Dissertations / Theses on the topic 'Signaux multivariés'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 31 dissertations / theses for your research on the topic 'Signaux multivariés.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
Barthélemy, Quentin. "Représentations parcimonieuses pour les signaux multivariés." Phd thesis, Université de Grenoble, 2013. http://tel.archives-ouvertes.fr/tel-00853362.
Full textBarthelemy, Quentin. "Représentations parcimonieuses pour les signaux multivariés." Thesis, Grenoble, 2013. http://www.theses.fr/2013GRENU008/document.
Full textIn this thesis, we study approximation and learning methods which provide sparse representations. These methods allow to analyze very redundant data-bases thanks to learned atoms dictionaries. Being adapted to studied data, they are more efficient in representation quality than classical dictionaries with atoms defined analytically. We consider more particularly multivariate signals coming from the simultaneous acquisition of several quantities, as EEG signals or 2D and 3D motion signals. We extend sparse representation methods to the multivariate model, to take into account interactions between the different components acquired simultaneously. This model is more flexible that the common multichannel one which imposes a hypothesis of rank 1. We study models of invariant representations: invariance to temporal shift, invariance to rotation, etc. Adding supplementary degrees of freedom, each kernel is potentially replicated in an atoms family, translated at all samples, rotated at all orientations, etc. So, a dictionary of invariant kernels generates a very redundant atoms dictionary, thus ideal to represent the redundant studied data. All these invariances require methods adapted to these models. Temporal shift-invariance is an essential property for the study of temporal signals having a natural temporal variability. In the 2D and 3D rotation invariant case, we observe the efficiency of the non-oriented approach over the oriented one, even when data are not revolved. Indeed, the non-oriented model allows to detect data invariants and assures the robustness to rotation when data are revolved. We also observe the reproducibility of the sparse decompositions on a learned dictionary. This generative property is due to the fact that dictionary learning is a generalization of K-means. Moreover, our representations have many invariances that is ideal to make classification. We thus study how to perform a classification adapted to the shift-invariant model, using shift-consistent pooling functions
Aminghafari, Mina. "Méthodes d'ondelettes en statistique des signaux temporels uni et multivariés." Paris 11, 2006. http://www.theses.fr/2006PA112045.
Full textThis thesis takes place in statistics and deals with the applications of wavelets to the univariate and multivariate signals. The first part is devoted to a multivariate extension of the well known wavelet denoising procedure widely examined for scalar valued signals. It combines a straightforward multivariate generalization of a classical one and principal component analysis. This new procedure exhibits promising behavior on classical bench signals and the associated estimator is found to be near minimax in the one-dimensional sense, for Besov balls. The method is finally illustrated by an application to multichannel neural recordings. The second part is devoted to the forecasting problem of a stationary or non-stationary one dimensional time series, using non-decimated wavelet transform. A new proposal method to prediction stationary data and stationary data contaminated by additive trend is proposed. It consists of generalizing a procedure whose idea is to select the wavelet coefficients built from the past observations then to directly estimate the forecasting equation by the regression of the process on the past wavelet coefficients. This scheme is extended to an arbitrary orthogonal wavelet and to the introduction a non-stationary component. The third part relates to a topic a little bit different from the others. We introduce a method for prior selection. This method can be considered as an alternative approach to the parametric empirical Bayes method for priorselection and can then be applied to the choice of threshold in the denoising procedure using wavelets
Fedotenkova, Mariia. "Extraction de composants multivariés des signaux cérébraux obtenus pendant l'anesthésie générale." Thesis, Université de Lorraine, 2016. http://www.theses.fr/2016LORR0189/document.
Full textNowadays, surgical operations are impossible to imagine without general anesthesia, which involves loss of consciousness, immobility, amnesia and analgesia. Understanding mechanisms underlying each of these effects guarantees well-controlled medical treatment. This thesis focuses on analgesia effect of general anesthesia, more specifically, on patients reaction to nociceptive stimuli. We also study differences in the reaction between different anesthetic drugs. The study was conducted on dataset consisting of 230 EEG signals: pre- and post-incision recordings obtained form 115 patients, who received desflurane and propofol. The first stage of the study comprise power spectral analysis, which is a widespread approach in signal processing. Spectral information was described by fitting the background activity, that exposes $1/f$ behavior, to power spectral density estimates of the EEG signals and measuring power contained in delta and alpha bands relatively to the power of background activity. A further improvement was done by expanding spectra with time information due to observed non-stationary nature of EEG signals. To obtain time-frequency representations of the signals we apply three different methods: scalogram (based on continuous wavelet transform), conventional spectrogram, and spectrogram reassignment. The latter allows to ameliorate readability of a time-frequency representation by reassigning energy contained in spectrogram to more precise positions. Subsequently, obtained spectrograms were used as phase space reconstruction in recurrence analysis and its quantification by complexity measure. Recurrence analysis allows to describe and visualize recurrent dynamics of a system and discover structural patterns contained in the data. Here, recurrence plots were used as rewriting grammar to turn an original signal into a symbolic sequence, where each symbol represents a certain state of the system. After computing three different complexity measures of resulting symbolic sequences they are used as features for classification. Finally, combining features obtained with power spectral analysis and recurrence symbolic analysis, we perform classification of the data using two classification methods: linear discriminant analysis and support vector machines. Classification was carried out on two-class problem, distinguishing between pre-/post-incision EEG signals, as well as between two different anesthetic drugs, desflurane and propofol
Lung-Yut-Fong, Alexandre. "Détection de ruptures pour les signaux multidimensionnels. Application à la détection d'anomalies dans les réseaux." Phd thesis, Télécom ParisTech, 2011. http://pastel.archives-ouvertes.fr/pastel-00675543.
Full textFrusque, Gaëtan. "Inférence et décomposition modale de réseaux dynamiques en neurosciences." Thesis, Lyon, 2020. http://www.theses.fr/2020LYSEN080.
Full textDynamic graphs make it possible to understand the evolution of complex systems evolving over time. This type of graph has recently received considerable attention. However, there is no consensus on how to infer and study these graphs. In this thesis, we propose specific methods for dynamical graph analysis. A dynamical graph can be seen as a succession of complete graphs sharing the same nodes, but with the weights associated with each link changing over time. The proposed methods can have applications in neuroscience or in the study of social networks such as Twitter and Facebook for example. The issue of this thesis is epilepsy, one of the most common neurological diseases in the world affecting around 1% of the population.The first part concerns the inference of dynamical graph from neurophysiological signals. To assess the similarity between each pairs of signals, in order to make the graph, we use measures of functional connectivity. The comparison of these measurements is therefore of great interest to understand the characteristics of the resulting graphs. We then compare functional connectivity measurements involving the instantaneous phase and amplitude of the signals. We are particularly interested in a measure called Phase-Locking-Value (PLV) which quantifies the phase synchrony between two signals. We then propose, in order to infer robust and interpretable dynamic graphs, two new indexes that are conditioned and regularized PLV. The second part concerns tools for dynamical graphs decompositions. The objective is to propose a semi-automatic method in order to characterize the most important patterns in the pathological network from several seizures of the same patient. First, we consider seizures that have similar durations and temporal evolutions. In this case the data can be conveniently represented as a tensor. A specific tensor decomposition is then applied. Secondly, we consider seizures that have heterogeneous durations. Several strategies are proposed and compared. These are methods which, in addition to extracting the characteristic subgraphs common to all the seizures, make it possible to observe their temporal activation profiles specific to each seizures. Finally, the selected method is used for a clinical application. The obtained decompositions are compared to the visual interpretation of the clinician. As a whole, we found that activated subgraphs corresponded to brain regions involved during the course of the seizures and their time course were highly consistent with classical visual interpretation
Michaud, François-Thomas. "Profilage protéomique par analyse multivariée de signaux LCMS appliqué en ingénierie cellulaire." Thesis, Université Laval, 2009. http://www.theses.ulaval.ca/2009/26366/26366.pdf.
Full textDumont, Jerome. "Fouille de dynamiques multivariées, application à des données temporelles en cardiologie." Phd thesis, Université Rennes 1, 2008. http://tel.archives-ouvertes.fr/tel-00364720.
Full textDumont, Jérôme. "Fouille de dynamiques multivariées : application à des données temporelles en cardiologie." Rennes 1, 2008. http://www.theses.fr/2008REN1S078.
Full textThis manuscript focuses on the problem of analysing dynamics of time series observed in cardiology. The proposed solution is divided into two steps. The first one consists in the extraction of useful information from the ECG by segmenting each beat with a wavelet decomposition algorithmn, adapted from the litterature. The difficult problem of optimising both thresholds and time windows is solved with evolutionary algorithms. The second step relies on Hidden Semi-Markovian models to represent the time series made up of the extracted variables. An algorithm of unsupervised classification is proposed to retrieve the natural groups. The application of this method to the detection of ischemic episodes and to the analysis of stress ECG from patients suffering from Brugada syndrome presents a higher performance than more tradionnal approaches
Altuve, Miguel. "Détection multivariée des épisodes d'apnée-bradycardie chez le prématuré par modèles semi-markovien cachés." Rennes 1, 2011. http://www.theses.fr/2011REN1S053.
Full textThis dissertation studies the early detection of apnea-bradycardia (AB) events in preterm infants. After defining the importance of AB detection from a clinical point of view, a methodological approach is proposed. It relies on a data mining process that includes data cleansing and feature extraction. In chapter 3, a novel method based on evolutionary algorithms, for optimizing the thresholds and the analysis windows, is proposed to adapt the algorithms of the ECG signal to the specific characteristics of preterm infants, very different from the EGC of adult. In chapter 4, a semi-Markovian approach is adapted for modeling of dynamics and several improvements are proposed : heterogeneous models, adaptation to online processing, optimization of experiments, are reported on simulated and read signals. They clearly highlight the importance of considering the dynamic of the signals. They also emphasize that with a suitable pre-treatment such as the quantification of observations and the introduction of delay between the observable, a significant gain in performance can be observed
Lefevre, Francis. "Classification automatique de signaux d'émission acoustique en relation avec les phénomènes qui leur ont donné naissance." Compiègne, 1985. http://www.theses.fr/1985COMPI182.
Full textRoyer, Jean-Jacques. "Analyse multivariable et filtrage des données régionalisées." Vandoeuvre-les-Nancy, INPL, 1988. http://www.theses.fr/1988NAN10312.
Full textLeni, Pierre-Emmanuel. "Nouvelles méthodes de traitement de signaux multidimensionnels par décomposition suivant le théorème de Superposition de Kolmogorov." Phd thesis, Université de Bourgogne, 2010. http://tel.archives-ouvertes.fr/tel-00581756.
Full textMielcarek, Didier. "Étude et développement de méthodes d'identification multi-variables : application à un procédé chimique." Vandoeuvre-les-Nancy, INPL, 1990. http://docnum.univ-lorraine.fr/public/INPL_T_1990_MIELCAREK_D.pdf.
Full textAhmadou, Mohamed Diaa. "Contribution au développement d'un dispositif robuste de détection-diffusion d'huiles essentielles à concentration contrôlée." Thesis, Université de Lorraine, 2015. http://www.theses.fr/2015LORR0187/document.
Full textControlThis work contributes to the design of a gas diffusion-sensing system controlling in real time the essential oil concentration in a confined atmosphere. The objective is to create reproducible exposure conditions of olfactory stimuli on living beings to test their neurosensory impacts. The main constraint is to measure with good accuracy and rapidity the odor concentration of a global atmosphere. We decided to use a gaseous detection device (electronic nose) based on commercial resistive metal oxide sensors coupled to a prior learning at fixed concentrations of pine essential oil. Experimental equipment was first developed in order to study, characterize and especially optimize the device performances to be achieved. Initially, the study of time gas sensor responses was used to optimize working measurement conditions: cycle of 75s gas exposure phase, followed by 350s pure air regeneration phase. First results allowed the classification of our sensors in terms of rapidity, sensitivity and drift levels. A systematic characterization measurement was made under various concentration variations: increasing, decreasing or random ones taking account of all possible forms of response drifts. To reduce errors due to the drifts, an original pretreatment was initiated by normalizing each sensor response value in relation with its corresponding conductance value at the end of regeneration phase. Two normalized features and also the maximum value of the derivative curve were defined for each time sensor response. The analysis by ACP and AFD classification methods of the database created using these three features show the difficulty in differentiating high concentrations, even by eliminating the two least efficient sensors. So, a completely new approach was proposed by combining the orthogonal signal correction technique (OSC) allowing to remove irrelevant information, and the Partial Linear Square regression method PLS, adapted in case of multi-collinearity and a large number of parameters. Using these two methods yields a much better discrimination of the high concentrations, maintaining the concentration prediction accuracy with a maximum stability of the regression model. Finally, the concentration prediction has been optimized by substituting representative parameters with the full response signal, the calculation time remaining low. A very good assessment of the gas concentration in all the used range was obtained. So we have developed a robust and accurate model for the calibration of our system thanks to a combination of original processing and analysis methods, allowing to achieve a reliable detection-diffusion prototype
Qin, Lei. "Online machine learning methods for visual tracking." Thesis, Troyes, 2014. http://www.theses.fr/2014TROY0017/document.
Full textWe study the challenging problem of tracking an arbitrary object in video sequences with no prior knowledge other than a template annotated in the first frame. To tackle this problem, we build a robust tracking system consisting of the following components. First, for image region representation, we propose some improvements to the region covariance descriptor. Characteristics of a specific object are taken into consideration, before constructing the covariance descriptor. Second, for building the object appearance model, we propose to combine the merits of both generative models and discriminative models by organizing them in a detection cascade. Specifically, generative models are deployed in the early layers for eliminating most easy candidates whereas discriminative models are in the later layers for distinguishing the object from a few similar "distracters". The Partial Least Squares Discriminant Analysis (PLS-DA) is employed for building the discriminative object appearance models. Third, for updating the generative models, we propose a weakly-supervised model updating method, which is based on cluster analysis using the mean-shift gradient density estimation procedure. Fourth, a novel online PLS-DA learning algorithm is developed for incrementally updating the discriminative models. The final tracking system that integrates all these building blocks exhibits good robustness for most challenges in visual tracking. Comparing results conducted in challenging video sequences showed that the proposed tracking system performs favorably with respect to a number of state-of-the-art methods
Moslem, Bassam. "Méthodes non paramétriques pour la classification dans les signaux non stationnaires : application à l'EMG utérin." Compiègne, 2011. http://www.theses.fr/2011COMP1981.
Full textUterine contraction monitoring provides important prognostic information during pregnancy and labor and can be used for an early detection of any sign of preterm labor. Current techniques used for monitoring the uterine contraction impose a compromise between accuracy and invasiveness. Recently, the uterine electrical activity has been proven to be representative of the uterine contractility. The uterine electromyogram (EMG), also called the electrohysterogram (EHG), is the bioelectrical signal associated with the uterine activity. Recorded noninvasively from the abdominal wall of pregnant women, uterine EMG gives valuable information about the function aspects of the uterine contractility. Numerous studies have analyzed the uterine recordings associated with pregnancy and labor: it has been proven that it is of interest to offer a good insight into the process of pregnancy and labor and may be also used to predict the risk of preterm labor. Our study focuses on feature extraction, pregnancy monitoring and signal classification. In the first part, we apply new signal processing techniques (spectral analysis, multiresolution analysis, nonlinear analysis…) in order to extract new features capable of provide the best characterization of the uterine EMG. Next, a pregnancy monitoring using the extracted features in presented. This study concerns different women recorded at several pregnancy terms. This approach is improved by applying the multiresolution analysis based on the wavelet packet transform. We searched for the best basis adapted for the problem of pregnancy monitoring. In order to benefit from the multichannel type of the recorded signals, we study the spatial variability of the electrical activity at different recording sites of the uterus. This multichannel-based approach allows us to know the way the electrical activity changes at throughout pregnancy over all the uterine muscle. In the last part, we present our work on classifying uterine EMG signals between two classes of contraction (pregnancy vs. Labor). A novel approach based on multisensor data fusion is presented. The high correct classification ratio (92%) obtained proves that this method may be the solution for the problem described
Wang, Tian. "Abnormal detection in video streams via one-class learning methods." Thesis, Troyes, 2014. http://www.theses.fr/2014TROY0018/document.
Full textOne of the major research areas in computer vision is visual surveillance. The scientific challenge in this area includes the implementation of automatic systems for obtaining detailed information about the behavior of individuals and groups. Particularly, detection of abnormal individual movements requires sophisticated image analysis. This thesis focuses on the problem of the abnormal events detection, including feature descriptor design characterizing the movement information and one-class kernel-based classification methods. In this thesis, three different image features have been proposed: (i) global optical flow features, (ii) histograms of optical flow orientations (HOFO) descriptor and (iii) covariance matrix (COV) descriptor. Based on these proposed descriptors, one-class support vector machines (SVM) are proposed in order to detect abnormal events. Two online strategies of one-class SVM are proposed: The first strategy is based on support vector description (online SVDD) and the second strategy is based on online least squares one-class support vector machines (online LS-OC-SVM)
Traore, Oumar Issiaka. "Méthodologie de traitement et d'analyse de signaux expérimentaux d'émission acoustique : application au comportement d'un élément combustible en situation accidentelle." Thesis, Aix-Marseille, 2018. http://www.theses.fr/2018AIXM0011/document.
Full textThe objective of the thesis is to contribute to the improvement of the monitoring process of nuclear safety experiments dedicated to study the behavior of the nuclear fuel in a reactivity initiated accident (RIA) context, by using the acoustic emission technique. In particular, we want to identify the physical mechanisms occurring during the experiments through their acoustic signatures. Firstly, analytical derivations and numerical simulations using the spectral finite element method have been performed in order to evaluate the impact of the wave travelpath in the test device on the recorded signals. A resonant frequency has been identified and it has been shown that the geometry and the configuration of the test device may not influence the wave propagation in the low frequency range. Secondly, signal processing methods (spectral subtraction, singular spectrum analysis, wavelets,…) have been explored in order to propose different denoising strategies according to the type of noise observed during the experiments. If we consider only the global SNR improvement ratio, the spectral subtraction method is the most robust to changes in the stochastic behavior of noise. Finally, classical multivariate and functional data analysis tools are used in order to create a machine learning algorithm dedicated to contribute to a better understanding of the phenomenology of RIA accidents. According to the method (multivariate or functional), the obtained algorithms allow to identify the mechanisms in more than 80 % of cases
El, Aabid Moulay Abdelaziz. "Attaques par canaux cachés : expérimentations avancées sur les attaques template." Paris 8, 2011. http://www.theses.fr/2011PA083394.
Full textIn the 90's, the emergence of new cryptanalysis methods revolutionized the security of cryptographic devices. These attacks are based on power consumption analysis, when the microprocessor is running the cryptographic algorithm. Especially, we analyse in this thesis some properties of the \emph{template attack}, and we provide some practical improvements. The analyse consists in a case-study based on side-channel measurements acquired experimentally from a hardwired cryptographic accelerator. The principal component analysis (PCA) is used to represent the \emph{templates} in some dimensions, and we give a physical interpretation of the \emph{templates} eigenvalues and eigenvectors. We introduce a method based on the thresholding of leakage data to accelerate the profiling or the matching stages. In this context, there is an opportunity to study how to best combine many attacks with many leakages from different sources or using different samples from a single source. That brings some concrete answers to the attack combination problem. Also we focus on identifying the problems that arise when there is a discrepancy between the \emph{templates} and the traces to match : the traces can be desynchronized and the amplitudes can be scaled differently. Then we suggest two remedies to cure the \emph{template} mismatches. We show that SCAs when performed with a multi-resolution analysis are much better than considering only the time or the frequency resolution. Actually, the gain in number of traces needed to recover the secret key is relatively considerable with repect to an ordinary attack
Hassan, Mahmoud. "Analysis of the propagation of uterine electrical activity applied to predict preterm labor : prédiction de menaces d'accouchement prématuré." Compiègne, 2011. http://www.theses.fr/2011COMP1948.
Full textUterine contractions are essentially controlled by two physiological phenomena: cell excitability and propagation of uterine electrical activity probably related to high and low frequencies of uterine electromyogram, called electrohysterogram -EHG-, respectively. All previous studies have been focused on extracting parameters from the high frequency part and did not show a satisfied potential for clinical application. The objective of this thesis is the analysis of the propagation EHG signals of during pregnancy and labor in the view of extracting tool for clinical application. A novelty of our thesis is the multichannel recordings by using 4x4 electrodes matrix posed on the woman abdomen. Monovariate analysis was aimed to investigate the nonlinear characteristics of EHG signals. Bivariate and multivariate analyses have been done to analyze the propagation of the EHG signals by detecting the connectivity between the signals. An increase of the nonlinearity associated by amplitude synchronization and phase desynchronization were detected. Results indicate a highest EHG propagation during labor than pregnancy and an increase of this propagation with the week of gestations. The results show the high potential of propagation’s parameters in clinical point of view such as labor detection and then preterm labor prediction. We proposed novel combination of Blind Source Separation and empirical mode decomposition to denoise monopolar EHG as a possible way to increase the classification rate of pregnancy and labor
Lindig, León Cecilia. "Classification multilabels à partir de signaux EEG d'imaginations motrices combinées : application au contrôle 3D d'un bras robotique." Thesis, Université de Lorraine, 2017. http://www.theses.fr/2017LORR0016/document.
Full textBrain-Computer Interfaces (BCIs) replace the natural nervous system outputs by artificial ones that do not require the use of peripheral nerves, allowing people with severe motor impairments to interact, only by using their brain activity, with different types of applications, such as spellers, neuroprostheses, wheelchairs, or among others robotics devices. A very popular technique to record signals for BCI implementation purposes consists of electroencephalography (EEG), since in contrast with other alternatives, it is noninvasive and inexpensive. In addition, due to the potentiality of Motor Imagery (MI, i.e., brain oscillations that are generated when subjects imagine themselves performing a movement without actually accomplishing it) to generate suitable patterns for scheming self-paced paradigms, such combination has become a common solution for BCI neuroprostheses design. However, even though important progress has been made in the last years, full 3D control is an unaccomplished objective. In order to explore new solutions for overcoming the existing limitations, we present a multiclass approach that considers the detection of combined motor imageries, (i.e., two or more body parts used at the same time). The proposed paradigm includes the use of the left hand, right hand, and both feet together, from which eight commands are provided to direct a robotic arm comprising fourteen different movements that afford a full 3D control. To this end, an innovative switching-mode scheme that allows managing different actions by using the same command was designed and implemented on the OpenViBE platform. Furthermore, for feature extraction a novel signal processing scheme has been developed based on the specific location of the activity sources that are related to the considered body parts. This insight allows grouping together within a single class those conditions for which the same limb is engaged, in a manner that the original multiclass task is transformed into an equivalent problem involving a series of binary classification models. Such approach allows using the Common Spatial Pattern (CSP) algorithm; which has been shown to be powerful at discriminating sensorimotor rhythms, but has the drawback of being suitable only to differentiate between two classes. Based on this perspective we also have contributed with a new strategy that combines together the CSP algorithm and Riemannian geometry. In which the CSP projected trials are mapped into the Riemannian manifold, from where more discriminative features can be obtained as the distances separating the input data from the considered class means. These strategies were applied on three new classification approaches that have been compared to classical multiclass methods by using the EEG signals from a group of naive healthy subjects, showing that the proposed alternatives not only outperform the existing schema, but also reduce the complexity of the classification task
Komaty, Ali. "Traitement et analyse des processus stochastiques par EMD et ses extensions." Thesis, Brest, 2014. http://www.theses.fr/2014BRES0107.
Full textThe main contribution of this thesis is aimed towards understanding the behaviour of the empirical modes decomposition (EMD) and its extended versions in stochastic situations
Kherdali, Khalil el. "Etude, conception et réalisation d'un radar ultrasonore." Montpellier 2, 1992. http://www.theses.fr/1992MON20122.
Full textNicollin, Florence. "Traitement de profils sismiques "ECORS" par projection sur le premier vecteur propre de la matrice spectrale." Grenoble INPG, 1989. http://www.theses.fr/1989INPG0101.
Full textBjörk, Anders. "Chemometric and signal processing methods for real time monitoring and modeling : applications in the pulp and paper industry." Doctoral thesis, KTH, Kemi, 2007. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-4383.
Full textVid framställning av pappersprodukter är kvaliteten på massan en viktig faktor för produktiviteten och kvalitén på slutresultatet. Det är därför viktigt att ha tillgång till tillförlitliga mätningar av massakvalitet i realtid. En möjlighet är att använda akustik- eller vibrationssensorer i lämpliga positioner vid enhetsoperationer i massaprocessen. Selektiviteten hos dessa mätningar är emellertid relativt låg i synnerhet om mätningarna är passiva. Därför krävs avancerad signalbehandling och multivariat kalibrering. Det nu presenterade arbetet har varit fokuserat på kalibreringsmetoder för extraktion av information ur akustiska mätningar samt på algoritmer för signalbehandling som kan ge förbättrad informationsselektivitet. Multivariata metoder som Principal Component Analysis (PCA), Partial Least Squares (PLS) and Orthogonal Signal Correction (OSC) har använts för visualisering och kalibrering. Signalbehandlingsmetoderna Fast Fourier Transform (FFT), Fast Wavelet Transform (FWT) och Continuous Wavelet Transform (CWT) har använts i utvecklingen av nydanande metoder för signalbehandling anpassade till att extrahera information ur signaler från vibrations/akustiska sensorer. En kombination av OSC och PLS applicerade på FFT-spektra från raffineringen i en Termo Mechnaical Pulping (TMP) process ger lägre prediktionsfel för Canadian Standard Freeness (CSF) än enbart PLS. Kombinationen av FFT och PLS har vidare använts för monitorering av malning av sulfatmassa och monitorering av silning. Ordinära FFT-spektra av t.ex. vibrationssignaler är delvis överlappande. För att komma runt detta har två signalbehandlingsmetoder utvecklats, Wavelet Transform Multi Resolution Spectra (WT-MRS) baserat på kombinationen av FWT och FFT samt Continuous Wavelet Transform Fibre Length Extraction (CWT-FLE) baserat på CWT. Tillämpning av WT-MRS gav enklare PLS-modeller med lägre prediktionsfel för CSF jämfört med att använda normala FFT-spektra. I en annan tillämpning på en massaström med relativt hög koncentration (Medium Consistency, MC) kunde prediktioner för CSF samt ljushet erhållas med prediktionsfel jämförbart med referensmetodernas fel. Metoden CWT-FLE validerades mot en kommersiell fiberlängdsmätare med god överensstämmelse. CWT-FLE-kurvorna skulle därför kunna användas i stället för andra fiberdistributionskurvor för processtyrning. Vidare användes CWT-FLE kurvor för PLS modellering av dragstyrka samt optiska egenskaper med goda resultat. Utöver de nämnda resultaten har en omfattande litteratursammanställning gjorts över området och relaterade applikationer.
QC 20100629
Faure, Cynthia. "Détection de ruptures et identification des causes ou des symptômes dans le fonctionnement des turboréacteurs durant les vols et les essais." Thesis, Paris 1, 2018. http://www.theses.fr/2018PA01E059/document.
Full textAnalysing multivariate time series created by sensors during a flight or a bench test represents a new challenge for aircraft engineers. Each time series can be decomposed univariately into a series of stabilised phases, well known by the expert, and transient phases that are merely explored but very informative when the engine is running. Our project aims at converting these time series into a succession of labels, designing transient and stabilised phases in a bivariate context. This transformation of the data will allow several perspectives: tracking similar behaviours or bivariate patterns seen during a flight, finding similar curves from a given curve, identifying the atypical curves, detecting frequent or rare sequences of labels during a flight, discovering hidden multivariate structures, modelling a representative flight, and spotting unusual flights. This manuscript proposes : methodology to automatically identify transient and stabilized phases, cluster all engine transient phases, label multivariate time series and analyse them. All algorithms are applied on real flight measurements with a validation of the results from expert knowledge
Sargent, Gabriel. "Estimation de la structure de morceaux de musique par analyse multi-critères et contrainte de régularité." Phd thesis, Université Rennes 1, 2013. http://tel.archives-ouvertes.fr/tel-00853737.
Full textMeriaux, Bruno. "Contributions aux traitements robustes pour les systèmes multi-capteurs." Electronic Thesis or Diss., université Paris-Saclay, 2020. http://www.theses.fr/2020UPASG009.
Full textOne of the objectives of statistical signal processing is the extraction of useful information from a set of data and a statistical model. For example, most of the methods for detecting/localizing targets in radar generally require the estimation of the covariance matrix. With the emergence of high-resolution systems, the use of a Gaussian model is no longer suited and therefore leads to performance degradations. In addition, prior information can be obtained by a prior study of the system, such as the structure of the covariance matrix. Taking them into account then improves the performance of the processing methods. First, we introduce new robust structured estimators of the covariance matrix, based on the family of elliptical distributions and the class of M-estimators. We analyze the asymptotic performances of the latter and we conduct a sensitivity analysis by considering the possibility of mismatches on the statistical model.Secondly, we propose a reformulation of the target detection problem using sparse subspace clustering techniques. We then study some theoretical properties of the optimization problem and we apply this methodology in a scenario of target detection in presence of jammers
Duchêne, Florence. "Fusion de Données Multicapteurs pour un Système de Télésurveillance Médicale de Personnes à Domicile." Phd thesis, Université Joseph Fourier (Grenoble), 2004. http://tel.archives-ouvertes.fr/tel-00007607.
Full textLasmar, Nour-Eddine. "Modélisation stochastique pour l'analyse d'images texturées : Approches Bayésiennes pour la caractérisation dans le domaine des transformées." Phd thesis, Université Sciences et Technologies - Bordeaux I, 2012. http://tel.archives-ouvertes.fr/tel-00809279.
Full text