Portail HAL des publications du LIRMM
Not a member yet
13279 research outputs found
Sort by
SARRA, histoire d'un modèle de simulation des cultures pour les zones intertropicales
No full textInternational audienceThe SARRA (Regional Analysis System for Agroclimatic Risks) model family has been under development by the French Agricultural Research Centre for International Development (CIRAD) since the 1970s. Used in intertropical regions, the early versions of the model were designed to assess the impact of water stress—that is, the lack of accessible water for crops—on production and seedling losses. Their objectives were to improve riskestimation, optimize sowing management, and serve as a regional-scale early warning system for food security. Since the 2000s, the model has been increasingly used to analyze the impacts of climate change on agroclimatic risks. Several developments have enabled the integration of, on the one hand, the genetic characteristics of crops, and on the other hand, a better consideration of spatio-temporal dynamics, particularly through the use of remote sensing. Additionally, SARRA has gradually evolved from a descriptive model into a scenario-based anticipation tool, requiring structural transformations and the adoption of modern programming and versioning techniques to ensure traceability and reproducibility of simulations. Recognized for their simplicity, parsimony, and robustness, these models have proven effective for monitoring agricultural seasons, managing foodsecurity, and supporting decision-making in the face of climate risks. This article provides a retrospective analysis of their development, highlighting major innovations and their potential to guide resilient agricultural practices in response to current climate challenges.La famille de modèles SARRA (Système d'Analyse Régional des Risques Agroclimatiques) est développée par le CIRAD depuis les années 1970. Utilisées dans les régions intertropicales, les premières versions du modèle permettaient d'évaluer l'impact des stress hydriques, c'est-à-dire le manque d'eau accessible par la culture, sur sa production ainsi que les pertes de semis. Elles avaient pour objectif d'améliorer l'estimation des risques, la gestion des semis et de servir de système d'alerte à l'échelle régionale pour la sécurité alimentaire. À partir des années 2000, le modèle a été sollicité pour analyser les effets du changement climatique sur les risques agroclimatiques. Plusieurs évolutions ont permis d'intégrer, d'une part, les caractéristiques génétiques des plantes et, d'autre part, une meilleure prise en compte des dynamiques spatio-temporelles, notamment par le biais de la télédétection. Par ailleurs, SARRA a progressivement évolué d'un modèle descriptif vers un outil d'anticipation et de simulation de scénarios, nécessitant des transformations structurelles et l'adoption de techniques modernes de programmation et de versioning afin d'assurer la traçabilité et la reproductibilité des simulations.Reconnus pour leur simplicité, parcimonie et robustesse, ces modèles se sont avérés efficaces pour le suivi des campagnes agricoles, la gestion de la sécurité alimentaire et la prise de décision face aux risques climatiques. Cet article propose une rétrospective de leur développement et met en lumière les innovations majeures ainsi que leur potentiel pour guider des pratiques agricoles résilientes face aux défis climatiques actuels
Robust Adaptive STA-Based Tracking Control of AUVs with Real-Time Experiments
No full textInternational audienceThe complex nonlinear dynamics, significant parametric uncertainties, and external disturbances, inherent to underwater vehicles require robust control schemes to ensure reliable operations. This paper introduces a novel adaptive gains variant of the Super-Twisting Algorithm (STA) and compares its performance with two existing adaptive STA-based controllers. Real-time experiments are proposed to demonstrate that the proposed controller enhances robustness toward external disturbances and parametric uncertainties, achieving better tracking performance. The paper also highlights the potential of robust controllers for underwater vehicle applications
Exome Sequencing Detects Uniparental Disomy of Chromosome 4 Revealing a LARP7 Pathogenic Variant Responsible for Alazami Syndrome: A Case Report
No full textInternational audienceAlazami syndrome is an autosomal recessive disease characterized by global developmental delay, growth restriction, and distinctive facial features. Fewer than 50 individuals are currently reported with biallelic loss of function variants in LARP7 . We report the case of a 3.5‐year‐old boy born from nonconsanguineous parents, presenting with syndromic global developmental delay. Exome sequencing identified a homozygous frameshift pathogenic variant in LARP7 . Parental analysis failed to detect the variant in the paternal sample, although the father's biological paternity was confirmed. Targeted secondary bioinformatic analyses at the LARP7 locus suggested a 45 Mb loss of heterozygosity (LOH), further confirmed by a single nucleotide polymorphism array that identified four LOH regions on chromosome 4, including one encompassing LARP7. This LOH exposes the recessive LARP7 pathogenic variant, resulting in the manifestation of Alazami syndrome. To our knowledge, this is the first reported case of Alazami syndrome due to uniparental disomy (UPD). UPD is a rare cause of autosomal recessive disorders. Its identification is crucial for genetic counseling to adjust recurrence risk for siblings. This case highlights the effectiveness and usefulness of bioinformatics algorithms applied to next generation sequencing in detecting such events
A novel adaptive approach to underwater ROV control: design and real-time experiments
No full textInternational audienceThis paper aims to develop and assess the efficacy of an Adaptive Backstepping Controller and its nonlinear variant in regulating the behavior of underwater Remotely Operated Vehicles, with a specific focus on the Leonard ROV. The performances of these controllers are compared against a conventional Proportional-Integral-Derivative controller across three distinct scenarios: nominal depth-yaw tracking, robustness testing against perturbations and adaptability to mass changes. To quantify the control performance in each scenario, the study employs two metrics: Root Mean Square Error and the Integral of the control input. These metrics provide insights into the accuracy and consistency of depth and yaw tracking over time. The results of the evaluation demonstrate the superior performance of the proposed methods, particularly in managing complex dynamics of underwater environments, where precise control is critical for mission success
"Only ChatGPT gets me": An Empirical Analysis of GPT versus other Large Language Models for Emotion Detection in Text
No full textformerly International World Wide Web ConferenceInternational audienceThis work investigates the capabilities of large language models (LLMs) in detecting and understanding human emotions through text. Drawing upon emotion models from psychology, we adopt an interdisciplinary perspective that integrates computational and affective sciences insights. The main goal is to assess how accurately they can identify emotions expressed in textual interactions and compare different models on this specific task. This research contributes to broader efforts to enhance human-computer interaction, making artificial intelligence technologies more responsive and sensitive to users' emotional nuances. By employing a methodology that involves comparisons with a state-of-the-art model on the GoEmotions dataset, we aim to gauge LLMs' effectiveness as a system for emotional analysis, paving the way for potential applications in various fields that require a nuanced understanding of human language
NVSRLO: A FeFET-Based Non-Volatile and SEU-Recoverable Latch Design with Optimized Overhead
No full textInternational audienceThis paper presents a FeFET-based non-volatile and single-event upset (SEU) recoverable latch, namely NVSRLO, which does not require any extra control signals. Simulation results show that the proposed latch provides non-volatility and SEU-recovery with optimized overhead. Compared with existing non-volatile latches, NVSRLO significantly reduces delay, power, and delay-power-area product at the cost of area
Area Optimized Architectures for Galois Counter Mode
No full textInternational audienceThis work presents new hardware architectures for the Galois Counter Mode of operation used for authenticated encryption/decryption. In common Galois Counter Mode implementations, Advance Encryption Standard in counter mode is used for encryption and decryption. Authentication tag is computed by repeated multiplication and addition of encrypted message in the field F 2 128 . Sub-quadratic Toeplitz matrix vector product based binary multiplier architectures are used for the implementation of binary field F 2 128 . For optimal implementation, encryption/decryption function and the multiplication in F 2 128 should be balanced in time, i.e., take approximately the same number of clock cycles to perform a single operation on one block. The architectures proposed in this paper are designed to offer low to medium throughputs with logic requirements that reflect the performance level. We report and compare implementation results of three different architectures using application specific integrated circuits
Design and experiments of an adaptive disturbance observer for tracking control of autonomous underwater vehicles
No full textInternational audienceUnderwater vehicles, controlled by human operators, often employ a simple internal control scheme, such as the Proportional Derivative (PD) controller. The choice of a PD controller is driven by its simplicity, ease of implementation, and acceptable reliability. However, this type of controller often exhibits limited robustness when facing parametric uncertainties (e.g., salinity variations) and external disturbances, such as sea waves. Motivated by these inherent limitations, we designed an adaptive disturbance observer using the Super-Twisting Algorithm. The primary objective of this observer is to enhance the robustness of the PD controller without compromising its simple and straightforward structure and minimizing the effort required for its tuning. Our proposed observer takes inspiration from the Extended State Observer (ESO) and is supported by a rigorous stability analysis using Lyapunov techniques. We also establish the stability of the closed-loop system involving the controller and observer. To showcase the effectiveness and reliability of our proposed method, we carried out a set of real-time experiments under various operating conditions
Leveraging Data Seasonality and Matrix Profile for Anomaly Detection: Application to Climate Time Series
No full textSeasonal time series analysis is fundamental in domains such as climate science, where detecting and understanding anomalies, patterns, and data changes are essential. The classical Matrix Profile approach does not consider the data's seasonality, failing to detect seasonal anomalies and patterns. This paper introduces the Interval Matrix Profile, a novel extension of the Matrix Profile specifically designed for analyzing periodic and seasonal time series data. The Interval Matrix Profile enables flexible interval-based comparisons across seasons, allowing the detection of anomalies that conventional approaches miss. We further propose the constrained k Nearest Neighbor Interval Matrix Profile, designed to identify anomalies that may appear across multiple periods, a common characteristic of abnormal climate events and extreme weather phenomena. Our approach leverages a scalable block-based algorithm that achieves significant performance gains through caching, vectorization, and parallelism. Additionally, we introduce a novel methodology to detect the first or last occurrence of a pattern, enabling the discovery of pattern emergence or disappearance within seasonal time series. The algorithms are demonstrated in case studies on temperature climate time series. They effectively capture seasonal anomalies and find pattern disappearance. Our results illustrate that the IMP consistently outperforms the classical Matrix Profile in the accuracy of seasonal anomaly detection and computational efficiency
Relational Concept Analysis: Where Formal Concepts Meet Description Logics
No full textNational audienceFormal Concept Analysis (FCA) originates from lattice theory and Galois connections. It has seen significant developments in the field of knowledge discovery. Two main artifacts derived from this theory are used in practice: conceptual structures (lattices or specific suborders) and sets of propositional logical rules with various properties. Graphical and Logical views complement each other. Relational Concept Analysis (RCA) has been introduced as an extension of FCA to account for datasets consisting of entities described by attributes and relationships between these entities, whereas FCA only considers entities described by Boolean attributes. The construction of concepts and rules is carried out using logical quantification operators, which immerses RCA into the domain of description logics. We will introduce the different notions, present some tools, and showcase a practical application