No description available.

[pt] Esse trabalho realiza uma extensa revisão bibliográfica da literatura, onde pontos de inovação são observados e novas soluções para os principais problemas no uplink são propostas. Baseados no algoritmo adaptativo Recursive Least Squares (RLS), são apresentadas duas variações regularizadas que conjuntamente detectam a atividade e decodificam os sinais dos dispositivos, sem a necessidade de estimação explícita do canal. Além disso, duas técnicas de detecção por listas são desenvolvidas de modo a refinar o processo de ajuste dos pesos do algoritmo RLS e assim mitigar possíveis propagações de erros. Dividindo o problema em grafos fatoriais, são propostos algoritmos baseados em troca de mensagens de modo a realizar conjuntamente a estimação de canal e detecção de atividade. Além de uma nova derivação das mensagens, são exploradas técnicas de agendamento dinâmico, com base na convergência do algoritmo. Por fim, é apresentada uma solução completa baseada na troca de mensagens, que realiza conjuntamente a estimação de canal, detecção de atividade e decodificação dos dados transmitidos. Resultados numéricos são fornecidos com o objetivo de comparar o desempenho dos algoritmos propostos aos existentes na literatura. Análises de complexidade computacional, evolução dos estados, convergência e ordem de diversidade também são realizadas, assim como a derivação da soma das taxas para o uplink. Por fim, são apresentadas as conclusões obtidas dos trabalhos realizados e discutidas direções para trabalhos futuros. / [en] This work, presents an extensive literature review that highlights innovation opportunities and presents novel solutions for the main uplink mMTC problems. Based on the adaptive Recursive Least Squares (RLS) algorithm, the proposed regularized techniques jointly performs activity detection and signal decoding, without the need to perform explicit channel estimation. In order to improve the detection performance, a list detection technique that uses two candidate-list schemes is developed. Rewriting the problem with factor-graphs, novel message-passing algorithms with dynamic scheduling that jointly estimates the channels and detects devices activity are proposed. Lastly, a complete message-passing solution is presented, where LDPC decoding beliefs are introduced in the system, in a way that the algorithm besides the channel estimation and activity detection, also jointly decodes the signals. In order to evaluate the proposed techniques, numerical results are provided as well as a computational complexity, state-evolution, convergence and a diversity analysis. Uplink sum-rate expressions that take into account metadata collisions, interference and a variable activity probability for each user are also derived. Finally, conclusions and future directions are discussed.

La stratégie de l’Union Européenne pour atteindre les objectifs climatiques, est d’augmenter progressivement la part d’énergies renouvelables dans le mix énergétique et d’utiliser l’énergie plus efficacement de la production à la consommation finale. Cela implique de mesurer les performances énergétiques du bâtiment et des systèmes associés, indépendamment des conditions climatiques et de l’usage, pour fournir des solutions efficaces et adaptées de rénovation. Cela implique également de connaître la demande énergétique pour anticiper la production et le stockage d’énergie (mécanismes de demande et réponse). L’estimation des besoins énergétiques et des performances énergétiques des bâtiments ont un verrou scientifique commun : l’identification expérimentale d’un modèle physique du comportement intrinsèque du bâtiment. Les modèles boîte grise, déterminés d’après des lois physiques et les modèles boîte noire, déterminés heuristiquement, peuvent représenter un même système physique. Des relations entre les paramètres physiques et heuristiques existent si la structure de la boîte noire est choisie de sorte qu’elle corresponde à la structure physique. Pour trouver la meilleure représentation, nous proposons d’utiliser, des simulations de Monte Carlo pour analyser la propagation des erreurs dans les différentes transformations de modèle et, une méthode de priorisation pour classer l’influence des paramètres. Les résultats obtenus indiquent qu’il est préférable d’identifier les paramètres physiques. Néanmoins, les informations physiques, déterminées depuis l’estimation des paramètres, sont fiables si la structure est inversible et si la quantité d’information dans les données est suffisante. Nous montrons comment une structure de modèle identifiable peut être choisie, notamment grâce au profil de vraisemblance. L’identification expérimentale comporte trois phases : la sélection, la calibration et la validation du modèle. Ces trois phases sont détaillées dans le cas d’une expérimentation d’une maison réelle en utilisant une approche fréquentiste et Bayésienne. Plus précisément, nous proposons une méthode efficace de calibration Bayésienne pour estimer la distribution postérieure des paramètres et ainsi réaliser des simulations en tenant compte de toute les incertitudes, ce qui représente un atout pour le contrôle prédictif. Nous avons également étudié les capacités des méthodes séquentielles de Monte Carlo pour estimer simultanément les états et les paramètres d’un système. Une adaptation de la méthode de prédiction d’erreur récursive, dans une stratégie séquentielle de Monte Carlo, est proposée et comparée à une méthode de la littérature. Les méthodes séquentielles peuvent être utilisées pour identifier un premier modèle et fournir des informations sur la structure du modèle sélectionnée pendant que les données sont collectées. Par la suite, le modèle peut être amélioré si besoin, en utilisant le jeu de données et une méthode itérative. / The European Union strategy for achieving the climate targets, is to progressively increase the share of renewable energy in the energy mix and to use the energy more efficiently from production to final consumption. It requires to measure the energy performance of buildings and associated systems, independently of weather conditions and user behavior, to provide efficient and adapted retrofitting solutions. It also requires to known the energy demand to anticipate the energy production and storage (demand response). The estimation of building energy demand and the estimation of energy performance of buildings have a common scientific: the experimental identification of the physical model of the building’s intrinsic behavior. Grey box models, determined from first principles, and black box models, determined heuristically, can describe the same physical process. Relations between the physical and mathematical parameters exist if the black box structure is chosen such that it matches the physical ones. To find the best model representation, we propose to use, Monte Carlo simulations for analyzing the propagation of errors in the different model transformations, and factor prioritization, for ranking the parameters according to their influence. The obtained results show that identifying the parameters on the state-space representation is a better choice. Nonetheless, physical information determined from the estimated parameters, are reliable if the model structure is invertible and the data are informative enough. We show how an identifiable model structure can be chosen, especially thanks to profile likelihood. Experimental identification consists of three phases: model selection, identification and validation. These three phases are detailed on a real house experiment by using a frequentist and Bayesian framework. More specifically, we proposed an efficient Bayesian calibration to estimate the parameter posterior distributions, which allows to simulate by taking all the uncertainties into account, which is suitable for model predictive control. We have also studied the capabilities of sequential Monte Carlo methods for estimating simultaneously the states and parameters. An adaptation of the recursive prediction error method into a sequential Monte Carlo framework, is proposed and compared to a method from the literature. Sequential methods can be used to provide a first model fit and insights on the selected model structure while the data are collected. Afterwards, the first model fit can be refined if necessary, by using iterative methods with the batch of data.

Die vorliegende Arbeit beschreibt ein generisches Verfahren zur wahrscheinlichkeitsbasierten Erkennung von Fahrsituationen und Fahrmanövern für sicherheitsrelevante Fahrerassistenzsysteme. Fahrsituationen und Manöver unterliegen einer gewissen Unsicherheit basierend auf der unterschiedlichen Situationswahrnehmung bzw. Manöverdurchführung der Fahrzeugführer. Diese Unsicherheitskomponente wird in den Ansatz zur Situations- und Manövererkennung mit einbezogen. Ein weiterer Unsicherheitsaspekt beruht auf den ungenauen Umgebungsinformationen auf denen die Situations- und Manövererkennung basiert. Beide Unsicherheitsursachen sind völlig unabhängig voneinander und werden aus diesem Grund separat betrachtet und modelliert. Zur Modellierung dieser beiden Unsicherheitsaspekte bedient sich der vorgestellte Ansatz der Fuzzy-Theorie, der Theorie der probabilistischen Netzen sowie Verfahren zur Fehlerfortpflanzung und Sensitivitätsanalyse. Nach der theoretischen Vorstellung dieser Methodiken wird in der Arbeit detailliert auf den Einsatz und das Zusammenspiel der einzelnen Verfahren zur Erkennung der Fahrsituationen und Fahrmanöver eingegangen. Die Umsetzbarkeit des vorgestellten Verfahrens wird am Beispiel der Notbremssituation gezeigt. Die Notbremssituation setzt sich aus unterschiedlichen Teilsituationen und Manövern zusammen. Die Erkennung der einzelnen Situationen und Manöver sowie die Zusammenführung zur übergeordneten Notbremssituation wurden mit Hilfe des vorgestellten Verfahrens realisiert. Zur Evaluierung der Erkennungsgüte wurden sowohl Messdaten aus dem Straßenverkehr als auch realitätsnahe Daten, aufgezeichnet auf Versuchsstrecken, herangezogen. / The present work describes a generic method for the probabilistic identification of driving situations and driving manoeuvres for safety relevant driver assistance systems. Driving situations and driving manoeuvres underlie a certain uncertainty based on the different situation perception and manoeuvre execution of the driver. This uncertainty component is considered in the approach for the situation and manoeuvre identification. An additional uncertainty aspect is based on the inaccurate environment information, the identification of driving situations and manoeuvres depend on. Both uncertainty aspects are completely independent and are considered and modelled separately for this reason. For modelling both of these uncertainty aspects the present approach is using the fuzzy theory, probabilistic networks, as well as methods for error propagation and sensitivity analysis. After introducing these techniques theoretically, the application and the interaction of the single methods to identify the driving situations and manoeuvres is described in detail. The practicability of the introduced proceeding is shown exemplarily on the emergency brake situation. The emergency brake situation consists of several situation and manoeuvre components. The identification of the single situations and manoeuvres as well as the combination to the higher emergency brake situation is realised with the introduced proceeding. Measuring data gathered on road traffic and close to reality data measured on a test track were used to evaluate the identification quality.

Terrestrial laser scanning systems provide quick three-dimensional data acquisition with high spatial resolution. The technology can serve multiple applications such as monitoring and 3D modeling but several problems arise when scanning long objects such as tunnels, bridges, and corridors. Uncertainties in the data tend to propagate when conducting sequential registration on a long strip of point clouds and the errors can reach unacceptable levels only with a few scans. This thesis aims to examine this propagation of uncertainty caused by long strip registration and its effects on point cloud accuracy and precision. It further aims to evaluate the results of different ground control point distributions used for georeferencing and finally to construct guidelines and recommendations for the planning of terrestrial laser scanning projects of long objects. The thesis was divided into an analytical- and an experimental study. The analytical study explored the theoretical effects of long strip registration on a point cloud’s uncertainty by simulating repeated target-based registration in a 95-meter-long and 2.5-meter-wide corridor. This resulted in standard uncertainties of the point cloud coordinates on dm-level after the last registration and a clear trend in exponential growth. Tie point distribution in the overlaps between scans proved to have a big effect on the precision of the final point cloud. In the experimental study, a corridor of the same dimension was scanned from 16 stations and georeferenced in a local control network established at the scan site. Indirect registration in two steps was used, first by performing automatic feature-based registration in RealWorks before the point cloud was georeferenced. The results of different ground control point (GCP) scenarios were compared and evaluated by studying check point residuals. All scenarios had numerous check points that failed verification as the residuals exceeded the limit set to 15 mm. An even distribution of almost one GCP in every scan scene yielded the best results with a maximum check point residual of 3 cm in 3D. When the GCPs were clustered at one side of the corridor the check point residuals reached dm level at the opposite end. Also, a linear trend of increasing residuals as a function of distance from the point to the nearest GCP was evident. The linear trend was less evident when the GCPs were distributed evenly across the corridor. Their distribution proved more important than their quantity. The vertical coordinate component and the horizontal component perpendicular to the long side of the object proved less precise and accurate after long strip registration. This was apparent in both the analytical and the experimental study. / Terrester laserskaning erbjuder snabb, tre-dimensionell datafångst med hög punkttäthet. Teknologin täcker flertalet användningsområden såsom monitorering och 3D-modellering men problem kan uppstå vid skanning av långa objekt såsom tunnlar, broar och korridorer. Osäkerheter i data tenderar att fortplantas när registrering av en lång rad punktmoln genomförs sekventiellt och felen kan nå oacceptabla nivåer bara med några få skanningar. Detta examensarbete syftar till att undersöka denna fortplantning av mätosäkerhet orsakad vid registrering av långsträckta objekt och dess påverkan på punktmolnets noggrannhet och precision. Vidare syftar arbetet till att utvärdera resultatet av olika distributioner av utgångspunkter vid georeferering och slutligen till att utforma riktlinjer och rekommendationer för planering av terrestra laserskanningsprojekt över långa objekt.  Arbetet delades upp i en analytisk- och en experimentell studie. Den analytiska studien undersökte de teoretiska effekterna av långsträckt registrering på ett punktmolns osäkerhet genom att simulera upprepad punktbaserad registrering i en 95-meter-lång och 2.5-meter-bred korridor. Detta resulterade i standardosäkerheter av punktmolnets koordinater på dm-nivå efter den sista registreringen samt en tydlig exponentiell ökning. Distributionen av konnektionspunkter i överlappen mellan skanningarna visade sig ha en stor effekt på precisionen av det slutliga punktmolnet. I den experimentella studien skannades en korridor av samma dimensioner från 16 stationer innan det georefererades till ett lokalt stomnät som etablerats på platsen. Indirekt georeferering i två steg användes, först genomfördes automatisk objektbaserad registrering i RealWorks innan punktmolnet georefererades. Resultatet av olika scenarion av utgångspunkter jämfördes och utvärderades genom att studera avvikelser hos kontrollpunkter. Alla scenarion hade flertalet kontrollpunkter vars avvikelser överskred den satta gränsen på 15 mm. En jämn distribution med upp till en utgångspunkt i för varje skanning medförde de bästa resultaten med en maximal avvikelse hos kontrollpunkterna på 3 cm i 3D. När utgångspunkterna var placerade på ena änden av korridoren så nådde avvikelserna dm-nivå på den motsatta änden. Dessutom så uppstod en tydlig linjär trend med ökande avvikelser som en funktion av avståndet till närmsta utgångspunkt. Den linjära trenden var mindre tydlig när utgångspunkterna var jämnt fördelade över korridoren. Deras distribution visade sig vara viktigare än deras kvantitet. Punktkoordinaternas höjdkomponent och plana komponent vinkelrätt mot objektets långsida visade sig vara mindre precis och noggrann efter långsträckt registrering. Detta var tydligt både i den analytiska och den experimentella studien.