Thesis (Ph. D.)--Louisiana State University and Agricultural and Mechanical College, 2004. / Includes bibliographical references.

Στην προκειμένη εργασία μελετάται όλο το θεωρητικό υπόβαθρο για τον προβλεπτικό έλεγχο για τις δύο κατηγορίες συστημάτων (Single Input-Single Output SISO, Multiple Input-Multiple Output MIMO). Αρχικά μελετάται η πρώτη μορφή προβλεπτικού ελέγχου που ήταν ο δυναμικός έλεγχος μητρών (DMC). Στην συνέχεια ακολουθεί το πρόβλημα του βέλτιστου προβλεπτικού ελέγχου διακριτού χρόνου όπως αυτό παρουσιάζεται και αναλύεται στην θεωρία του κ. Μaciejowski. Αμέσως μετά μελετάται πάλι το πρόβλημα εύρεσης βέλτιστου προβλεπτικού ελέγχου διακριτού χρόνου αλλά με την χρησιμοποίηση των διακριτών ορθοκανονικών συναρτήσεων βάσης Laguerre όπως αναλύεται από τον κ. Wang στο βιβλίο του. Στις δύο επόμενες ενότητες παρουσιάζονται οι ορθοκανονικές συναρτήσεις βάσης Laguerre συνεχούς χρόνους καθώς και μια άλλη κατηγορία, οι συναρτήσεις Κautz και αναλύεται ο τρόπος που υπολογίζεται ο προβλεπτικός έλεγχος συνεχούς χρόνου με τη χρήση αυτών. Αφού ο αναγνώστης αποκτήσει τις γνώσεις που χρειάζονται πάνω στον προβλεπτικό έλεγχο, ακολουθεί μια πρακτική εφαρμογή πάνω σε ένα ελικόπτερο 2 βαθμών ελευθερίας της Quanser. Εκεί αρχικά αφού περιγραφεί πλήρως η διάταξη μελετάμε τα προβλήματα ελέγχου πρώτα με Pole Placement στην συνέχεια με LQR καθώς και με την χρησιμοποίηση εκτιμητών κατάστασης όπως κάποιο παρατηρητή (observer) ή φίλτρο Kalman πάντα με τη βοήθεια του Μatlab και του Simulink. Επίσης όλη η θεωρία του ΜPC που μελετήσαμε έχει εφαρμοσθεί επιτυχώς σε προσομοίωση στο Μatlab και Simulink. Παρουσιάζονται ο κώδικας που χρειάζεται κάθε φορά καθώς και ενδιαφέροντα αποτέλεσματα για την απόκριση της διεργασίας. Επιπλέον μελετήθηκε το toolbox του Matlab για τον προβλεπτικό έλεγχο (MPC Toolbox). Τέλος οι παραπάνω έλεγχοι εφαρμόσθηκαν κατευθείαν στην πραγματική διεργασία (μη γραμμική) και τα αποτελέσματα ήταν ικανοποιητικά. / This work presents all the necessary theory for the Model Predictive Control for both system categories (Single Input-Single Output SISO, Multiple Input-Multiple Output MIMO). To start, the earliest form of MPC called dynamic matrix control (DMC) is studied. Then the optimal Model Predictive Control for discrete time is analyzed based on the theory that Maciejowski presented. Afterwards the same problem is studied using the discrete time Laguerre orthonormal base functions and the optimal Model Predictive Control is computed as presented in Wang’s theory. In the next two chapters the reader will be guided through the continuous time Laguerre and Kautz orthonormal base functions and how they are used in computing the optimal continuous time Model Predictive Control. Since the reader has acquired all the necessary knowledge about MPC, a practical approach on the Quanser’s two degrees of freedom helicopter follows. Initially, after we have fully described the plant and all its components, we study the control problems using the pole placement and LQR techniques along with state estimators such as observers and Kalman filter, always in the Matlab and Simulink enviroment. Next, the MPC approaches we’ve studied are applied successfully, again using Matlab and Simulink. In every case, all the necessary programs and results are presented in detail. Addionally, the Matlab MPC Toolbox is studied along with its results for the problem. Finally all those controls are applied directly to the real nonlinear plant successfully and the results are discussed.

Flying animals such as bats, birds, and moths are actively studied by researchers wanting to better understand these animals’ behavior and flight characteristics. Towards this goal, multi-view videos of flying animals have been recorded both in lab- oratory conditions and natural habitats. The analysis of these videos has shifted over time from manual inspection by scientists to more automated and quantitative approaches based on computer vision algorithms. This thesis describes a study on the largely unexplored problem of 3D pose estimation of flying animals in multi-view video data. This problem has received little attention in the computer vision community where few flying animal datasets exist. Additionally, published solutions from researchers in the natural sciences have not taken full advantage of advancements in computer vision research. This thesis addresses this gap by proposing three different approaches for 3D pose estimation of flying animals in multi-view video datasets, which evolve from successful pose estimation paradigms used in computer vision. The first approach models the appearance of a flying animal with a synthetic 3D graphics model and then uses a Markov Random Field to model 3D pose estimation over time as a single optimization problem. The second approach builds on the success of Pictorial Structures models and further improves them for the case where only a sparse set of landmarks are annotated in training data. The proposed approach first discovers parts from regions of the training images that are not annotated. The discovered parts are then used to generate more accurate appearance likelihood terms which in turn produce more accurate landmark localizations. The third approach takes advantage of the success of deep learning models and adapts existing deep architectures to perform landmark localization. Both the second and third approaches perform 3D pose estimation by first obtaining accurate localization of key landmarks in individual views, and then using calibrated cameras and camera geometry to reconstruct the 3D position of key landmarks. This thesis shows that the proposed algorithms generate first-of-a-kind and leading results on real world datasets of bats and moths, respectively. Furthermore, a variety of resources are made freely available to the public to further strengthen the connection between research communities.

CoordenaÃÃo de AperfeiÃoamento de Pessoal de NÃvel Superior / Diante da necessidade em aperfeiÃoar as tecnologias existentes para a conversÃo de energia elÃtrica em sistemas de alta potÃncia, este trabalho tem por finalidade projetar, analisar e implementar experimentalmente uma tÃcnica de modulaÃÃo com o intuito de reduzir o conteÃdo harmÃnico da tensÃo de saÃda em um inversor multinÃvel com capacitor flutuante de trÃs nÃveis. Algumas das diversas tÃcnicas de modulaÃÃo (PSPWM, LSPWM, HE-PWM e CSV-PWM) foram analisadas e implementadas no controlador digital FPGA a fim de comparar suas caracterÃsticas de desempenho com a tÃcnica de modulaÃÃo proposta. AlÃm disso, foi realizado o estudo de perdas da topologia de inversor multinÃvel com capacitor flutuante de trÃs nÃveis e a anÃlise teÃrica da distorÃÃo harmÃnica total da modulaÃÃo proposta. Finalmente, o desenvolvimento digital das tÃcnicas mostrou resultados coerentes, com formas de onda obtidas experimentalmente com alta qualidade de resoluÃÃo. A comparaÃÃo entre as estratÃgias de modulaÃÃo em termos de DHT resultou positivamente à modulaÃÃo proposta, cujos resultados experimentais de DHT nas tensÃes de linha na saÃda apresentaram o melhor desempenho para toda a faixa de Ãndices de modulaÃÃo comparadas Ãs tÃcnicas PSPWM, LSPWM-POD e CSV-PWM e uma reduÃÃo de atà 4,5% em relaÃÃo à HE-PWM. Isto comprova o estudo teÃrico realizado e sua aplicaÃÃo no inversor multinÃvel com capacitor flutuante de trÃs nÃveis. / Given the need to improve the existing technologies for electrical energy conversion into high power systems, this works purpose to design, analyze and implement a modulation technique that aims to reduce the output voltage harmonic content on the three-level flying capacitor multilevel inverter. Some of the various conventional modulation techniques (PSPWM, LSPWM, HE-PWM e CSV-PWM) have been analyzed and implemented in FPGA controller in order to compare their performance features with the proposed modulation technique. Furthermore, the losses study of the three-level flying capacitor multilevel inverter topology and the total harmonic distortion theoretical analysis of the proposed modulation technique have been made. Finally, the digital implementation of the techniques showed consistent results with experimentally obtained waveforms with high quality resolution. The comparison between the modulation strategies in the THD rates resulted positively for the proposed modulation, which THD experimental results in the line output voltage showed the best performance for all range of modulation indexes compared to techniques PSPWM, LSPWM-POD and CSV-PWM and the reduction of up to 4.5% better than HE-PWM. This proved the theoretical study done and its application in three-level flying capacitor multilevel inverter.

Orientador: Maria Regina Wolf Maciel / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Quimica / Made available in DSpace on 2018-08-11T20:12:49Z (GMT). No. of bitstreams: 1 Lima_DaniloRibeirode_M.pdf: 2097787 bytes, checksum: 174055436552208f0eeb28c29f157213 (MD5) Previous issue date: 2008 / Resumo: Biodiesel (ésteres alquílicos) é um combustível limpo derivado de fontes renováveis como óleo vegetal e gordura animal. O biodiesel é biodegradável, possui elevado ponto de fulgor, não-tóxico e produz menos COx e resíduos aromáticos do que o diesel de petróleo. Atualmente, devido ao elevado preço do óleo de soja e de canola, a produção de biodiesel a partir de matérias-primas de baixo custo como óleo residual pode aumentar a competitividade econômica do biodiesel frente ao diesel de petróleo, além de ser uma maneira de reciclar resíduos. A etanólise do óleo residual coletado em restaurante comercial foi estudada utilizando os catalisadores hidróxido de sódio e de potássio. Com teor de acidez superior a 3,8%, a transesterificação do óleo residual apresentou problemas comuns a esse tipo de matéria-prima, como formação de sabão e dificuldade de separação de fases. Assim, os planejamentos experimentais foram realizados em condições específicas de modo a resolver estes problemas. O óleo residual foi caracterizado pelo teor de acidez, umidade, composição de ácidos graxos, índice de iodo, viscosidade cinemática e massa específica. Primeiramente, foi realizado um planejamento fatorial fracionário para determinação das variáveis (temperatura, razão molar etanol:óleo, tipo e quantidade de catalisador e tempo de reação) de maior influência na conversão. O hidróxido de sódio formou uma mistura pastosa e homogênea após a destilação do etanol e, por isso, foi escolhido o catalisador hidróxido de potássio para os próximos planejamentos. A temperatura de reação de 30ºC também se mostrou a mais adequada. Além disso, os modelos estatísticos apresentaram curvatura, sendo necessária a adição de pontos axiais nos próximos planejamentos. Posteriormente, foram realizados planejamentos do tipo composto central com as variáveis mais significativas obtidas em planejamentos anteriores, visando não somente a obtenção da maior conversão, mas também bons índices de rendimento de ésteres. Os resultados mostraram que uma quantidade de KOH de 1,3%, temperatura de 30°C, razão molar etanol:óleo residual de 8:1, tempo de 50 minutos foram adequados para uma maior conversão e rendimento de ésteres obtidos. As curvas de contorno e superfícies de resposta mostraram as melhores condições de reação de cada planejamento de experimentos. Por fim, as melhores condições de operação foram tomadas para a obtenção dos dados cinéticos (energia de ativação e constantes da taxa) / Abstract: Biodiesel (alkyl esters) is a clean burning fuel derived from a renewable lipid feedstock such as vegetable oil or animal fat. It is biodegradable, non-inflammable, nontoxic and produce much lesser COx, sulfur dioxide and unburned hydrocarbons than petroleum-based diesel. Nowadays, due to the price of virgin oil such as soybean and canola oils, the use of low-cost feedstock such as waste frying oils in a base-catalyzed process should help make biodiesel competitive in price with petroleum diesel, beyond being a suitable way to reuse waste materials. Waste frying oil ethanolysis was studied using sodium hydroxide and potassium hydroxide as catalysts. The waste oil was collected in restaurant boxes. The waste oil showed free fatty acids content up to 3.8% and typical problems related to this condition as soap formation and phase separation interference were observed. Then, experimental designs were carried out in specific conditions to solve such problems. The properties of waste oil studied were acid value, water content, fatty acids composition, iodine value, kinematic viscosity and density. At first, a fractional factorial experimental design was chosen to analyze the most significant factors (temperature, ethanol to oil molar ratio, type and amount of catalyst and time) on the conversion oil to biodiesel. The use of sodium hydroxide as catalyst resulted in a dense and homogeneous phase after ethanol distillation, then potassium hydroxide was chosen to the subsequent experimental designs. The temperature of 30ºC was found more suitable as well. Moreover, curvature was detected in the fractional factorial experimental designs studied and star points were included in the following ones. Afterwards, central composite designs were carried out with the most significant factors obtained in previous designs not only looking for the higher conversion rate but also good ester yield. The reaction was optimized using 1.3% KOH, 30°C, ethanol to oil molar ratio of 8:1 and 50 minutes of stirring. The surface response methodology was used to study the best conditions related to each design. At last, a computer-aided tool of this system was used to model the kinetic of biodiesel production in order to calculate kinetic parameters such as reaction rate constants and the activation energy / Mestrado / Desenvolvimento de Processos Químicos / Mestre em Engenharia Química

L’émergence de véhicules aériens sans pilote, généralement appelés drones, petits et bon marché favorise leur utilisation dans le domaine des applications civiles. Ces drones sont équipés de différents capteurs et ont la capacité de communiquer via des liaisons sans fil et ont la particularité de se déplacer librement dans l’espace, révolutionnant la gestion des applications de surveillance. Un réseau ad hoc de drones, Flying Ad hoc Networks (FANET) en anglais, est composé d’une flotte de drones autonomes et est utilisé lors de missions dans des environnements hostiles pour la surveillance ou l’inspection de sites dangereux ou inconnus. Les FANETs peuvent être également utilisés pour suivre et filmer des événements spéciaux comme une course de vélos ou un match de football, dans ce cas les liaisons doivent garantir un minimum de qualité. Les FANETs nécessitent une attention particulière en ce qui concerne l’économie d’énergie des UAV dont les ressources en énergie sont limitées, mais aussi d’être immunisés contre des attaques malveillantes. Au cours de cette thèse, nous nous sommes concentrés sur le problème de suivi d’une cible mobile utilisant une flotte de drones pour la filmer. Étant donné que la cible se déplace, les drones doivent la suivre en continu, et une liaison vers la station terrestre doit être disponible. Dans ce contexte, nous proposons une solution qui permet la coordination d’un ensemble de drones afin de maintenir un chemin optimal entre la cible et la station terrestre. Notre solution se révèle efficace en matière de gain en temps et en énergie. Nous avons également proposé une solution basée sur des protocoles hiérarchiques pour économiser plus d’énergie dans le processus de communication avec la station terrestre. Nous avons également développé une autre solution qui permet d’économiser plus d’énergie en forçant les nœuds égoïstes à participer dans le réseau et d’assurer le relais de paquets lorsqu’ils sont sollicités. En effet, si un nœud égoïste refuse de router des paquets d’autres nœuds, cela induit une charge supplémentaire pour le reste des nœuds du réseau. Nous avons validé l’apport de l’ensemble de nos solutions par évaluation de performances à l’aide de simulations. / The emergence of small and inexpensive Unmanned Aerial Vehicles (UAVs) promotes their use in several applications. UAVs are usually equipped with different sensors and have the ability to communicate via wireless connections. Their capability to fly freely in the space offers new opportunities to monitoring and tracking applications. A Flying Ad hoc Network (FANET) is composed of a fleet of autonomous UAVs and is used for monitoring applications in hostile environments, surveillance or site inspection. FANETs could also be used for filming special events such as bike races or soccer matches, so, the connections must guarantee a minimum of quality of service. In FANETs, saving energy of UAVs that have limited battery is very challenging and protecting the network from malicious attacks is even more difficult. In this thesis, we focus on tracking and filming a moving target using a fleet of UAVs. Since the target is moving, the UAVs have to follow it continuously, and a path to the ground station must be available. In this context, we propose an efficient solution that allows the coordination of the UAVs to maintain an optimal path between the target and the ground station. The proposed solution is time and energy efficient. We also propose a solution based on hierarchical protocols to save more energy in the communication process with the ground station. Another solution that allows energy saving is to force selfish nodes to participate in the network to route received packets towards their destination. Indeed, a selfish node is concerned only about its own welfare, refusing to route packets of other node, causing an extra charge for the rest of nodes in the network. We validate our solutions through simulation campaigns.

The tropical forests of South and Southeast Asia hold the highest gliding squirrel diversity but our knowledge of species diversity, ecology and major threats is limited. The present study was undertaken in Arunachal Pradesh, Northeast India between June 2011 and March 2015 to address the paucity of data available on gliding squirrels. Based on field and literature surveys, 14 species of gliding squirrels were detected in the state of Arunachal Pradesh. However, species such as Biswamoyopterus biswasi, which is reported as endemic to Namdapha National Park, were not detected. The high gliding squirrel diversity in this region could be related to a diversity of forest types and its location between the Himalayas and the Indomalayan region. Encounter rates with four different species revealed that Petaurista petaurista was most frequently detected in Namdapha National Park. Major threats include hunting for traditional medicine, cultural purposes or bushmeat, and habitat loss due to forest degradation caused by shifting cultivation. In addition, more intensive studies on population, ecology and conservation status are needed in order to design species and site specific conservation action plans in this region which represents the highest diversity of gliding squirrels globally.

The study of dynamics and performance of flying discs is motivated by how variations in their design features influence the aerodynamic characteristics and flight performance, particularly range. Reviews in the literature focus on the development of fundamental research in flying discs and on the performance of sports projectiles. Theoretical background on disc dynamics, range and sensitivity are given. This work investigates disc aerodynamics by performing wind tunnel experiments to measure aerodynamic loads on a set of generic parametric discs and a set of commercial golf discs. The parametric discs isolate various geometrical features such as thicknesses, cavities, leading edge curvatures and cambers to study the influence of these parameters on aerodynamic characteristics. The commercial golf discs consist of three different categories known as putter, mid-range and driver: these discs are used for short, medium and long flight range, respectively. To study the performance, aerodynamic data of the golf discs are used as inputs in numerical simulations to predict their flight trajectories and range. Effects of launch attitude on range are assessed to study the optimal initial launch conditions to achieve maximum range. Further, the simulation provides a direct approach to analyse range sensitivity on launch parameters. Results from the experiments show a number of significant findings on disc geometry. First, a cavity is fundamentally important for a disc to have satisfactory flying qualities: the presence of the cavity produces a significant aft shift in the aerodynamic position to minimize the pitching moment about the centre of the disc. Hence, the disc will have a minimal tendency to roll about the flight axis. Second, the thickness of a disc has a significant effect on its profile drag: increasing the thickness increases the profile drag. Third, a disc with a positive camber produces a relatively higher lift-to-drag ratio (CL/CD) compared to one with no camber. Fourth, the effect of tapering a flat leading or trailing edge of a disc leads to a reduction in its lift (within the angles of attack tested in the study) with a significant reduction in its drag. Fifth, it is shown that peak lift-to-drag ratio of a free flying disc is not necessarily a good indicator of performance because the angle of attack (and hence, lift-to-drag ratio) varies widely through a typical flight. Furthermore, a disc with a significant pitching moment will roll significantly about the flight path direction, further reducing the achievable range for a given lift or drag characteristic. Finally, a novel method to quantify disc sensitivity with respect to changes in launch conditions has been developed. The simulations show that the range sensitivity of each disc with respect to launch pitch angle varies significantly, with discs design for long range being much more sensitive (and therefore harder to throw accurately) than discs designed for short range.

The notion of airplane stability and control being a balancing act between stability and control has been around as long as aeronautics. The Wright brothers’ first successful flights were born of the debate, and were successful at least in part because they spent considerable time teaching themselves how to control their otherwise unstable airplane. This thesis covers four aspects of handling for large transport aircraft: large size and the accompanying low frequency dynamics, the way in which lifting surfaces and control system elements are modelled in flight dynamics analyses, the cockpit feel characteristics and details of how pilots interact with them, and the dynamic instability associated with Pilot Induced Oscillations. The dynamics associated with large transport aircraft are reviewed from the perspective of pilot-in-the-loop handling qualities, including the effects of relaxing static stability in pursuit of performance. Areas in which current design requirements are incomplete are highlighted. Issues with modelling of dynamic elements which are between the pilot’s fingers and the airplane response are illuminated and recommendations are made. Cockpit feel characteristics are examined in detail, in particular, the nonlinear elements of friction and breakout forces. Three piloted simulation experiments are described and the results reviewed. Each was very different in nature, and all were designed to evaluate linear and nonlinear elements of the cockpit feel characteristics from the pilot’s point of view. These included understanding the pilot’s ability to precisely control the manipulator itself, the pilot’s ability to command the flight path, and neuro-muscular modelling to gain a deeper understanding of the range of characteristics pilots can adapt to and why. Based on the data collected and analyzed, conclusions are drawn and recommendations are made. Finally, a novel and unique PIO prediction criterion is developed, which is based on control-theoretic constructs. This criterion identifies unique signatures in the dynamic response of the airplane to predict the onset of instability.

The performance of pilots in the aerospace environment is a critical factor in the success of modern air and space travel. Various methods of evaluating performances of pilots have been implemented and the search for improved means of evaluation is an ongoing process. Multiple factors influencing performance have been identified in the past. However, as the demands on the pilot’s performances varies with changing technology, so does the need to identify new risk factors, as well as ranking old and new factors in order of effect on performance. Aim The descriptive study aims to identify and rank risk factors affecting the performance of pilots as assessed by the Penalty Scoring System at a Precision World Flying Championship. Methods and materials Pilots participating at the 2008 World Precision Championship in Ried-Kircheim in Austria were requested to complete questionnaires regarding possible factors that could affect performance stress factors. Each questionnaire required the subject to answer 14 questions, relating to 17 possible factors. These questionnaires were linked to the participant’s individual score as per the official competition results. Results Out of a total number (n = 178) of pilot performances during a week period, 88 % (n=157) completed questionnaires. Only 57% (n=89) of these performances were included in the study, due to administrative difficulties preventing the accurate linking of performances to penalty scores. Out of the 17 possible risk factors, 4 factors (23 %) were identified as being significantly associated with the Penalty Scoring System. Age proved the most consistent factor, the younger pilots (youngest aged 21) performing consistently better than the older ones (oldest aged 67), even if the older pilots may have had more experience. Experience also proved reliable as a factor predicting outcome, as the performances of the moderate experienced group (having competed in 3 or less previous World championships) was associated with a lower penalty score. The mood of the pilots on the day of competing proved to be an effective way of predicting outcome, with a good mood associated with a lower penalty score. Any medical condition or medication used, were associated with a higher penalty score. The remaining factors (n=13) showed no association, although some (n=5) factors, like sleep deprivation and alcohol are known risk factors. Conclusions The study succeeds in showing an association between the Penalty Scoring System and 4 factors (Age, Experience, Mood and Medical conditions) affecting the performance of pilots. Although not the aim of this study, the conclusion can be made that the Penalty Scoring System may be a valuable tool in identifying risk factors affecting pilot’s performance. / Dissertation (MSc)--University of Pretoria, 2011. / School of Health Systems and Public Health (SHSPH) / Unrestricted