Dissertations / Theses on the topic 'Identification des systèmes'

A nonlinear identification scheme is provided for a LTI-system with a feedback-nonlinearity, which depends on the input and LTI-system output. This is especially the case for MEMS, where the electrostatic field depends on the displacement and input voltage. The fact, that the algorithm only requires a matrix inversion and singular value decomposition, makes it possible to use the identification scheme for online-estimation. There is also no other a-priori knowledge about the system, except the order, needed.

Direkt viktoptimering (Direct Weight Optimization, DWO) är en ickeparamterisk systemidentifieringsmetod. DWO bygger på att man skattar ett funktionsvärde i en viss punkt genom en viktad summa av mätvärden, där vikterna optimeras fram. Det faktum att DWO har en inparameter som man måste veta i förväg leder till att man på något sätt vill skatta denna inparameter. Det finns många sätt man kan göra denna skattning på men det centrala i denna uppsats är att skatta inparametern lokalt. Fördelen med detta är att metoden anpassar sig om till exempel systemet ändrar beteende från att variera långsamt till att variera snabbare. Denna typ av metoder brukar kallas adaptiva metoder.Det finns flera metoder för att skatta en inparameter lokalt och anpassningen till DWO är redan klar för ett fåtal som lämpar sig bra. Det är dock inte undersökt vilken av dessa metoder som ger det bästa resultatet för just DWO. Syftet med denna uppsats är alltså att ta reda på hur man lokalt kan skatta en inparameter till DWO på bästa sätt och om DWO är en bra grund att basera en adaptiv metod på.Det har visat sig att DWO kanske är för känslig för en lokalt vald inparameter för att vara en bra grund att basera en adaptiv metod på. Däremot utmärker sig en av metoderna för att skatta inparametern genom att vara mycket bättre än de andra metoderna när den kanske inte borde vara det. Varför den är så bra kan vara ett bra ämne för vidare forskning.

Direct Weight Optimization (DWO) is a nonparametric system identification meth\-od. In DWO the value of a function in a certain point is estimated by a weighted sum of measured values. The weights are obtained as a solution to a convex optimization problem. DWO has a design parameter which has to be chosen or estimated a priori. There are many ways to estimate this parameter. The main focus of this thesis is to estimate this parameter locally. The advantage of estimating the parameter locally is that the estimate will adapt if the system changes behavior from slowly varying to rapidly varying. Estimation methods of this type are usually called adaptive estimation methods.There are a number of adaptive estimation methods and the adaptation of some of these methods to DWO has already been done. There are however no evaluation studies done. The goal with this thesis is therefore to find out how to estimate the parameter in DWO in the best way and to find out whether DWO is a good base for an adaptive method.It turned out that DWO might be too sensitive to local changes in the design parameter to be a good base for an adaptive method. However, one of the adaptive estimation methods stands out from the rest because it is much better than the other methods when it, perhaps, should not. Why this method is good might be a good subject for further research.

Thesis (Ph. D.)--West Virginia University, 2004.
Title from document title page. Document formatted into pages; contains v, 132 p. : ill. (some col.). Includes abstract. Includes bibliographical references (p. 128-132).

Magnetically levitated microrobotic systems have shown a great deal of promise for micromanipulation tasks. A new large-gap magnetic suspension system has recently been developed at the University of Waterloo in order to develop microrobotic systems for various applications. In order to achieve motion with the system, a model is needed in order to facilitate the design of various aspects of the system, such as the microrobot and the controller. In order to derive equations of motion for the system attempts were made to characterize the force produced by the magnetic drive unit in terms of a simple analytical equation. The force produced by the magnetic drive unit was estimated with the aid of a finite element model. The derived equations were able to predict the general trend of the force curves, and with sufficient parameter tweaking the error between the force estimated by the finite element model and the force estimated by the analytical equation could be minimized. System models describing the motion of the system in the horizontal and vertical directions are identified and compared to the actual system response. The vertical position response is identified through a least squares parameter estimate of the closed loop response combined with a partial reconstruction of the root locus diagram, with the model structure based on the known dynamics of a simplified form of magnetic levitation. This model was able to provide a reasonable prediction of the system response for a variety of PID controllers under a variety of input conditions. The horizontal models are identified using a least-squares parameter estimate of the open loop characteristics of the system. The horizontal models are able to provide a reasonable prediction of the system response under PD and PID control. Full spatial motion of a microrobot prototype is demonstrated over a working range of 20x22x30 mm³, with PID controller parameters and reference trajectories adjusted to minimize disturbances. The RMS error at steady state is on the order of 0. 020 mm for vertical positioning and 0. 008 mm for horizontal positioning. A linear quadratic regulator implemented for vertical position control was able to reduce the vertical position RMS error to 0. 014 mm.

While the characteristic dimensions of electronic systems scale down to micro- or nano-world, their performance is greatly influenced. Micro-fabrication process or variations of the operating situation such as temperature, humidity or pressure are usual cause of dispersion. Therefore, it seems essential to co-integrate self-testing or self-adjustment routines for these microdevices. For this feature, most existing system parameter estimation methods are based on the implementation of high-resolution digital measurements of the system's output. Thus, long design time and large silicon areas are needed, which increases the cost of the micro-fabricated devices. The parameter estimation problems based on binary outputs can be introduced as alternative self-test identification methods, requiring only a 1-bit Analog-to-Digital Converter (ADC) and a 1-bit Digital-to-Analog Converter (DAC).In this thesis, we propose a novel recursive identification method to the problem of system parameter estimation from binary observations. An online identification algorithm with low-storage requirements and small computational complexity is derived. We prove the asymptotic convergence of this method under some assumptions. We show by Monte Carlo simulations that these assumptions do not necessarily have to be met in practice in order to obtain an appropriate performance of the method. Furthermore, we present the first experimental application of this method dedicated to the self-test of integrated micro-electro-mechanical systems (MEMS). The proposed online Built-In Self-Test method is very amenable to integration for the self-testing of systems relying on resistive sensors and actuators, because it requires low memory storage, only a 1-bit ADC and a 1-bit DAC which can be easily implemented in a small silicon area with minimal energy consumption.

I linfociti T CD4+ orchestrano la risposta immunitaria differenziandosi in vari sottogruppi di cellule T effettrici. Nella prima parte del lavoro, abbiamo studiato il ruolo dei long non-coding RNAs nel processo differenziativo delle cellule T mediante un'analisi completa del trascrittoma con tecnologia RNA-Seq di tredici sottopopolazioni linfocitarie altamente purificate da cellule primarie umane e abbiamo identificato oltre 500 nuovi lincRNAs. Dai nostri risultati si evince che i lincRNAs sono preferenzialmente espressi in specifiche sottopopolazioni linfocitarie e che i loro pattern di espressione cambiano durante il differenziamento delle cellule T. Abbiamo inoltre caratterizzato funzionalmente Linc-MAF-4, un lincRNA 'signature' delle cellule CD4+ Th1, e abbiamo dimostrato che la downregolazione di Linc-MAF-4 aumenta i livelli di espressione del fattore di trascrizione MAF e spinge il processo differenziativo delle cellule CD4+ verso un profilo di espressione TH2. Dopo aver evidenziato la centralitá del ruolo dei lncRNAs in cellule T primarie da donatori sani, cerchiamo ora di caratterizzare il loro coinvolgimento nei processi regolatori dei linfociti infiltranti i tumori (TIL). Una caratterizzazione approfondita dei meccanismi molecolari alla base delle caratteristiche funzionali dei TIL possono condurre ad una comprensione del loro ruolo nei meccanismi di fuga del tumore dalla risposta immunitaria e consentire l'identificazione di nuovi bersagli terapeutici per una efficace modulazione di queste cellule nel cancro. Dal momento che si sa molto poco riguardo all'espressione dei lncRNAs nei TIL, abbiamo isolato cellule CD4+ Th1, Th17 e Tregs infiltranti sia tumore che tessuto sano, e linfociti da organi linfoidi secondari e sangue periferico di pazienti affetti da Non-small cell lung cancer (cancro al polmone) e cancro del colon-retto. Abbiamo analizzato il trascrittoma di queste cellule con tecnologia RNA-Seq e definito una serie di lncRNAs che sono specificamente espressi in linfociti infiltranti il tumore.
CD4+ T lymphocytes orchestrate immune responses by differentiating into various subsets of effector T cells. We addressed the role of regulatory long non-coding RNAs in T-cell differentiation and plasticity performing a comprehensive transcriptome analysis by RNA-seq of thirteen highly purified human primary lymphocyte subsets and identified more than 500 new lincRNAs. We found that lincRNAs are preferentially expressed in specific lymphocyte subsets and that their expression patterns change during T-cell differentiation. Furthermore, we functionally characterized linc-MAF-4, a Th1 CD4+ signature lincRNA, and found that linc-MAF-4 down regulation increases the expression levels of transcription factor MAF and skews CD4+ differentiating cells towards a Th2 like expression profile. After assessing the role of lncRNAs in primary T cells from healthy donors, we seek to characterize their involvement in regulatory processes for tumor infiltrating lymphocytes (TIL). An in-depth characterization and understanding of the molecular mechanisms underlying the functional features of TIL may lead to a comprehension of their role in tumor immune escape and allow the identification of new therapeutic targets for the effective modulation of these cells in cancer. Since very little is known on the expression of lncRNAs in TILs, we isolated CD4+ Th1, Th17 and Tregs cells infiltrating both tumor and healthy tissue as well as lymphocytes from lymphoid Issues and peripheral blood of Non-Small- Cell-Lung and Colorectal cancer patients. We analysed these cells by RNA-seq and defined a set of lncRNAs that are specifically expressed in TIL subsets.

Modelling neuronal interactions using a directed network can be used to provide insight into the activity of the brain during experimental tasks. Magnetoencephalography (MEG) allows for the observation of the fast neuronal dynamics necessary to characterize the activity of sources and their interactions. A network representation of these sources and their con- nections can be formed by mapping them to nodes and their connection strengths to edge weights. Dynamic Causal Modelling (DCM) presents a Bayesian framework to estimate the parameters of these networks, as well as the ability to test hypotheses on the structure of the network itself using Bayesian model comparison. DCM uses a neurologically-informed representation of the active neural sources, which leads to an underdetermined system and increased complexity in estimating the network parameters. This work shows that inform- ing the MEG DCM source location with prior distributions defined using a MEG source localization algorithm improves model selection accuracy. DCM inversion of a group of can- didate models shows an enhanced ability to identify a ground-truth network structure when source-localized prior means are used.
Master of Science

Les modèles mathématiques sont à l’essence même de chaque domaine de l’ingénierie que ce soit pour le contrôle, la prédiction et le diagnostic. Dans cette thèse, nous abordons l’identification des systèmes, c’est-à-dire la modélisation d’un système dynamique avec des données expérimentales. Plus particulièrement, nous considérons l’identification par erreur de prédiction (Prediction Error). Afin d’obtenir un modèle identifié précis, les données doivent satisfaire une propriété fondamentale : l’informativité. Dans la littérature, l’informativité des données a été largement étudiée pour l’identification des systèmes linéaires une-entrée une-sortie. Peu de résultats sont néanmoins présents pour le cas des systèmes linéaires plusieurs-entrées plusieurs-sorties (multiple-inputs multiple-outputs en anglais ou MIMO) à l’heure où les systèmes deviennent de plus en plus complexe. Ainsi, nous nous focalisons en premier à développer des conditions pour vérifier que l’informativité est garantie pour l’identification des systèmes MIMO linéaires en boucle ouverte et en boucle fermée. Cependant, la plupart des systèmes réels ont des comportements non-linéaires. Heureusement, l’identification par Prediction Error peut s’avérer être un outil efficace pour identifier certaines classes de systèmes non-linéaires. C’est le cas, par exemple, de la classe des systèmes dits Hammerstein, i.e., des systèmes dont la non-linéarité se trouve à l’entrée du système. Dans cette thèse, nous étudions l’informativité des données pour l’identification d’une classe particulière de systèmes Hammerstein. Cette étude se motive par un système réel étudié dans cette thèse : le gyroscope MEMS. Un gyroscope MEMS est un micro-capteur inertiel mesurant des vitesses de rotation. Ils se distinguent des gyroscopes optiques par ses multiples avantages tels que sa petite taille, son faible coût et sa faible consommation d’énergie. Cependant, il reste moins précis que les gyroscopes optiques. Le gyroscope MEMS est alors placé en boucle fermée afin d’augmenter sa précision. Afin de faire la synthèse d’un correcteur optimal, il nous faut déterminer un modèle précis des dynamiques du gyroscope MEMS. Dans la littérature, les modèles des dynamiques du gyroscope MEMS sont souvent incomplet. Une méthode de modélisation basée sur l’identification des systèmes est alors proposée pour obtenir un modèle plus complet et précis. Nous observons alors que l’étude de l’informativité des données faite précédemment s’applique dans ce cas pratique
Mathematical models have a crucial place in every engineering field. They can be used for several purposes such as the design of a controller, the prediction, the health monitoring of a system, etc. In this thesis, we deal with system identification which is the scientific field consisting in the modeling of a system with experimental data. More particularly, we will consider the Prediction Error method. In order to get an accurate identified model, the data must guarantee one fundamental property which is the informativity. The data informativity has been largely studied for the identification of linear single-input single-output systems. However, few results can be found for the identification of linear multiple-inputs multiple-outputs (MIMO) systems. This is inconvenient since the systems get more and more complex. Hence, in the first part of this thesis, we focus on developing new conditions to verify the data informativity for the open-loop and closed-loop identification of linear MIMO systems. However, most of real-life systems have nonlinear dynamics. Fortunately, Prediction Error identification can be used as an efficient tool for the modeling of some classes of nonlinear systems such as Hammerstein systems, i.e., systems where the nonlinearity is found at the input of the system. In this thesis, we study a particular class of Hammerstein systems. The motivation of this study comes from the real-life considered in this thesis : the MEMS gyroscope. A MEMS gyroscope is a micro-sensor that measures angular rates. It has several advantages such as its small size, its low energy consumption and its cheap price. However, it is less accurate than its optical counterpart. In order to tackle this accuracy issue, the MEMS gyroscope is put in closed-loop. Of course, we want to design an optimal controller. For this purpose, we need to derive an accurate model of the dynamics of the MEMS gyroscope. In the literature, the proposed models are not enough complete. Therefore, in this thesis, we develop an identification method that yields an accurate and complete model of the dynamics of the MEMS gyroscope. We observe that the previous study of the data informativity can be applied to this real-life problem

Background. Multiple system atrophy (MSA) is a rare, sporadic disease characterized by autonomic failure and a various combination of parkinsonism and cerebellar dysfunction. Currently, development of new treatment strategies in MSA is hampered by the lack of reliable diagnostic and disease-progression biomarkers. The aim of this study was to investigate brain microstructural abnormalities in MSA through diffusion and neuromelanin-sensitive magnetic resonance imaging (MRI) and their relationship with clinical manifestations. Methods. Clinical evaluation and MRI were performed on 11 MSA patients, 19 Parkinson’s Disease (PD) and 18 healthy controls (HC). MRI scans included structural, diffusion (dMRI) and neuromelanin-sensitive sequences. dMRI was applied through a novel technique, neurite orientation dispersion and density imaging (NODDI). Compared to previous dMRI techniques, NODDI allows the simultaneous evaluation of the integrity of the intracellular and extracellular compartments, while gathering information on the orientation of axons and dendrites. Neuromelanin-sensitive MRI was used to quantitively investigate the integrity of substantia nigra (SN) and locus coeruleus (LC). Results. Median duration of symptoms in MSA patients was 3 years (range 1-6). Age was not significantly different across subgroups. Compared to PD, MSA patients had reduced neurite density index (NDI) in the middle cerebellar peduncle (MCP) and in the pons (Mann-Whitney U=44.0, p=0.019 and U=52.0, p=0.050), indicating white matter degeneration in these locations, and increased free water fraction (FWF), indicating grey matter loss, in the putamen, caudate and cerebellar lobule VI grey matter (U=146.0, p=0.019; U=145.0, p=0.021; U=154.0, p=0.006 respectively). Neuromelanin content was not different in SN and LC between PD and MSA, although this was reduced in the posterior SN and intermediate part of LC compared to HCs (Kruskal Wallis H=11.363, p=0.003 and H=13.788, p=0.001), indicating similar, significant degeneration of these nuclei in both conditions. No significant correlations were found between motor scores and MRI parameters in the SN, putamen, and MCP and pons. LC neuromelanin loss in the rostral and/or intermediate sections was significantly associated with greater cognitive, depressive and REM sleep behaviour disorder (RBD) symptoms scores in MSA. Symptoms of dysautonomia were not associated with diffusion or neuromelanin content measures. Conclusion. Multimodal MRI with diffusion and neuromelanin evaluation may help define structural abnormalities in the early stages of MSA. NODDI seems a promising technique to simultaneously evaluate multiple microstructural parameters in critical locations of MSA pathology, such as the basal ganglia, cerebellum, and pons. Neuromelanin content evaluation is useful for defining SN and LC degeneration, although this occurs similarly in MSA and PD. In MSA, LC degeneration is associated with greater depressive, cognitive and RBD symptoms. Longitudinal investigations are needed to establish whether these MRI parameters may serve as disease-progression biomarkers.

Modéliser la réponse mécanique des matériaux consiste à établir une relation (un modèle) entre contraintes et déformations,dépendant de paramètres identifiés à partir de données expérimentales issues d’essais mécaniques. D’une part, l’identification basée sur des essais homogènes n’apporte pas d’information sur la réponse du matériau soumis à des sollicitations complexes. D’autre part, l'identification basée sur des essais multiaxiaux est plus coûteuse numériquement et impose le choix du modèle a priori. Récemment, a émergé la possibilité de représenter le comportement mécanique des matériaux par une base de données plutôt que via une loi de comportement au travers de la « Data Driven Computational Mechanics ». Dans ce cadre, l’algorithme Data-Driven Identification (DDI) développé par Leygue et al. (2018) permet d’estimer le champ de contrainte lors d’un essai multiaxial.Cette thèse propose d’explorer la réponse cinématique puis mécanique complète de membranes élastomères sollicitées en grandes déformations multiaxiales grâce à un montage expérimental original mettant en jeu un hexapode. La méthode DDI est ensuite utilisée pour déterminer la réponse en contraintes du matériau lors de ces essais. Deux développements sont finalement proposés : une méthode d’identification alliant DDI (sans modèle) et modèles de comportement, et une proposition d’amélioration des géométriesd’éprouvettes pour les essais multiaxiaux
Modeling the mechanical response of materials involves the derivation of a relationship (a model) between stresses andstrains, depending on parameters. These parameters are identified from experimental data obtained from mechanical tests. On the one hand, identification based on simple tests (uniaxial tension, for example) provides no information on the response of materials subjected to complex loading conditions. On the other hand, identification based on multiaxial tests is more costly numerically and requires the model to be chosen at the outset of the procedure. Recently, the possibility of representing themechanical behaviour of materials by a database rather than via a behaviour law has emerged through "Data-Driven Computational Mechanics". On this basis, the Data-Driven Identification (DDI) algorithm developed by Leygue et al. (Computer Methods in Applied Mechanics and Engineering, 331, 184-196 (2018)) can be used to estimate the stress fieldduring a multiaxial test. The present thesis explores the complete kinematic and mechanical response of elastomer membranes subjected to multiaxial large strain, using an original experimental set-up involving a hexapod. The DDI method is then used to determine the stress response of the material during these multiaxial tests. Two developments are finally presented: an identification method combining DDI (model-free) and standard constitutive models, and a proposal forimproving sample geometries for multiaxial tests

Syftet med examensarbetet är att diskutera och belysa hur några skolor arbetar med attidentifiera och stötta elever med dyslexi och sätta det i relation till skolans styrdokument. Föratt kunna undersöka dessa frågeställningar användes semistrukturerade intervjuer med sexstycken lärare som alla arbetar på olika skolor. Resultaten visar att skolorna arbetar med attkartlägga och följa upp eleverna och att lärarna försöker kompensera elevers svårighetergenom olika pedagogiska stöd och tekniska hjälpmedel. Skolorna arbetar inkluderande och deanpassar undervisningen och lärandemiljön utifrån elevernas olika behov och svårigheter.Resultaten visar också det finns dilemman som lärarna mött under sitt arbete med att hjälpaeleverna med dyslexi så som att skolan ska vara kompensatorisk men samtidigt varainkluderande, tidsbrist, hjälpmedlen kan vara utpekande och att eleverna behöver få endiagnos för att skolan ska få extra resurser.
The aim of the study is to discuss and illustrate how some schools are working to identify andsupport pupils with dyslexia and how it relates to their policy documents. To investigate thesequestions, semi-structured interviews were chosen as the method. Six teachers, all working atdifferent schools, were interviewed. The results show that the school are working to identifyand follow up students, and that educators are trying to compensate for pupils difficultiesthrough a variety of educational supports and assistive technology. Schools working in aninclusive and personalizing the teaching and learning environment according to puppilsdifferent needs and difficulties. The results also show there are dilemmas that educators facedin their effort to help students whit dyslexia as that the school should be compensatory but atthe same time inclusive, lack of time, that the us of aids can become stigmatizing, andstudents need a diagnosis for the school to get extra resources.

As waste is separated at the source, the success of the Swedish recycling system largely depends on an active participation of households. However, especially young people were found to not consistently follow their local recycling schemes. A recent and promising approach to tackle such suboptimal household recycling behaviour (HRB) is the use of persuasive technology. To understand and further its context-specific potential, this research aims to explore the key elements of persuasive technology which aspires to promote sustainable household recycling behaviour among adults in Sweden. The chosen methodology is research through design. Based on the results of a literature review and online survey among target users (N=50), a mobile phone application was designed in an iterative manner. Through these activities, the following key elements were established: (1) easy access to information about optimal household recycling behaviour, (2) employment of several motivational strategies, (3) recognition of differences between local recycling schemes, (4) regard of users as equals and (5) use of a readily accessible technology channel. The impact of these elements depends on the users’ ability to carry out the target behaviour and therefore on a well-functioning recycling system. The technological format of persuasive technology interventions was found to spark the target users’ curiosity. Using this as a ‘carrot’, a well-designed content is argued to encourage repeated use and a reflection process to help break unsustainable household recycling habits.
Eftersom avfall separeras vid dess källa, beror Sveriges återvinningssystems framgång i stor utsträckning på aktivt deltagande hushåll. Studier har visat att särskilt unga inte följer lokala återvinningsrutiner på ett konsekvent sätt. Ett nytt och lovande tillvägagångssätt för att hantera detta suboptimala hushållsåtervinningsbeteende (English: household recycling behaviour (HRB)) är användningen av övertygande teknik. Syftet med denna studie är att utforska de viktigaste delarna av övertygande teknik som strävar till att främja ett hållbart hushållsåtervinningsbeteende hos vuxna i Sverige. Den valda metodiken är forskningsdriven design. Baserat på resultat från en litteraturstudie samt en online enkätundersökning fokuserad på målgruppen unga (N=50), designades en mobilapplikation genom ett iterativt tillvägagångssätt. Studien identifierade följande nyckelelement för en övertygande design i domänen: (1) Lättillgänglig information angående optimalt hushållsåtervinningsbeteende, (2) Användning av flera motivationsstrategier, (3) Identifiering av skillnader mellan lokala återvinningsrutiner, (4) Betrakta användare som jämställda och (5) Användning av lättillgänglig teknikkanal. Effekterna av dessa element är beroende av att användarna även har möjlighet att utföra den önskade sopsorteringen, och därför på ett välfungerande återvinningssystem. Designlösningen att använda sig av övertygande teknik i form av en app visade sig väcka målanvändarnas nyfikenhet. Studien visar att en väldesignad app kan fungera som en 'morot' för att uppmuntra användning över en längre tid, och skapa en reflektionsprocess som kan bryta ohållbara återvinningsvanor.

Based on the description and definition of technology and processes falling within the vision of the fourth industrial revolution with the aim of creating intelligent factories, this diploma thesis deals with the principles of the Industry 4.0 concept in Hilti's production plant with a focus on transport and supply of production equipment. The aim of the work is to create a comprehensive proposal that takes into account all the necessary aspects associated with upgrading the existing state of inter-operational transport in a particular production line to fully automated, flexible and autonomous transport of materials and products in the context of Industry 4.0. A prerequisite for creating a design is the connection of automatically guided vehicles (AGVs) serving individual transport orders. The selection of the vehicle was made taking into account the safety of movement, the method of charging, the system and network integrity of existing and proposed technologies and components. The intention is not only to automate the inter-operational service, but also on the basis of the created automation concept, the ability to autonomously procure the flow of material and products. The mathematical calculation of capacity planning in the production line helped to determine the total load and the number of vehicles needed for continuous procurement of transport requirements. The result of the design part is also the design of specific transport routes and transport conditions that AGV vehicles must comply with in order to maintain a high level of safety. Transparency and a constant overview of transported products is provided by the presented scheme for identification of production batches, Auto-ID system. The financial efficiency of the whole project elaborated in the diploma thesis is evaluated as payable after 4 years from the implementation of the proposal. The financial efficiency of the whole project elaborated in the diploma thesis is evaluated as payable after 4 years from the implementation of the proposal due to high labor costs.

In automatic speech recognition systems (ASRs), training is a critical phase to the system?s success. Communication media, either analog (such as analog landline phones) or digital (VoIP) distort the speaker?s speech signal often in very complex ways: linear distortion occurs in all channels, either in the magnitude or phase spectrum. Non-linear but time-invariant distortion will always appear in all real systems. In digital systems we also have network effects which will produce packet losses and delays and repeated packets. Finally, one cannot really assert what path a signal will take, and so having error or distortion in between is almost a certainty. The channel introduces an acoustical mismatch between the speaker's signal and the trained data in the ASR, which results in poor recognition performance. The approach so far, has been to try to undo the havoc produced by the channels, i.e. compensate for the channel's behavior. In this thesis, we try to characterize the effects of different transmission media and use that as an inexpensive and repeatable way to train ASR systems.

Wydział Biologiczno-Chemiczny. Instytut Chemii
Identyfikacja związków organicznych w próbkach pochodzenia naturalnego o skomplikowanym składzie matrycy stanowi jedno z największych wyzwań współczesnej chemii analitycznej. Metoda chromatografii gazowej ze spektrometrią mas (GC/MS) pozwala podjąć to wyzwanie. Podstawą identyfikacji analitów w tej metodzie jest jednoczesne użycie dwóch parametrów analitycznych: widma mas i chromatograficznego indeksu retencji (RI). Korzystanie z wartości RI dostępnych w źródłach literaturowych nie daje oczekiwanych wyników ze względu na zbyt małą liczbę zgromadzonych danych oraz ze względu na zbyt słabą odtwarzalność RI. Niniejsza rozprawa dotyczy zaprojektowania, stworzenia i praktycznego wdrożenia systemu komputerowej identyfikacji związków organicznych metodą GC/MS. Celem głównym było stworzenie takiego narzędzia, które umożliwi zbieranie RI i widm w sposób automatyczny: szybko i z wykluczeniem błędów wprowadzania ręcznego. Rozprawa ma formę komentarza do pięciu artykułów naukowych o łącznym współczynniku Impact Factor (2012) równym 18,739. Wszystkie artykuły dotyczą publikacji wartości chromatograficznych indeksów retencji. W trzech artykułach prowadzona jest analiza rozpoznawcza, która polega na zastosowaniu metod przewidywania wartości RI dla substancji, które nie posiadały dotychczas tych parametrów zmierzonych. Niniejsze artykuły powstały z częściowym wspomaganiem przez opracowany system. W języku programowania Pascal (z pakietu Delphi 5.0) zbudowano relacyjną bazę danych (RI_mz), która przechowuje informacje o próbkach, chromatogramach, widmach mas i RI zidentyfikowanych analitów. W środowisku MSD Chemstation (Agilent Technologies) napisano programy (makra) umożliwiające obustronną wymianę danych pomiędzy zarejestrowanymi chromatogramami a bazą RI_mz. Zgromadzone dane dotyczą analitów zidentyfikowanych w próbkach pochodzenia naturalnego: ekstrakty z tkanek roślin, eksudaty, olejki eteryczne, ekstrakty produktów pszczelich. W bazie RI_mz zebrano też dane dostępnych w naszym laboratorium substancji wzorcowych oraz zsyntetyzowanych ich pochodnych. Wszystkie pomiary wykonywano w jednakowych warunkach z użyciem kolumny chromatograficznej tego samego typu (polidimetylosiloksan z 5% domieszką grup fenylowych). Zapewniona została w ten sposób wysoka powtarzalność uzyskanych wyników. System został wdrożony do praktycznego użycia i działa w laboratoriach trzech uczelni: Uniwersytetu w Białymstoku, Uniwersytetu Medycznego w Białymstoku oraz Politechniki Białostockiej.
The identification of organic compounds coming from the natural samples with complex matrices is one of the greatest challenges of modern analytical chemistry. The gas chromatography-mass spectrometry (GC/MS) method allows to undertake this challenge. In this method, the basis of analytes identification is the simultaneous use of two analytical parameters: mass spectrum and chromatographic retention index (RI). The use of available in the literature RI values does not give expected results due to too small number of collected data and because of too poor reproducibility of RI. The presented dissertation deals with the design, development and practical implementation of the computer system of organic compounds identification with the GC/MS method. The main objective was to create a tool that enables the collection of RI and spectra in an automatic way: rapidly and without errors coming from the manual data entering. The dissertation takes the form of the commentary on the five research articles with the total Impact Factor (2012) equal to 18,739. In all of the articles the values of chromatographic retention indices were published. In three articles, the prospecting analysis is performed. It is based on the implementation of RI prediction methods for substances which did not have this parameter measured yet. These articles were created, partially, with the aid of the developed system. The Pascal programming language (from Delphi 5.0 package) was used to built the relational database (RI_mz) which stores the information about samples, chromatograms, mass spectra and RI for identified analytes. In the MSD Chemstation (Agilent Technologies) environment, there were written programs (macros) enabling two-way data exchange between registered chromatograms and RI_mz database. The collected data relate to the analytes identified in the samples of natural origin: plant tissues extracts, exudates, essential oils, bees products extracts. In the RI_mz database there were also collected the data for the reference substances available in our laboratory, and their synthesized derivatives. All measurements were performed under the same conditions with the use of chromatographic column of the same type (polydimethylsiloxane with 5% phenyl groups). In such a way, high reproducibility of the results was ensured. The system was implemented for practical use and it works in the laboratories of three universities: University of Bialystok, Medical University of Bialystok and Bialystok University of Technology.