Search

Background Bariatric surgery can influence the presentation, diagnosis, and management of gastrointestinal cancers. Oesophago-Gastric (OG) malignancies in patients who have had a prior bariatric procedure have not been fully characterised. Objective To characterise OG malignancies after bariatric procedures. Setting University Hospital, United Kingdom. Methods We performed a retrospective, multi-centre observational study of patients with OG malignancies after bariatric surgery to characterise this condition. Results This study includes 170 patients from 75 centres in 25 countries who underwent bariatric procedures between 1985 and 2020. At the time of the bariatric procedure, the mean age was 50.2 ± 10 years and the mean weight 128.8 ± 28.9 kg. Females comprised 57.3% (n=98) of the population. Most (n=64) patients underwent a Roux-en-Y Gastric Bypass (RYGB) followed by Adjustable Gastric Band (AGB) (n = 46) and Sleeve Gastrectomy (SG) (n = 43). Time to cancer diagnosis after bariatric surgery was 9.5 ± 7.4 years and mean weight at diagnosis was 87.4 ± 21.9 kg. The time lag was 5.9 ± 4.1 years after SG compared to 9.4 ± 7.1 years after RYGB and 10.5 ± 5.7 years after AGB. One-third of patients presented with metastatic disease. The majority of tumours were adenocarcinoma (82.9%). Approximately 1 in 5 patients underwent palliative treatment from the outset. Time from diagnosis to mortality was under one year for most patients who died over the intervening period. Conclusions The OGMOS (Oesophago-Gastric Malignancies after Obesity/ Bariatric Surgery) study presents the largest series to date of patients developing OG malignancies after bariatric surgery and attempts to characterise this condition.

International audience; IntroductionSteganography, the art to hide information inside host media like pictures and movies, and steganalysis, its countermeasure attempting to detect the presence of an hidden information within an innocent-looking document, are frequently reported as promising information security techniques for telemedicine. For the past few years, in the race between image steganography and steganalysis, deep learning has emerged as a very promising alternative to steganalyzer approaches based on rich image models combined with ensemble classifiers. A key knowledge of image steganalyzer, which combines relevant image features and innovative classification procedures, can be deduced by a deep learning approach called convolutional neural networks (CNN). This kind of deep learning networks is so well-suited for classification tasks based on the detection of variations in 2D shapes that it is the state-of-the-art in many image recognition problems.Materials and methodsWe design a CNN-based steganalyzer for images obtained by applying steganography with a unique embedding key. The proposed CNN has a quite different shape compared to the ones resulting from the earlier works, and it is able to provide high detection accuracy for several steganographics tools when the same stego key is reused during the embedding process. The convolutional part of our proposal starts by a global filtering, using a single filter, followed by a second convolutional layer that produces a reduced set of high-level features (256 features for 512 × 512 pixels input images) thanks to the use of large filters.ResultsThe proposed architecture embeds less convolutions, with much larger filters in the final convolutional layer, and is more general: it is able to deal with larger images and lower payloads. For the “same embedding key” scenario, our proposal outperforms all other steganalyzers, in particular the existing CNN-based ones, and defeats many state-of-the-art image steganography schemes. The information encoded by the final vector of features is so discriminating that the classifier part can be reduced to only two output neurons. We finally evaluated the detection ability of the CNN against two spatial domain steganographic schemes and a frequency domain one. More precisely, we designed a perfect steganalyzer for embedding payloads of 0.4 bit per pixel, and for all the steganographic tools investigated in this article (working either in spatial or in frequency domains). Rather interesting results have been obtained too, albeit to a lesser extent, for a payload value of 0.1 bpp.Discussion and conclusionsThe obtained results are very encouraging, and they outperform all the previous deep learning proposals for steganalysis. A first step in the design of a universal detector has been achieved too, as we are able to detect HUGO based hidden messages even when a WOW steganographier has been used during the training stage. These results allow us to propose to add fragile watermarks on media like pictures or pdf medical documents, to guarantee the authenticity of the material: any attempt of modification of the support will alter the watermark, proving by doing so the modification. Another application is to add personal and medical information inside medical images.RésuméIntroductionLa stéganographie, l’art de dissimuler de l’information au sein de médias numériques tels que les images et les vidéos, et la stéganalyse, sa contre-mesure tentant de détecter la présence d’une information secrète dans un document semblant innocent, sont fréquemment signalés comme étant des techniques de sécurité intéressantes pour la télémédecine. Au cours des dernières années, l’apprentissage profond (deep learning) a émergé dans la compétition entre stéganographie et stéganalyse, paraissant une alternative intéressante dans le cadre des stéganalyseurs basés sur des modèles raffinés d’image couplés avec des classifieurs d’ensembles. On peut alors combiner une sélection bien choisie de caractéristiques d’images avec des procédures innovantes de classification, au travers d’une approche d’apprentissage approfondie moderne appelée les réseaux de neurones convolutionnels (RNC). Ces réseaux, qui sont faits pour de la classification basée sur de la détection de variations dans les formes 2D, produisent la plupart du temps les meilleurs résultats dans divers problèmes de reconnaissance d’images.Matériel et méthodesNous avons programmé un stéganalyseur basé sur les RNC, pour des images stéganographiées avec une seule et même clé secrète. Le RNC proposé a une structure assez différente des réseaux développés jusqu’ici, et il est capable de fournir un fort taux de détection sur de nombreux logiciels de stéganographie, sous l’hypothèse que l’adversaire n’utilise qu’une seule clé d’embarquement. La partie « convolutionnelle » de notre approche commence par un filtrage global n’utilisant qu’un seul filtre, suivi par une seconde couche convolutionnelle, qui produit un ensemble réduit de caractéristiques de haute qualité (256 caractéristiques pour 512 × 512 pixels) grâce à l’utilisation de larges filtres.RésultatsL’architecture proposée embarque un plus faible nombre de convolutions que ce qui se trouve habituellement dans la littérature. Les filtres sont plus larges dans la dernière couche de convolution, et le réseau dans son ensemble est plus général : il permet de considérer des images plus larges, et de plus faibles taux d’embarquement. Pour le scénario d’une seule clé, nos résultats sont meilleurs que tous les autres stéganaliseurs, dont ceux basés sur des RNC, et nous pouvons mettre en échec de nombreux outils de stéganographie. L’information encodée par le vecteur final de caractéristiques est si discriminant, que la partie de classification peut se réduire à deux neurones de sortie. Nous avons finalement évalué la capacité de détection de notre RNC contre deux algorithmes de stéganographie fonctionnant dans le domaine spatial, et un dans le domaine fréquentiel. Nous avons obtenu un stéganalyseur parfait pour un taux d’embarquement de 0,4 bit par pixel, et ce pour tous les algorithmes considérés dans cet article. Des résultats intéressants, quoique moins surprenants, ont aussi été obtenus avec un taux d’embarquement de 0,1 bpp.Discussion et conclusionLes résultats obtenus sont très encourageants, et ils améliorent les résultats obtenus jusqu’à présent dans l’apprentissage approfondis pour la stéganalyse. Ces résultats nous permettent d’envisager l’utilisation de la stéganographie pour la télémédecine, comme l’ajout d’une marque fragile dans des documents médicaux, pour en garantir leur authenticité, ou l’insertion des informations de patients dans des images médicales.KeywordsTelemedicine security; Steganography; Steganalyzis; Deep learningMots clésSécurité pour la télémédecine; Stéganographie; stéganalyse; Apprentissage approfondis

Over the past decade, reported incidence of human metapneumovirus (hMPV) has increased owing to the use of molecular assays for diagnosis of respiratory viral infections in cancer patients. The seasonality of these infections, differences in sampling strategies across institutions, and small sample size of published studies make it difficult to appreciate the true incidence and impact of hMPV infections. In this systematic review, we summarized the published data on hMPV infections in hematopoietic cell transplant recipients and patients with hematologic malignancy, focusing on incidence, hMPV-associated lower respiratory tract infection (LRTI), mortality, prevention, and management with ribavirin and/or intravenous immunoglobulins. Although the incidence of hMPV infections and hMPV-associated LRTI in this patient population is similar to respiratory syncytial virus or parainfluenza virus and despite lack of directed antiviral therapy, the mortality rate remains low unless patients develop LRTI. In the absence of vaccine to prevent hMPV, infection control measures are recommended to reduce its burden in cancer patients.

Hematopoietic cell transplant (HCT) recipients have lower immune response to influenza vaccination and are susceptible to lower respiratory tract infection (LRI) and death. We determined clinical characteristics and outcomes of laboratory-confirmed influenza, including 2014/H3N2 infection, in 146 HCT recipients. An immunodeficiency scoring index (ISI) was applied to identify patients at high risk for LRI and death. Thirty three patients developed LRI (23%), and 7 died within 30 days of diagnosis (5%). Most patients received antiviral therapy (83%); however, only 18% received it within 48 hours of symptom onset. The incidence of LRI was significantly higher in the ISI high-risk group than in the low-risk group (P < 0.001). Receiving early antiviral therapy was associated with a substantial reduction in LRI for all ISI risk groups with the greatest risk reduction observed in the high-risk group. When compared to previous seasons, no significant differences in patient outcomes were observed during the 2014/H3N2 season; however, antiviral therapy was more promptly initiated in the latter season. The ISI that was originally developed for respiratory syncytial virus may help identifying HCT recipients at risk for progression to LRI and mortality following influenza infection. These patients should be monitored more closely. Early initiation of antiviral therapy for influenza in HCT recipients, irrespective of the ISI risk group, may improve morbidity as well as mortality.

Equivalent open chains generating the same motion as RC dyad with parallel axes are introduced based on equivalencies stemming from the algebraic structure of Lie group of the set of Euclidean displacements. These equivalent chains can be used to replace the RC limbs to deduce a family of RC // RC ― ― –like linkages. Cylindrical translation and exceptional mobility of 1-DoF translational linkages having RC limbs or equivalencies are revealed using the Lie-group property of the rigid-body displacement set. The intersections of two circular cylinders offer classification of various types of RC // RC ― ― linkages and the like. The discontinuous mobility of motion bifurcation in special RC // RC ― ― -like linkages is disclosed. General cases for the same phenomenon occur when an offset of translation displaces two RC subchains or equivalencies. At a singular posture of bifurcation, the coupler-bar locally has 2-DoF of infinitesimal translation and two working modes arise at the position of possible double point on the bicylindrical curve generated by two RC open sub-chains or equivalencies. Finally, the special case of Koenigs joint is described for the potential application in machinery.

Hydrogels have a wide range of applications from medical devices to tissue engineering to industrial health products. The numerical modeling of these porous structures can provide insight into their in-situ performance. However, an accurate constitutive model to replicate the stress-strain behavior of the swelling process is essential in ensuring the efficacy of the simulation. This study presents a strain dependent constitutive model for describing the finite deformation of superabsorbent polymers undergoing solvent induced swelling. Like many elastic materials, stretch-induced softening and hardening of these polymers occurs at large deformations. In order to incorporate the deformation dependent modulus into the new model, the shear modulus of sodium polyacrylate gels were measured using a rheometer as a function of swelling. Through numerical simulations on spherical gels, the effect of both cross-link density and the dependent modulus is investigated against a control model. As experimental quantification of the initial porosity is difficult and often not consistent between samples, Monte-Carlo simulations are implemented to experimentally verify the new model. Furthermore, this model is applied to experimental uniaxial tension data of incompressible rubber to allow comparison to other strain hardening constitutive models.

The circular translation generated by a pseudo-planar parallelogram with four screw pairs having equal pitches and parallel axes is introduced and a special Delassus parallelogram producing the same motion is obtained through the addition of a second pseudo-planar parallelogram with another value of its equal pitches. A mechanical generator of spherical translation stemming from the circular translation generator is revealed. Combining a spherical translation generator and a pseudo-planar or special Delassus parallelogram leads to a new family of 6-screw (6H) paradoxical linkages with circular translation in a purely geometric way. Their architectural type belongs to the structural group with two alternate pairs of joint-axes parallel and one pair of opposite joint-axes parallel. Moreover, the one-degree-of-freedom (1-DoF) movability is conditioned by metric constraints. The already published 6R linkages having circular translation are special types of 6H with all screw-pitches zero. The parallel arrangement of three special 6R linkages produces a 3-UU parallel platform with circular translation (U denotes a universal joint with intersecting joint axes of revolute pairs). A more overconstrained m-UU case (m ≥ 4) with special joint-axes arrangements is disclosed too.

ObjectiveThe incidence of postpartum anemia is high. Current therapy consists of iron supplementation or blood transfusions, based on the assumption that these treatments improve health status (HS) and reduce fatigue. The aim of this study was to compare HS and fatigue in postpartum women with and without anemia.Study designThis prospective cohort study was performed in The Netherlands between April 2008 and August 2010 and involved 112 anemic (hemoglobin [Hb] < 10.5 g/dL) and 108 non-anemic (Hb ≥ 10.5 g/dL) women. The anemic women received oral iron supplementation. Within 48 h and 5 weeks after delivery, HS was measured using the 36 item Short-Form Health Survey (SF-36) and fatigue was measured using the Checklist Individual Strength (CIS). ANOVA for repeated measures was used to compare HS and fatigue scores among groups and across time.ResultsAfter adjustment for confounding variables, there were no differences in any of the HS and fatigue scores. HS and fatigue seem to be more influenced by a complicated delivery than by anemia. HS and fatigue scores significantly improved over time in all women.ConclusionHS and fatigue were not different among women with and without postpartum anemia.Keywords: Health status, Fatigue, Anemia, Iron deficiency, Postpartum

Bricard one-DoF motion (abbreviated to BM) is a type of spatial motion in which all points of the moving body have spherical trajectories in the three-dimensional (3D) Euclidean space. This paper focuses on the mechanical realization of BM. The invariance characteristics and a visualization of three types of BM are provided in detail and the criterion for a possible full revolution is clarified too. The primitive Bricard motion generators (BMGs) are first proposed and geometrically verified. With the synthetic reasoning via the self-conjugation of any cylindrical 1-DoF motion and the invariance property of BM, two or more-loop BMGs are then synthesized systematically. These BMGs include isoconstrained mechanisms and overconstrained ones and the latter are exceptional and paradoxical chains. Besides, the platforms having BM, such as the Wren platform, LADD motion converter, algebraic screw joint, Griffis–Duffy (GD) type platform, etc. are clearly explained. A generalization of the GD platform and the general Verheyen-type mechanism with or without offset are newly revealed as well.

International audience; As any well-designed information security application uses a very large quantity of good pseudorandom numbers, inefficient generation of these numbers can be a significant bottleneck in various situations. In previous research works, a technique that applies well-defined discrete iterations, satisfying the reputed Devaney's definition of chaos, has been developed. It has been proven that the generators embedding these chaotic iterations (CIs) produce truly chaotic random numbers. In this new article, these generators based on chaotic iterations are redesigned specifically for Field Programmable Gate Array (FPGA) hardware, leading to an obvious improvement of the generation rate. Analyses illustrate that statistically perfect and chaotic random sequences are produced. Additionally, such generators can also be cryptographically secure. To show the effectiveness of the method, an application in the information hiding domain is finally proposed.

A novel special motion called helical Cardan motion is introduced. The well known Cardanic motion is a special case with a zero pitch. A Cardan motion is a planar cycloidal motion, which can be produced by the internal rolling of a circle on a fixed circle whose diameter is twice the moving circle diameter. The newly studied motion is a combination of a Cardan motion with a helical motion. Any point lying on the surface of a cylinder having the smaller rolling circle as its cross section traces out a planar sinusoidal curve. Frame-free vector calculations are used to establish the essential properties of the special motion. Its corresponding characteristics are further utilized to make new derivations of related paradoxical chains, namely the new isosceles triangle HHH P chain, the HHHH rhomboid of Delassus and the two-loop Sarrus-type 6H-1P chain. A generalization of Sarrus' mechanism for the generation of a rectilinear translational motion is derived as well.

The product of two Schoenflies motion subgroups of the group of general displacements characterizes a noteworthy type of 5-dimensional (5D) displacement set called double Schoenflies or X–X motion. It includes any spatial translation and any two sequential rotations whose axes are parallel to two given independent vectors. The canonical irreducible representation of the 5D set of X–X motion is utilized to study various intersection sets of three distinct X–X motions. Three remarkable possibilities: 3-DoF translation motion, 3-DoF translation and 1-DoF infinitesimal rotation, and 3-DoF translation and 2-DoF infinitesimal rotation are discussed. Then we disclose the corresponding mechanisms with a mixture of finite and infinitesimal DoFs. Novel parallel mechanisms with 3-DoF finite translation and 2- or 1-DoF infinitesimal rotation are proposed. Parallel mechanisms whose infinitesimal mobility is specific of a singularity locus are derived, too. Meanwhile, it is established that the well-known SNU 3-UPU parallel manipulator with 3-DoF finite translation also has a mobility of 2-DoF infinitesimal rotation for a set of singular postures including its home configuration.

Based on the geometric properties of the oblique circular or elliptic cylinder and an invariant subset of translations, a family of related chains having a paradoxical mobility is systematically synthesized. First of all, vector calculation is used to describe the plane symmetry operation. The plane symmetry property of an oblique circular (OC) cylinder, which has circular directrices, is verified. Next, we introduce a mechanical generator of 2-DoF translation along the surface of an OC cylinder and derive a novel PaPPa paradoxical chain, where Pa denotes the composite joint of a 4-R hinged parallelogram with parallel axes. Then, we establish its corresponding RR ¯ P RR ¯ ¯ chain. The implementation of pseudo-planar and Delassus parallelograms leads to more general chains of type PHPPH, where PH designates the composite joint of a 4-H parallelogram with four parallel axes. The related HH ¯ P HH ¯ ¯ chain having two couples of H pairs with parallel axes and equal pitches separately is a new 1-DoF paradoxical linkage.

Schoenflies motion is often termed X-motion for conciseness. A set of X-motions with a given direction of its axes of rotations has the algebraic properties of a Lie group for the composition product of rigid-body motions or displacements. The product of two X-subgroups, which is the mathematical model of a serial concatenation of two kinematic chains generating two distinct X-motions, characterizes a noteworthy type of 5-dimensional (5D) displacement set called double Schoenflies motion or X–X motion for brevity. This X–X motion set is a 5D submanifold of the displacement 6D Lie group. Such a motion type includes any spatial translation (3T) and any two sequential rotations (2R) provided that the axes of rotation are parallel to two fixed independent vectors. This motion set also contains the rotations that are products of the foregoing two rotations. In the paper, some preliminary fundamentals on the 4D X-motion are recalled; the 5D set of X–X motions is emphasized. Then implementing serial arrays of one-dof Reuleaux pairs and hinged parallelograms, we enumerate all serial mechanical generators of X–X motion, which have no redundant internal mobility. Based on the group-theoretic concepts, one can differentiate two families of irreducible representations of an X–X motion. One family is realized by twenty-one open chains including the doubly planar motion generators as special cases. The other is generally classified into eight major categories in which one hundred and six distinct open chains generating X–X motion are revealed and nineteen more ones having at least one parallelogram are derived from them. Meanwhile, these kinematic chains are graphically displayed for a possible use in the structural synthesis of parallel manipulators.

In bottom-up proteomics, rapid and efficient protein digestion is crucial for data reliability. However, sample preparation remains one of the rate-limiting steps in proteomics workflows. In this study, we compared the conventional trypsin digestion procedure with two accelerated digestion protocols based on shorter reaction times and microwave-assisted digestion for the preparation of membrane-enriched protein fractions of the human pathogenic bacterium Staphylococcus aureus. Produced peptides were analyzed by Shotgun IPG-IEF, a methodology relying on separation of peptides by IPG-IEF before the conventional LC-MS/MS steps of shotgun proteomics. Data obtained on two LC-MS/MS platforms showed that accelerated digestion protocols, especially the one relying on microwave irradiation, enhanced the cleavage specificity of trypsin and thus improved the digestion efficiency especially for hydrophobic and membrane proteins. The combination of high-throughput proteomics with accelerated and efficient sample preparation should enhance the practicability of proteomics by reducing the time from sample collection to obtaining the results.

In this paper, we focus on some examples of practical applications of the primitive Schonflies-motion generators that are named also the primitive X-motion generators for conciseness. We address especially the serial SCARA-type robots and the 3-dof translational parallel manipulators. Based on the limb architectures of primitive X-motion generators, we systematically introduce general architectures of serial or parallel manipulators for the purpose of practical applications. The brief account of various achievements of some of the X-motion generators that are already described in the recent literature is also proposed and their obviously singular postures are preliminarily obtained.

A noteworthy type of motion called Schoenflies motion and often termed X-motion for brevity is presented. A specified set of X-motions is endowed with the algebraic structure of a four-dimensional (4D) Lie group. This 4D displacement Lie subgroup includes any translation and any rotation provided that the axis of rotation is parallel to a given direction. In the paper, some preliminary fundamentals about the Lie group of displacements are recalled; the 4D Lie subgroup of X-motion is emphasized. Then serial concatenations of one-dof Reuleaux pairs and hinged parallelograms lead to the enumeration of all possible general architectures of mechanical generators for a given X subgroup. Meanwhile, their corresponding embodiments are graphically displayed for a future use in the structural synthesis of parallel manipulators. These generators are sorted into four classes based on the number of prismatic pairs. In total, forty-three distinct mechanical generators of X-motion are revealed and eighty-two ones having at least one parallelogram are also derived from them. Some chains that are defective generators of X-motion are also identified through an approach based on the group dependency.

International audience; This paper presents an incremental learning algorithm for feed-forward neural networks used as approximators of real world data. This algorithm allows neural networks of limited size to be obtained, providing better performances. The algorithm is compared to two of the main incremental algorithms (Dunkin and cascade correlation) in the respective contexts of synthetic data and of real data consisting of radiation doses in homogeneous environments.

We develop here a new class of stochastic models of gene evolution in which the mutations are chaotic, i.e. a random subset of the 64 possible trinucleotides mutates at each evolutionary time t according to some substitution probabilities. Therefore, at each time t, the numbers and the types of mutable trinucleotides are unknown. Thus, the mutation matrix changes at each time t. The chaotic model developed generalizes the standard model in which all the trinucleotides mutate at each time t. It determines the occurrence probabilities at time t of trinucleotides which chaotically mutate according to three substitution parameters associated with the three trinucleotide sites. Two theorems prove that this chaotic model has a probability vector at each time t and that it converges to a uniform probability vector identical to that of the standard model. Furthermore, four applications of this chaotic model (with a uniform random strategy for the 64 trinucleotides and with a particular strategy for the three stop codons) allow an evolutionary study of the three circular codes identified in both eukaryotic and prokaryotic genes. A circular code is a particular set of trinucleotides whose main property is the retrieval of the frames in genes locally, i.e. anywhere in genes and particularly without start codons, and automatically with a window of a few nucleotides. After a certain evolutionary time and with particular values for the three substitution parameters, the chaotic models retrieve the main statistical properties of the three circular codes observed in genes. These applications also allow an evolutionary comparison between the standard and chaotic models.

Direct dose calculations are a crucial requirement for Treatment Planning Systems. Some methods, such as Monte Carlo, explicitly model particle transport, others depend upon tabulated data or analytic formulae. However, their computation time is too lengthy for clinical use, or accuracy is insufficient, especially for recent techniques such as Intensity-Modulated Radiotherapy. Based on artificial neural networks (ANNs), a new solution is proposed and this work extends the properties of such an algorithm and is called NeuRad. Prior to any calculations, a first phase known as the learning process is necessary. Monte Carlo dose distributions in homogeneous media are used, and the ANN is then acquired. According to the training base, it can be used as a dose engine for either heterogeneous media or for an unknown material. In this report, two networks were created in order to compute dose distribution within a homogeneous phantom made of an unknown material and within an inhomogeneous phantom made of water and TA6V4 (titanium alloy corresponding to hip prosthesis). All NeuRad results were compared to Monte Carlo distributions. The latter required about 7 h on a dedicated cluster (10 nodes). NeuRad learning requires between 8 and 18 h (depending upon the size of the training base) on a single low-end computer. However, the results of dose computation with the ANN are available in less than 2 s, again using a low-end computer, for a 150 × 1 × 150 voxels phantom. In the case of homogeneous medium, the mean deviation in the high dose region was less than 1.7%. With a TA6V4 hip prosthesis bathed in water, the mean deviation in the high dose region was less than 4.1%. Further improvements in NeuRad will have to include full 3D calculations, inhomogeneity management and input definitions.

The only power source of nodes in sensor networks is usually provided by batteries which have a determined lifetime. In order to increase the total network lifetime, neighbors nodes can alternate between active (receive, transmit...) and sleep mode. The objective is to find sets of active nodes that ensure as much as possible the zone coverage. In this paper, we will introduce a new method for nodes scheduling in sensor networks. We exploit a mobile beacon localization approach to construct sets of active nodes. The mobile beacon follows the Hilbert space filling curve, that has several properties. It divides the area into unit squares by the mean of linear ordering. In our approach, we chose the Hilbert curve order, in a way that, every unit square will be covered by one and only one sensor node and all the reminder nodes inside this unit square will go to sleep until the first node dies. It is an energy saved approach because nodes communicate only with the mobile beacon. The experimental studies we conducted, confirmed the effectiveness of our approach

Osmolytes are small neutral solutes that provide osmotic pressure to the cell without interacting with or denaturing enzymes and cofactors. An increase in the concentration of osmolytes in mammalian renal cells during hyperosmotic stress has been demonstrated to assist in cell volume regulation. More recently uptake of taurine, a common osmolyte, has been reported in tilapia and the carp epithelial cell line, epithelioma papulosum cyprini (EPC), exposed to hyperosmotic stress. The current report investigated whether taurine transporters were present in anadromous salmon and might assist in adaptation to the marine environment. Cloning of the cDNA coding for the taurine transporter in salmon revealed a high degree of identity with the deduced amino acid sequences of taurine transporters from tilapia (93%), carp (90%) and canine (79%). Exposure of isolated tissues to hyperosmotic conditions revealed marked upregulation (2–5 fold) of transcript for the taurine transporter in branchial lamellae, kidney and heart tissue with maximum expression at 125 mM supplemental NaCl and from 24–72 h. Juvenile salmon transferred from freshwater to seawater exhibited a similar elevation in accumulation of taurine transporter mRNA which continued at least 72 h. Supplementation of hyperosmotic tissue culture medium with 5–30 μM taurine failed to reduce cell mortality or alleviate suppression of protein synthesis in branchial lamellae. Oral administration of taurine in the diet of juvenile salmon prior to and during seawater transfer reduced the quantity of taurine transporter mRNA in smolts but not parr. The results indicate that salmon possess a transporter for the osmolyte taurine and suggest it to be one component of an array of mechanisms involved in adaptation to the dehydrating marine environment.

This paper presents a dynamical study of an imaginary Bragg grating. It shows that when the lasing threshold is reached, the standard transmission matrix approach is no longer valid and a more comprehensive model taking in account gain saturation is required. The grating has then been studied using a direct numerical solution of the time-dependant coupled-mode equations including gain saturation effects. This description is valid for analyzing imaginary Bragg gratings operating both below and above the lasing threshold, and it has demonstrated to provide a physically consistent solution, avoiding the group delay divergences and abnormalities which are predicted by the transfer-matrix model.

International audience; Numerical simulation is a common approach to understand many phenomena, usually yielding a computationally intensive problem. To overcome insufficient computer capacity and computational speed, a grid computing environment is a suitable approach. In this paper we focus on the development of parallel algorithms to solve a 3D transport model in such a context. The solver is based on the multisplitting Newton method that provides a coarse-grained scheme. Algorithms are implemented using JACE, a grid-enabled Java Asynchronous Computing Environment. This programming environment allows users to design synchronous and asynchronous parallel iterative algorithms as well. Experiments are carried out on a heterogeneous grid environment in which the behaviour of both parallel iterative algorithms is analysed. The results allow us to draw some conclusions about the use of the programming library JACE and the design of parallel iterative algorithms in a grid computing environment.

The paper introduces four families of three-DOFs translational-rotational Parallel-Kinematics Mechanisms (PKMs) as well as the mobility analysis of such families using Lie group theory. Two of these families are mechanisms with one-rotational two-translational degrees of freedom (DOFs) and each of the other two has one-translational two-rotational DOFs. Four novel mechanisms are presented and discussed as representatives of these four families. Although these mechanisms are asymmetric, the components used to realise them are very similar and, hence, there is no great departure from the favourable modularity of parallel-kinematics mechanisms.

Based on the Lie-group-algebraic properties of the displacement set, the 3-dof generators of planar gliding motion (G) are introduced. Two serial generators of 3-dof planar gliding produce a 5-dof motion including 3 translations and 2 rotations (3T-2R) as it is established using the composition product of two Lie subgroups of G-motion. All possible 3T-2R 5-dof limbs generating double-planar (G-G) kinematic bonds are comprehensively enumerated. There are a total of 21 mechanical generators of G-G motion with distinct architectures. The parallel setting of three 5-dof G-G limb make up a translational parallel manipulator (TPM), which produces a motion subset of spatial translations. Moreover, related families of TPMs are also deducted using the commutation of factors in any products of spatial translations and rotations around an axis. By that way, some novel examples of 3-dof TPM including doubly planar kinematic chains are synthesized systematically.

The subject of this paper is to show the high efficiency of asynchronism for parallel iterative algorithms in the context of grid computing, that is to say, with machines scattered on a broad geographical scale. The question is: does asynchronism help to reduce the communication penalty and the overall computation time of such a given algorithm? The asynchronous programming model is evaluated on two test problems representing two important classes of scientific applications: a stationary linear problem and a non-stationary non-linear problem. They are implemented with a multi-threaded environment and tested on a set of distant heterogeneous machines. Several experiments have been performed allowing us to compare the performances of such asynchronous algorithms and also to analyze their behavior and extract the main possible optimizations for their use in a grid computing context.

The paper deals with the structural synthesis of asymmetrical non-overconstrained 3-dof spherical parallel mechanisms (SPM). Most of SPM described in the literature are composed of three identical limbs, and they can be called symmetrical SPM. New SPM composed of three structurally distinct limbs, which we name asymmetrical SPM, are disclosed. This paper addresses the type synthesis of asymmetrical non-overconstrained SPM. Firstly, a mobility analysis is presented in order to find the degree of freedom of the limbs. In a second step the geometrical conditions of the limb assembly that achieves the spherical motion are established. Finally, limb architectures of asymmetrical non-overconstrained SPM are enumerated. The resulting kind of scientific catalogue is comprehensive under the hypothesis of non-paradoxical mobility. Moreover SPM 4-dof limbs without idle pairs were never revealed before.