International Professional University of Technology in Nagoya Repository
Not a member yet
15131 research outputs found
Sort by
Robust MPC for actuator-fault tolerance using set-based passive fault detection and active fault isolation
No full textInternational audienc
H{\inf} observer-based decentralised fuzzy control design for nonlinear interconnected systems: an application to vehicle dynamics
No full textInternational audienceIn this study, a decentralised output learning control strategy for a class of nonlinear interconnected systems is studied. Based on Takagi-Sugeno fuzzy (TS) model to approximate the considered interconnected nonlinear systems, a decentralised observer-based control scheme is designed to override the external disturbances such that theH∞ performance is achieved. The appealing attributes of this approach include: (1) the closed-loop system exhibits a robustness against nonlinear interconnections and external disturbance, (2) by one-step procedure, the gainmatrices of observer and controller are obtained on a single step. In simulation results, the controller design is evaluated on the steering stability of a carwhere the nonlinear model describes the side slip, roll and yaw motions of the automotive vehicle equipped with four-wheel-steering and active suspension
Is a point-wise dissipation rate enough to show ISS for time-delay systems?
No full textInternational audienceIn this paper, we address the question whether input-to-state stability (ISS) of nonlinear time-delay systems is guaranteed when a Lyapunov-Krasovskii func- tional (LKF) satisfies a dissipation inequality in which the dissipation rate in- volves solely the present value of the state. We do not yet confirm or infirm this conjecture, but rather identify growth restrictions, on the LKF or on the vector field ruling the dynamics, under which it holds true. An example taken from the neuroscience literature illustrates our findings. We also list a series of robustness properties that naturally hold under this point-wise dissipation inequality, and indicate possible research directions to confirm our conjecture
Sketching for nearfield acoustic imaging of heavy-tailed sources
No full textInternational audienceWe propose a probabilistic model for acoustic source localization with known but arbitrary geometry of the microphone array. The approach has several features. First, it relies on a simple nearfield acoustic model for wave propagation. Second, it does not require the number of active sources. On the contrary, it produces a heat map representing the energy of a large set of candidate locations, thus imaging the acoustic field. Second, it relies on a heavy-tail alpha-stable probabilistic model, whose most important feature is to yield an estimation strategy where the multichannel signals need to be processed only once in a simple on- line procedure, called sketching. This sketching produces a fixed-sized representation of the data that is then analyzed for localization. The resulting algorithm has a small computational complexity and in this paper, we demonstrate that it compares favorably with state of the art for localization in realistic simulations of reverberant environments
Modelling Semantic Context of OOV Words in Large Vocabulary Continuous Speech Recognition
No full textInternational audienceThe diachronic nature of broadcast news data leads to the problem of Out-Of-Vocabulary (OOV) words in Large Vocabulary Continuous Speech Recognition (LVCSR) systems. Analysis of OOV words reveals that a majority of them are Proper Names (PNs). However PNs are important for automatic indexing of audio-video content and for obtaining reliable automatic transcriptions. In this paper, we focus on the problem of OOV PNs in diachronic audio documents. To enable recovery of the PNs missed by the LVCSR system, relevant OOV PNs are retrieved by exploiting the semantic context of the LVCSR transcriptions. For retrieval of OOV PNs, we explore topic and semantic context derived from Latent Dirichlet Allocation (LDA) topic models, continuous word vector representations and the Neural Bag-of-Words (NBOW) model which is capable of learning task specific word and context representations. We propose a Neural Bag-of-Weighted Words (NBOW2) model which learns to assign higher weights to words that are important for retrieval of an OOV PN. With experiments on French broadcast news videos we show that the NBOW and NBOW2 models outperform the methods based on raw embeddings from LDA and Skip-gram models. Combining the NBOW and NBOW2 models gives a faster convergence during training. Second pass speech recognition experiments, in which the LVCSR vocabulary and language model are updated with the retrieved OOV PNs, demonstrate the effectiveness of the proposed context models
Preserving architectural pattern composition information through explicit merging operators
No full textInternational audienceComposable software systems have been proved to support the adaptation to new requirements thanks to their flexibility. A typical method of composable software development is to select and combine a number of patterns that ad-dress the expected quality requirements. Therefore, pattern composition has become a crucial aspect during software design. One of the shortcomings of existing work about pattern composition is the vaporization of composition in-formation which leads to the problem of traceability and reconstructability of patterns. In this paper we propose to give first-class status to pattern merg-ing operators to facilitate the preservation of composition information. The approach is tool-supported and an empirical study has also been conducted to highlight its effectiveness. By applying the approach on the composition of a set of formalized architectural patterns, including their variants, we have shown that composed patterns have become traceable and reconstructable
Some singular sample path properties of a multiparameter fractional Brownian motion
No full textInternational audienceWe prove a Chung-type law of the iterated logarithm for a multiparameter extension of the frac-tional Brownian motion which is not increment stationary. This multiparameter fractional Brow-nian motion behaves very differently at the origin and away from the axes, which also appears in the Hausdorff dimension of its range and in the measure of its pointwise Hölder exponents. A functional version of this Chung-type law is also provided
Algorithm design and analysis in wireless networks
No full textAlgorithms are perhaps the most fundamental and fascinating elements in computer science asa whole. Networks and networked systems are no exception. This habilitation thesis summarizesmy research during the last eight years on some algorithmic problems of both fundamental andpractical importance in modern networks and networked systems, more specifically, wireless networks.Generically, wireless networks have a number of common features which form a commonground on which algorithms for wireless networks are designed. These features include the lackof network-wide coordination, large number of nodes, limited energy and computation resource,and the unreliable wireless links. These constraints and considerations make the algorithmic studyfor wireless networks an emerging research field requiring new tools and methodologies, some ofwhich cannot be drawn from existing state-of-the-art research in either algorithm or networkingcommunity.Motivated by this observation, we aim at making a tiny while systematic step forwards in thedesign and analysis of algorithms that can scale elegantly, act efficiently in terms of computationand communication, while keeping operations as local and distributed as possible. Specifically, weexpose our works on a number of algorithmic problems in emerging wireless networks that aresimple to state and intuitively understandable, while of both fundamental and practical importance,and require non-trivial efforts to solve. These problems include (1) channel rendezvous and neighbordiscovery, (2) opportunistic channel access, (3) distributed learning, (4) path optimizationand scheduling, (5) algorithm design and analysis in radio-frequency identification systems.Methodologically, most of our analysis is systematically articulated as follows.- Theoretical performance bound. After formulating the target problem, we analytically characterizethe performance of the optimal solution as well as some natural and intuitive algorithmsin some cases. These results usually give us pertinent insights on the structuralproperties of the problem including the theoretical limit and the performance gap betweenthe limit and any algorithm that is not carefully devised.- Optimum or approximation algorithm design. Guided by the theoretical results establishedin the first step, we then direct our efforts to the design and analysis of efficientalgorithms for the target problem. By efficient we mean that our algorithms produce eitherthe optimum solution, or, in case where the problem is NP-hard, constant-factor or logarithmicapproximations in polynomial or quasi-polynomial time.- Further extension and generalization. Once we have established a complete frameworksolving or approximately solving the problem, we further analyze the lessons that can belearnt from the analysis process and demonstrate how our framework can be extended oradapted to address a generic class of problems in a wider range of applications presentingsimilar structural properties
Dynamique de la combustion dans un foyer annulaire multi-injecteurs diphasique
No full textThese last decades have seen many innovations in the field of combustion to reduce fuel consumption and pollutant emissions. New types of injector, for example LPP - Lean Premixed Prevaporized, have then been developed to reduce the fuel / air ratio and aim to pre-vaporize the fuel upstream of the combustion in order to mix it better with the air coming from the compressor. Unfortunately this architecture makes annular chambers more sensitive to unsteady phenomena which disturb the functioning of the system, increase the heat flows towards the walls of the chamber, induce vibrations of structures, cause cyclic fatigue of mechanical parts and in extreme cases lead to irreversible damage. The objective of this thesis is to continue the effort undertaken at the EM2C laboratory on this topic and more particularly on the dynamics of combustion in annular chambers comprising a set of injectors. The thesis concerns more particularly the case where the injection of the fuel takes place in liquid form. This configuration reproduces, in idealized form, what can be found in practice in aeronautical engines. It is also a configuration studied at the fundamental level. The chamber, known as MICCA-Spray, is equipped with 16 swirled injectors that can be powered by liquid or gaseous fuel, thus enabling two-phase or fully premixed combustion. The system has quartz walls giving optical access to the flame zone. It is also equipped with a set of diagnostics such as microphones, photomultipliers and high-speed imaging systems.Ces dernières décennies ont vu apparaître de nombreuses innovations dans le domaine de la combustion afin de réduire la consommation et les émissions polluantes. De nouveaux types d'injecteur, de type LPP - Lean Premixed Prevaporized, ont été mis au point permettant de diminuer le rapport combustible/air et visent à pré-vaporiser le carburant en amont de la combustion afin de mieux le mélanger à l'air issu du compresseur. Cette architecture permet une amélioration de la consommation et des émissions polluantes, mais rend les foyers annulaires plus sensibles à des phénomènes instationnaires qui perturbent le fonctionnement du système, accroissent les flux de chaleur vers les parois de la chambre, induisent des vibrations de structures, entrainent une fatigue cyclique des pièces mécaniques et dans des cas extrêmes conduisent à des dommages irréversibles. L'objectif est de poursuivre l'effort engagé au laboratoire EM2C sur ce thème et plus particulièrement sur celui de la dynamique de la combustion dans les chambres annulaires. La thèse concerne plus spécialement le cas où l'injection du combustible s'effectue sous forme liquide. La configuration reproduit sous forme idéalisée celle que l'on trouve en pratique dans les moteurs aéronautiques. La chambre, désignée sous le nom de MICCA-Spray, est équipée de 16 injecteurs swirlés pouvant être alimentés par un combustible liquide ou gazeux, permettant ainsi une combustion diphasique ou prémélangée. Le système possède des parois en quartz donnant un accès optique à la zone de flamme. Il est aussi équipé d'un ensemble de diagnostics tels des microphones, des photomultiplicateurs ainsi que des systèmes d'imagerie à haute cadence
A NEW PERCEPTUAL ASSESSMENT METHODOLOGY FOR SELECTIVE HEVC VIDEO ENCRYPTION
No full textInternational audienceVideo data security is one of the most research topic in the recent years. It is widely used in the multimedia applications such as video-conferencing, Video on Demand and Pay-TV services. Although many video encryption methods and objective measurements have been employed, few real time schemes and no subjective studies have been proposed. In this paper we investigate a set of selective video encryption schemes by encrypting only a few parameters in HEVC video streams. Firstly, we carried out an in-depth subjective study of three proposed selective HEVC video encryption schemes. A panel of ob servers has participated in this test campaign in order to evaluate the degree of visibility of the encrypted videos at different bitrates. Experimental results are presented and analysed, showing therefore that two proposed selected encryption schemes allow a high perceptual security level by masking the whole details of the video content, while the third scheme achieves a high security level, with a content nearly unidentifiable. In addition, subjective scores can be used as ground truth for assessing selective video encryption methods, instead of classical objective signal-based metrics, which are not correlated with human judgment