Esta dissertação é o resultado de um estudo sobre a Teoria de Grandes Desvios, no qual deu-se ênfase ao contexto de variáveis aleatórias independentes e identicamente distribuídas (v.a's i.i.d.). Com a finalidade de discutir a lguns dos principais resultados sobre Grandes Desvios, inicialmente apresentamos as definições e propriedades básicas. ~uma etapa seguinte apresentamos o teorema de Cramér-Chernoff para v .a's i. i .d. a valores em R . A seguir enunciamos um princípio de Grandes Desvios para quaisquer abertos e fechados de R. Numa etapa posterior estendemos o teorema de Cramér-Chernoff para v.a's i.i.d. a valores em Rd. No Capítulo final apresentamos, de maneira sintética, outras extensões do teorema de Cramér-Chernoff, tais como: o teorema de Sanov para v .a's i.i .d. e o princípio de Grandes Desvios para Cadeias de Markov finit as. Além disso, apresentamos algumas aplicações de Grandes Desvios em Estatística Matemática. / This thesis is the result of a study about the Large Deviations Theory, in which the context of independent and identically distributed (i.i.d.) random variables was emphasized. Aiming to discuss some of the main results about Large Deviations, first we present the definitions and the basic properties. In a second moment, we present Cramér-Chernoff's theorem for i.i.d random variables in R. Mter, we enunciate a Large Deviations principie for any open and closed sets in R . In another stage, we extend Cramér-Chernoff's theorem for i.i.d. random variables in Rd. In Jast chapter, we present, briefly, other extensions of CramérChernoff's theorem, such as Sanov's theorem for i.i.d. random variables and the Large Deviations principie for finite state Markov Chains. Besides, we present some applications of Large Deviations in Mathematical Statisitics.

Esta dissertação é o resultado de um estudo sobre a Teoria de Grandes Desvios, no qual deu-se ênfase ao contexto de variáveis aleatórias independentes e identicamente distribuídas (v.a's i.i.d.). Com a finalidade de discutir a lguns dos principais resultados sobre Grandes Desvios, inicialmente apresentamos as definições e propriedades básicas. ~uma etapa seguinte apresentamos o teorema de Cramér-Chernoff para v .a's i. i .d. a valores em R . A seguir enunciamos um princípio de Grandes Desvios para quaisquer abertos e fechados de R. Numa etapa posterior estendemos o teorema de Cramér-Chernoff para v.a's i.i.d. a valores em Rd. No Capítulo final apresentamos, de maneira sintética, outras extensões do teorema de Cramér-Chernoff, tais como: o teorema de Sanov para v .a's i.i .d. e o princípio de Grandes Desvios para Cadeias de Markov finit as. Além disso, apresentamos algumas aplicações de Grandes Desvios em Estatística Matemática. / This thesis is the result of a study about the Large Deviations Theory, in which the context of independent and identically distributed (i.i.d.) random variables was emphasized. Aiming to discuss some of the main results about Large Deviations, first we present the definitions and the basic properties. In a second moment, we present Cramér-Chernoff's theorem for i.i.d random variables in R. Mter, we enunciate a Large Deviations principie for any open and closed sets in R . In another stage, we extend Cramér-Chernoff's theorem for i.i.d. random variables in Rd. In Jast chapter, we present, briefly, other extensions of CramérChernoff's theorem, such as Sanov's theorem for i.i.d. random variables and the Large Deviations principie for finite state Markov Chains. Besides, we present some applications of Large Deviations in Mathematical Statisitics.

Esta dissertação é o resultado de um estudo sobre a Teoria de Grandes Desvios, no qual deu-se ênfase ao contexto de variáveis aleatórias independentes e identicamente distribuídas (v.a's i.i.d.). Com a finalidade de discutir a lguns dos principais resultados sobre Grandes Desvios, inicialmente apresentamos as definições e propriedades básicas. ~uma etapa seguinte apresentamos o teorema de Cramér-Chernoff para v .a's i. i .d. a valores em R . A seguir enunciamos um princípio de Grandes Desvios para quaisquer abertos e fechados de R. Numa etapa posterior estendemos o teorema de Cramér-Chernoff para v.a's i.i.d. a valores em Rd. No Capítulo final apresentamos, de maneira sintética, outras extensões do teorema de Cramér-Chernoff, tais como: o teorema de Sanov para v .a's i.i .d. e o princípio de Grandes Desvios para Cadeias de Markov finit as. Além disso, apresentamos algumas aplicações de Grandes Desvios em Estatística Matemática. / This thesis is the result of a study about the Large Deviations Theory, in which the context of independent and identically distributed (i.i.d.) random variables was emphasized. Aiming to discuss some of the main results about Large Deviations, first we present the definitions and the basic properties. In a second moment, we present Cramér-Chernoff's theorem for i.i.d random variables in R. Mter, we enunciate a Large Deviations principie for any open and closed sets in R . In another stage, we extend Cramér-Chernoff's theorem for i.i.d. random variables in Rd. In Jast chapter, we present, briefly, other extensions of CramérChernoff's theorem, such as Sanov's theorem for i.i.d. random variables and the Large Deviations principie for finite state Markov Chains. Besides, we present some applications of Large Deviations in Mathematical Statisitics.

La connaissance de l'environnement marin est nécessaire pour un grand nombre d'applications dans le domaine de l'acoustique sous-marine comme la communication, la localisation et détection sonar et la surveillance des mammifères marins. Il constitue le moyen principal pour éviter les interférences néfastes entre le milieu naturel et les actions industriels et militaires conduites en zones côtières.Notre travail de thèse se place dans un contexte de sonar actif avec des fréquences allant de 1 kHz à 10 kHz pour des distances de propagations allant de 1 km à plusieurs dizaines de kilomètres. Nous nous intéressons particulièrement aux environnements de propagation grands fonds, à l'utilisation des antennes industrielles comme les antennes de flancs, les antennes cylindriques et les antennes linéaires remorquées, et à l'utilisation de signaux large bande afin de travailler avec des résolutions en distance et en vitesse très élevées. Le travail de recherche présenté dans ce mémoire est dédié à la recherche de nouveaux paramètres discriminants pour la classification de cible sous-marine en sonar actif et notamment à l'estimation de l'immersion instantanée.Cette étude présente : (1) les calculs de nouvelles bornes de Cramer-Rao pour la position d'une cible en distance en et en profondeur, (2) l'estimation conjointe de la distance et de l'immersion d'une cible à partir de la mesure des temps d'arrivées et des angles d'élévations sur une antenne surfacique et (3) l'estimation conjointe de la distance, de l'immersion et du gisement d'une cible à partir de la mesure des temps d'arrivées et des pseudo-gisements sur une antenne linéaire remorquée.Les méthodes développées lors de cette étude ont été validées sur des simulations, des données expérimentales à petite échelle et des données réelles en mer. / The knowledge of the marine environment is required for many underwater applications such as communications, sonar localization and detection, and marine mammals monitoring. It enables preventing harmful interference between the natural environment and industrial and military actions in coastal areas.This thesis work concentrates upton the context of active sonar with frequencies from 1 kHz to 10 kHz and long propagation ranges from 1 km to several tens of kilometers. We also concentrates upon deep water environment, the use of industrial arrays such as cylindrical arrays, flank arrays and linear towed arrays, and the use of large time-bandwidth signals in order to obtain high distance and speed resolutions. This research work is dedicated to the research of new features for the underwater target classification in active sonar, and specifically to the instantaneous target-depth estimation.This thesis presents: (1) calculations of new Cramer-Rao bounds for the target-position in range and in depth, (2) the joint estimation of the target-depth and the target-range from the arrival time and elevation angle measures with a surface array, (3) the joint estimation of the target-depth, the target-range and the target-bearing from the arrival time and pseudo-bearing angle measures with a linear towed array.The methods presented in this manuscript have been benchmarked on simulation, on reduced-scale experimental data and real marine data.

To date the majority of research in the area of cooperative communications focuses on maximizing throughput and reliability while assuming perfect channel state information (CSI) and synchronization. This thesis, seeks to address performance enhancement and system parameter estimation in cooperative networks while relaxing these idealized assumptions. In Chapter 3 the thesis mainly focuses on training-based channel estimation in multi-relay cooperative networks. Channel estimators that are capable of determining the overall channel gains from source to destination antennas are derived. Next, a new low feedback and low complexity scheme is proposed that allows for the coherent combining of signals from multiple relays. Numerical and simulation results show that the combination of the proposed channel estimators and optimization algorithm result in significant performance gains. As communication systems are greatly affected by synchronization parameters, in Chapter 4 the thesis quantitatively analyzes the effects of timing and frequency offset on the performance of communications systems. The modified Cramer-Rao lower bound (MCRLB) undergoing functional transformation, is derived and applied to determine lower bounds on the estimation of signal pulse amplitude and signal-to-noise ratio (SNR) due to timing offset and frequency offset, respectively. In addition, it is shown that estimation of timing and frequency offset can be decoupled in most practical settings. The distributed nature of cooperative relay networks may result in multiple timing and frequency offsets. Chapters 5 and 6 address multiple timing and frequency offset estimation using periodically inserted training sequences in cooperative networks with maximum frequency reuse, i.e., space-division multiple access (SDMA) networks. New closed-form expressions for the Cramer-Rao lower bound (CRLB) for multiple timing and multiple frequency offset estimation for different cooperative protocols are derived. The CRLBs are then applied in a novel way to formulate training sequence design guidelines and determine the effect of network protocol and topology on synchronization parameter estimation. Next, computationally efficient estimators are proposed. Numerical results show that the proposed estimators outperform existing algorithms and reach or approach the CRLB at mid-to-high SNR. When applied to system compensation, simulation results show that application of the proposed estimators allow for synchronized cooperation amongst the nodes within the network. / Thesis (Ph.D, Electrical & Computer Engineering) -- Queen's University, 2010-07-29 16:52:50.272

La parution en 1750 de l'Introduction à l'analyse des lignes courbes algébriques, unique ouvrage publié de Gabriel Cramer, professeur de mathématiques à l'Académie de Genève, est un jalon important dans l'histoire de la géométrie des courbes et de l'algèbre. Il s'est passé près de dix années entre le moment où le Genevois a écrit les premières lignes de son traité des courbes, à l'automne 1740, et sa publication effective : il s'agit de l'œuvre d'une vie.Ce livre a une histoire, à la fois intellectuelle et matérielle, qui s'inscrit dans les contextes scientifiques, professionnels, académiques et sociaux dans lesquels évoluent son auteur puis ses lecteurs. De la genèse d'un texte manuscrit en devenir dont nous suivrons les évolutionsau cours du processus d'écriture et au gré des événements de la vie de son auteur, aux différentes lectures mathématiciennes et historiennes du texte publié qui en sont faites dans les quelque deux siècles qui ont suivi sa publication, je souhaite ici écrire, pour autant que cette expression ait un sens - ce que je m'attacherai à démontrer - une « biographie » de l'Introduction de Gabriel Cramer / The publication in 1750 of the Introduction à l'analyse des lignes courbes algébriques, the only published work by Gabriel Cramer, professor of mathematics at the Geneva Academy, is an important milestone in the history of geometry of curves and algebra. Almost ten years passed between the time when the Genevan wrote the first lines of his treatise on curves in the autumn of 1740 and its actual publication, making it a lifetime achievement.This book has a history, both intellectual and material, which takes place in the scientific, professional, academic and social contexts in which evolve its author and its readers. From the genesis of a handwritten text as a work in progress of which we will follow the evolutions during the process of writing and according to the events of its author's life, to the various mathematicians and historians' readings of the published text which are made in the two centuries following its publication, I would like to write here, insofar as this expression makes sense - which I shall endeavour to demonstrate - a « biography » of Gabriel Cramer's Introduction

abstract: Signal compressed using classical compression methods can be acquired using brute force (i.e. searching for non-zero entries in component-wise). However, sparse solutions require combinatorial searches of high computations. In this thesis, instead, two Bayesian approaches are considered to recover a sparse vector from underdetermined noisy measurements. The first is constructed using a Bernoulli-Gaussian (BG) prior distribution and is assumed to be the true generative model. The second is constructed using a Gamma-Normal (GN) prior distribution and is, therefore, a different (i.e. misspecified) model. To estimate the posterior distribution for the correctly specified scenario, an algorithm based on generalized approximated message passing (GAMP) is constructed, while an algorithm based on sparse Bayesian learning (SBL) is used for the misspecified scenario. Recovering sparse signal using Bayesian framework is one class of algorithms to solve the sparse problem. All classes of algorithms aim to get around the high computations associated with the combinatorial searches. Compressive sensing (CS) is a widely-used terminology attributed to optimize the sparse problem and its applications. Applications such as magnetic resonance imaging (MRI), image acquisition in radar imaging, and facial recognition. In CS literature, the target vector can be recovered either by optimizing an objective function using point estimation, or recovering a distribution of the sparse vector using Bayesian estimation. Although Bayesian framework provides an extra degree of freedom to assume a distribution that is directly applicable to the problem of interest, it is hard to find a theoretical guarantee of convergence. This limitation has shifted some of researches to use a non-Bayesian framework. This thesis tries to close this gab by proposing a Bayesian framework with a suggested theoretical bound for the assumed, not necessarily correct, distribution. In the simulation study, a general lower Bayesian Cram\'er-Rao bound (BCRB) bound is extracted along with misspecified Bayesian Cram\'er-Rao bound (MBCRB) for GN model. Both bounds are validated using mean square error (MSE) performances of the aforementioned algorithms. Also, a quantification of the performance in terms of gains versus losses is introduced as one main finding of this report. / Dissertation/Thesis / Masters Thesis Computer Engineering 2019

Recently, indoor localization becomes a hot topic no matter in industry or academic field. Smart phones are good candidates for localization since they are carrying various sensors such as GPS, Wi-Fi, accelerometer, barometer and etc, which can be used to estimate the current location. But there are still many challenges for 3D indoor geolocation using smart phones, among which the map selection and 3D performance evaluation problems are the most common and crucial. In the indoor environment, the popular outdoor Google maps cannot be utilized since we need maps showing the layout of every individual floor. Also, layout of different floors differ from one another. Therefore, algorithms are required to detect whether we are inside or outside a building and determine on which floor we are located so that an appropriate map can be selected accordingly. For Wi-Fi based indoor localization, the performance of location estimation is closely related to the algorithms and deployment that we are using. It is difficult to find out a general approach that can be used to evaluate any localization system. On one hand, since the RF signal will suffer extra loss when traveling through the ceilings between floors, its propagation property will be different from the empirical ones and consequently we should design a new propagation model for 3D scenarios. On the other hand, properties of sensors are unique so that corresponding models are required before we analyze the localization scheme. In-depth investigation on the possible hybrid are also needed in case more than one sensor is operated in the localization system. In this thesis, we firstly designed two algorithms to use GPS signal for detecting whether the smart device is operating inside or outside a building, which is called outdoor-indoor transition detection. We also design another algorithm to use barometer data for determining on which floor are we located, which is considered as a multi-floor transition detection. With three scenarios designed inside the Akwater Kent Laboratory building (AK building) at Worcester Polytechnic Institute (WPI), we collected raw data from an Android phone with a version of 4.3 and conducted experimental analysis based on that. An efficient way to quantitatively evaluate the 3D localization systems is using Cramer-Rao Lower Bound (CRLB), which is considered as the lower bound of the estimated error for any localization system. The characteristics of Wi-Fi and barometer signals are explored and proper models are introduced as a foundation. Then we extended the 2D CRLB into a 3D format so that it can fit the our 3D scenarios. A barometer-assisted CRLB is introduced as an improvement for the existing Wi-Fi Receive Signal Strength (RSS)-only scheme and both of the two schemes are compared with the contours in every scenario and the statistical analysis.

The purpose of this thesis is to derive and analyze the theoretical limits for estimatingthe symboltiming delayof Multiple-Input Multiple-Output (MIMO)systems. Two main N X M system models are considered, where N represents the number of transmit antennas and M denotes the number of receive antennas, the 2 X 2 system used by S.-A. Yangand J. Wu and the 4 X 4system used by Y.-C. Wu and E. Serpedin. The second model has been extended to take into account the symbol time-varying fading. The theoretical estimation limits are shown by several bounds: modified Cramer-Rao bound (MCRB), Cramer-Rao bound (CRB) and Barankin bound (BB). BB will be exploited to obtain accurate information regarding the necessary length of data to obtain good estimation. Two scenarios for synchronization are presented: data-aided (DA) and non-data-aided (NDA). Two models for the fading process are considered: block fading and symbol time-varying fading, respectively, the second case being assumed to be Rayleigh distributed. The asymptotic Cramer-Rao bounds for low signal-to-noise ratio (low-SNR) and for high-SNR are derived and the performance of several estimators is presented. The performance variation of bounds and estimators is studied byvarying different parameters, such as the number of antennas, the length of data taken into consideration during the estimation process, the SNR, the oversampling factor, the power and the Doppler frequency shift of the fading.

In certain applications involving direction of arrival (DOA) estimation the operator may have a-priori information on some of the DOAs. This information could refer to a target known to be present at a certain position or to a reflection. In this study, the authors investigate a methodology for array processing that exploits the information on the known DOAs for estimating the unknown DOAs as accurately as possible. Algorithms are presented that can efficiently handle the case of both correlated and uncorrelated sources when the receiver is a uniform linear array. The authors find a major improvement in estimator accuracy in feasible scenarios, and they compare the estimator performance to the corresponding theoretical stochastic Cramer-Rao bounds as well as to the performance of other methods capable of exploiting such prior knowledge. In addition, real data from an ultra-sound array is applied to the investigated estimators. / <p>QC 20130107</p>