Portail HAL des publications du LIRMM
Not a member yet
13279 research outputs found
Sort by
Design and Implementation of a Native Socket Interface for the Bundle Protocol
No full textInternational audienceAs space communication architectures expand beyond low-Earth orbit, traditional Internet protocols-designed for low-latency, continuous connectivity-fail to meet the needs of these challenging environments. Delay-Tolerant Networking (DTN), and in particular the Bundle Protocol (BP), addresses these constraints through a store-and-forward model that tolerates long delays and disruptions. However, application-level adoption of BP remains limited due to its non-standard APIs, which diverge from the widely-used POSIX socket interface. This paper introduces a BP POSIX Socket abstraction that exposes BP communication as a native socket type via a new address family (AF_BP) within the Linux networking stack. The proposed architecture integrates BP sockets with existing system tools and kernel primitives while delegating protocol processing to a userspace daemon via Netlink. We validate the approach through the development of a Rust-based messaging application. Our results demonstrate that minimal code changes are required to adapt existing applications to DTN environments, and that user experience can be preserved-even improved-through delayaware interface designs. The BP socket abstraction thus presents a practical and extensible path for broadening the operational reach of DTN technologies
An Improved Threshold Homomorphic Cryptosystem Based on Class Groups
No full textInternational audienceWe present distributed key generation and decryption protocols for an additively homomorphic cryptosystem based on class groups, improving on a similar system proposed by Braun, Damgård, and Orlandi at CRYPTO '23. Our key generation is similarly constant round but achieves lower communication complexity than the previous work. This improvement is in part the result of relaxing the reconstruction property required of the underlying integer verifiable secret sharing scheme. This eliminates the reliance on potentially costly proofs of knowledge in unknown order groups. We present a new method to batch zero-knowledge proofs in unknown order groups which strengthens these improvements. We also present a protocol which is proven secure against adaptive adversaries in the single inconsistent player (SIP) model. Our protocols are secure in the universal composability (UC) framework and provide guaranteed output delivery. We demonstrate the relative efficiency of our techniques by presenting the running times and communication costs associated with our implementation of the statically secure protocol and provide a direct comparison with alternate state of the art constructions
Triple-Node-Upset Recovery High-Impedance-State-Insensitive and Single-Event-Transient Filtering Latch for Aerospace Applications
No full textInternational audienceThis article proposes a 4 × 4-device-Matrix-based Cost-optimized Triple-node-upset (TNU)-recovery High-impedance-state (HIS)-insensitive and Single-event-transient (SET) filtering Latch, namely MCTHSL, designed for aerospace applications. The MCTHSL comprises a 4 × 4 device matrix to completely provide TNU recovery, and thus, it is HIS insensitive. The input of a Schmitt trigger (ST) is split to create an input-split Schmitt trigger, namely ISST, to simultaneously provide the functions of a C-element (CE), an SET filtering device, and a delay element. The matrix comprises 15 mutually interlocking two-input CEs and an ISST to store values, recover from TNUs, and filter SETs with HIS insensitivity and cost optimization. Simulation results demonstrate the aforementioned features and moderate sensitivity to the impacts of process, voltage, and temperature variations of our proposed MCTHSL in comparison with the existing latches with TNU recovery. The features are achieved at the cost of moderate overhead in terms of area, delay, and power
Tip-Growing Robots: Design, Theory, Application
No full textInternational audienceGrowing robots apically extend through material eversion or deposition at their tip. This endows them with unique capabilities such as follow the leader navigation, long-reach, inherent compliance, and large force delivery bandwidth. Tip growing robots can therefore conform to sensitive, intricate, and difficult-to-access environments. This review paper categorizes, compares, and critically evaluates state-of-the-art growing robots with emphasis on their designs, fabrication processes, actuation and steering mechanisms, mechanics models, controllers, and applications. Finally, the paper discusses the main challenges that the research area still faces and proposes future directions
Closure Versus Dualization: A Comparison of Software Tools for Formal Concept Analysis
No full textInternational audienceFormal concept analysis revolves around the notion of formal concept. Many tools have been developed to compute these concepts, based on the fact that the extents and intents are closed under the closure operators given by the context. However, concepts are also dual to the complement of the incidence relation of the context, a fact that is rarely discussed in the community. In this paper, we compare tools developed in the FCA community against dualization-based tools on runtime in both real and artificial datasets. The experimental observations show that dualization-based tools are competitive and more resistant to increases in context density
A Distraction Knee-Brace and a Robotic Testbed for Tibiofemoral Load Reduction During Squatting
No full textInternational audienceWe design and evaluate a new knee distraction unloader brace. The proposed device conforms to the nonlinear behavior of the tibiofemoral contact force during squat motions, by means of patient-custom cams. Using pneumatic cylinders as springs, the unloading assistance provided by the brace is tailored to the patient's pathology and adjusted during the rehabilitation process. To assess the performance of our orthosis, various tests are conducted to evaluate its efficiency in terms of tibiofemoral contact load reduction. For this purpose, a robotic test-bench, equipped with a robotic arm, emulates upper leg motion under applied forces (hybrid force-motion control). A pseudo-leg is attached to the robot end-effector, and the orthosis is mounted onto it. The test bench is instrumented with two six degrees of freedom force-torque sensors. Using these force sensors as ground truth, tibiofemoral contact force measurements are obtained with and without our orthosis and compared. A pair of cams is fabricated based on data from a patient whose information is retrieved from the Orthoload database. Experimental results demonstrate a contact force reduction of up to 100% within the force range corresponding to the robot's maximum capacity
A new fractional‐order L1 adaptive feedback control of dynamical systems: Design, stability analysis, and application
No full textInternational audienceThis study focuses on the development of a new fractional‐order adaptive feedback control strategy of dynamical systems. Specifically, the original adaptive controller is revisited based on the consideration of fractional‐order filters using the Oustaloup recursive approximation. To show the effectiveness and superiority of the proposed fractional‐order filter, a comparative study is conducted against classical integer‐order filters based on both time analysis using performance metrics such as the integral of absolute error (IAE) and energy consumption, and frequency analysis, including Nyquist criterion and stability margins. The proposed control scheme is validated, on an electro‐hydraulic system (EHS), through several numerical simulation scenarios to show enhanced performance and robustness. Results indicate up to 86% improvement in tracking performance compared to the original controller
Evaluating deep learning models for plant protein function prediction
No full textNational audiencePredicting the functions of proteins remains a critical yet challenging task in computational biology. Advances in high-throughput sequencing, the expansion of protein databases, and the continuous development of artificial intelligence have led to the emergence of many computational methods dedicated to protein function prediction. In this study, we evaluated the performance of four state-of-the-art models -DeepGOPlus, DeepGraphGO, DeepGOZero, and DeepGOSE -using experimentally annotated proteins from the UniProt-KB/Swiss-Prot database. We also trained and tested these models on species-specific datasets from Arabidopsis thaliana and Oryza sativa to investigate their potential and applicability in plant protein studies. Our results showed that DeepGOPlus consistently achieved the best evaluation scores across all datasets. DeepGOSE and DeepGOZero performed comparably and only marginally outperformed Deep-GraphGO in certain training attempts. Further analysis revealed that dataset stratification into training, validation, and testing sets introduced variations in Gene Ontology annotation specificity, which may have influenced model performance
Largest planar graphs of diameter 3 and fixed maximum degree - connection with fractional matchings
No full textThe degree diameter problem asks for the maximum possible number of vertices in a graph of maximum degree ∆ and diameter D. In this paper, we focus on planar graphs of diameter 3. Fellows, Hell and Seyffarth (1995) proved that for all ∆ ⩾ 8, the maximum number np ∆,D of vertices of a planar graph with maximum degree at most ∆ and diameter at most 3 satisfies 9∆/2 -3 ⩽ np ∆,3 ⩽ 8∆ + 12. We show that the lower bound they gave is optimal, up to an additive constant, by proving that there exists c > 0 such that np ∆,3 ⩽ 9∆/2 + c for every ∆ ⩾ 0. Our proof consists in a reduction to the fractional maximum matching problem on a specific class of planar graphs, for which we show that the optimal solution is 9/2 , and characterize all graphs attaining this bound
DNA replication timing and Copy Number Variations are confounders of ATAC-Seq and RNA-DNA interaction data
No full textInternational audienceMeasuring chromatin-RNA contacts is the first step to understand their role(s) in gene expression. Several technologies have been developed to detect genome-wide RNA-DNA interactions, each with different protocols and computational analysis methods. While these methods use various statistical tests, they all include RNA-focused corrections to account for confounding factors such as spurious trans-chromosomal mRNA-DNA interactions, RNA abundance and distance between the RNA-emitting gene and RNA-receiving DNA region. However, the RNA-DNA interaction counts can also be biased by the amount of DNA regions that can be sequenced and which varies along the chromosomes due notably not only to Copy Number Variations (CNVs) but also DNA replication timing (RT). DNA is replicated in a precise spatiotemporal program, which is local, cell-type specific and conserved in evolution, that causes some genomic regions to be duplicated earlier than others during the cell cycle. Consequently, depending on the number of dividing cells, early-replicating regions are more likely to be overrepresented in sequencing data, especially in proliferating cells, leading to uneven DNA coverage across the genome. Using RADICL-seq data1 from the FANTOM consortium across multiple cell types, we show that, despite RNA-centric normalization, chromatin-RNA contact counts remain correlated with both CNVs and replication timing. Since both CNVs and RT can be inferred from Whole Genome Sequencing (WGS) 2,3, we propose this data to estimate and correct for local DNA abundance. To evaluate our approach's validity, we use ATAC-seq data, which strongly correlates with RADICL-seq signals. We demonstrate that the newly predicted peaks are more biologically relevant