autres configurations et améliore le processus d'évacuation par rapport au scénario avec un plan initial de plus de 30 % en termes de temps de dégagement du réseau. En outre, une analyse des performances est effectuée pour les scénarios d'évacuation avec différents taux de pénétration des véhicules connectés dans le réseau VANET
Straightforward Way to Provide Antibacterial Properties to Healthcare Textiles through N -2-Hydroxypropyltrimethylammonium Chloride Chitosan Crosslinking
No full textInternational audienceA straightforward and cost-effective way to coat polypropylene fibers, designed for healthcare textiles, was developed through chitosan crosslinking for antibacterial purposes. As polypropylene is an inert material, the goal was to physically trap the inert fibers through a network of crosslinked chitosan or a quaternized derivative (to enhance the antibacterial action). First, chitosan or its quaternized derivative was physically deposited by impregnation or spraying. Then, chitosan was crosslinked in glutaraldehyde solution followed by its quaternized derivative with a diisocyanate. Coated fabrics were characterized by infrared spectroscopy (IR), weight gain measurements, and scanning electron microscopy (SEM) coupled with energy dispersive X-ray (EDX). This led us to conclude that spraying provides uniform deposition while maintaining the fabric’s porosity. Acidic washing allowed us to prove that chitosan and its quaternized derivative were successfully immobilized on the fabric. Biological assays were conducted against two major strains of bacteria responsible for nosocomial infections: Staphylococcus aureus (Gram-positive) and Escherichia coli (Gram-negative). Chitosan-crosslinked samples did not show significant antibacterial activity, but the quaternized derivative allowed a significant decrease in S. aureus colonization. This study presents a simple and scalable process to coat inert fabrics with a polymer containing reactive functions potentially available to graft various additional antimicrobial agents
Macrodéchets depuis les surfaces urbaines jusqu'aux exutoires pluviaux
No full texthttps://www.youtube.com/watch?v=V8fb9y5HuisLa consommation et production de macrodéchets, notamment les macroplastiques, sont concentrées dans les zones urbaines. Une fraction de ces macrodéchets peut se retrouver abandonnée sur les surfaces urbaines et ensuite, rester sur place et entrainer une pollution des sols, ou être transférés vers les réseaux hydrographiques par les eaux de ruissellement. Les solutions curatives se multiplient aux exutoires pluviaux mais une meilleure évaluation de la source, des taux de fuite et du transfert de ces macrodéchets est nécessaire pour optimiser les actions à la source. Les résultats confirment que les eaux pluviales urbaines transfèrent des flux considérables de macrodéchets. Toutefois, ces flux représentent une faible fraction de ceux qui sont abandonnés sur les surfaces urbaines, eux-mêmes dérisoires par rapport aux proportions de déchets qui sont générées dans ces zones. Bien que la prépondérance de macrodéchets de petite taille et plastiques favorise leur transfert, notamment lors d’évènements météorologiques intenses, les actions des services de nettoiement et l’accumulation des macrodéchets dans les zones de végétation des surfaces urbaines comme dans les structures de certains avaloirs pluviaux constituent de nombreux obstacles limitant leur transfert
Topological Node2vec: Enhanced Graph Embedding via Persistent Homology
No full textFor associated repository, see https://github.com/killianfmeehan/topological_node2vecInternational audienceNode2vec is a graph embedding method that learns a vector representation for each node of a weighted graph while seeking to preserve relative proximity and global structure. Numerical experiments suggest Node2vec struggles to recreate the topology of the input graph. To resolve this we introduce a topological loss term to be added to the training loss of Node2vec which tries to align the persistence diagram (PD) of the resulting embedding as closely as possible to that of the input graph. Following results in computational optimal transport, we carefully adapt entropic regularization to PD metrics, allowing us to measure the discrepancy between PDs in a differentiable way. Our modified loss function can then be minimized through gradient descent to reconstruct both the geometry and the topology of the input graph. We showcase the benefits of this approach using demonstrative synthetic examples
Les émissions de CO2 de l’industrie française et le « ciblage carbone » des politiques publiques
No full textNote IPP n°102PrésentationQui sont les entreprises françaises qui contribuent le plus aux émissions de CO2 de l’industrie française, et quelles sont leurs caractéristiques ? Quels dispositifs de réduction des émissions de CO2 s’appliquent à elles, et selon quelles modalités ? Quel rôle insoupçonné peuvent avoir les outils fiscaux généraux, a priori sans visée environnementale ? Cette note répond successivement à ces trois questions, et propose un premier cadre d’analyse pour l’évaluation ex ante des mesures de politiques publiques à destination des entreprises. Nous documentons la distribution de l’intensité carbone dans le tissu industriel français, ainsi que les tarifications effectives du carbone auxquelles sont soumis différents types d’entreprises. Enfin, nous examinons le ciblage carbone implicite de différents dispositifs fiscaux sans visée environnementale.Résultats clésLes émissions de CO2 du secteur industriel sont extrêmement concentrées ; 10 % de la valeur ajoutée représentent 75 % des émissions de CO2.Cette forte concentration est en grande partie tirée par des effets sectoriels ; la métallurgie, la chimie, les minéraux métalliques (comme le ciment), et le papier/carton sont les secteurs les plus intenses en CO2.Deux grands régimes de tarification effective du CO2 cohabitent dans l’industrie : celle des établissements soumis au marché du carbone (SCEQE, 70 % des émissions), dont la tarification effective augmente avec le temps ; et celle des établissements hors SCEQE (30 % des émissions), gelée de 2018 à 2024.En 2019, la tarification effective du CO2 des entreprises les plus émettrices est plus faible (31€/tCO2e) que celle des entreprises les moins émettrices (47€/tCO2e). En 2022, elle est plus élevée (84€/tCO2e contre 60€/tCO2e).L’allocation de quotas gratuits, dont le volume représente, en 2022, 90 % des émissions réalisées par le secteur industriel, abaisse considérablement le poids effectif du marché carbone pour les entreprises qui y sont soumises.Par leur ciblage implicite, les dispositifs fiscaux sans visée environnementale peuvent avoir un effet sur les émissions industrielles totales.En 2019, le niveau de la contribution économique territoriale (les « impôts de production », fortement allégés dans le plan France Relance) est substantiellement plus élevé pour les 10 % des entreprises les plus intenses en CO2 (3 % de la valeur ajoutée), que pour les 10 % les moins intenses (1,2 %). Une suppression de ces impôts bénéficie donc davantage aux entreprises très émettrices
Numerical homogenization of fibrous composites with equivalent inclusion method
No full textThe equivalent inclusion method is proposed for the homogenization of fibrous composites, as an alternative to costly full-field methods and less accurate mean-field or effective-field approaches. General methodology was presented in our previous study [14], where a discretized polarization field in fibers was introduced, which took advantage of inhomogeneity slenderness. Here, a method for computing apparent 3d conductivities in periodic conditions is proposed and implemented for numerical homogenization of fibrous composites. A wide range of aspect ratios and volume fractions are investigated, and the data are used for providing analytical formulas which fit the curves
Income Inequality in South Korea, 1933-2022: Evidence from Distributional National Accounts
No full textThis study presents “Distributional National Accounts (DINA)" for South Korea. We combine household survey micro data, tax data, and national accounts to construct annual pretax income inequality series which is coherent with macro aggregates. We show the distribution of pretax national income over the period from 1933 to 2020, with detailed breakdown by age, gender, and income composition in the years from 1996 to 2020. This series allows for a much richer analysis of the long-run income inequality trend in South Korea than previous work based on fiscal tabulation (N.-N. Kim, 2018), which only includes top income shares and misses an increasing component of tax-exempted capital income in recent years. Our new series suggests that after the Asian financial crisis in 1997, income inequality has worsened due to the rise of tax-exempted capital income concentration at the top. Additionally, South Korea is characterized with relatively higher gender inequality and lower old-age income shares compared to the United States and France. Compared to other East Asian countries, South Korea exhibits relatively lower levels of income inequality, mostly due to the fact that its national income growth was more equally distributed in the early stages of economic take-off in the 1980s, even though income inequality has worsened over the last three decades. Rather strikingly, despite similar economic backgrounds and development trajectories, there is a huge gap in pretax national income inequality between Taiwan and South Korea. This gap stems mostly from the fact that the distribution of capital income in Taiwan has been much more unequal than that in Korea