[pt] A ocorrência de erros em surto é observada principalmente em canais sem fio. Para a análise e melhor entendimento deste tipo de erro, a fim de se melhorar os projetos de sistemas de comunicações digitais, uma modelagem mais precisa, de canais com esta característica, torna-se necessária. Uma diversidade de métodos de estimação de parâmetros tem sido estudada, principalmente aquelas baseadas nos Modelos Escondidos de Markov (HMM do ingês). Em geral cada método é focado em um sistema de comunicações específico, sobre uma camada específica. Neste trabalho é proposto um novo método baseado em um HMM com uma estrutura particular, que permite a dedução de expressões analíticas para todas as estatísticas de interesse. A estrutura do modelo proposto permite a geração de eventos que ocorrem numa sequência binária de dados sujeita a surtos de erro, de acordo com a definição de surtos de erro do CCITT. O modelo proposto possui um número fixo de apenas sete parâmetros, mas o seu número de estados cresce com um de seus parâmetros, que aumenta a precisão, mas não a complexidade. Este trabalho adotou técnicas de otimização, associadas aos métodos de Máxima Verossimilhança e Particle Swarm Optimization (PSO) a fim de realizar a estimação dos parâmetros do modelo proposto. Os resultados demonstram que o modelo proposto permite a caracterização precisa de canais com memória de diversas origens. / [en] The occurrence of error busts is mainly observed in wireless channels. For analysis and a better understanding of such errors, in order to improve the design of communication systems, an accurate modeling of channels with this characteristic is necessary. A lot of parameter estimation methods have been studied, mainly the ones based on Hidden Markov Models (HMM). In general each method is focused in a specific communication system, on a specific layer. In this work it is proposed a new method based on a HMM with particular structure that allows the deduction of analytical expressions for all statistics of interest. The structure of the proposed model permits the generation of events that occur in a binary data sequence subject to bursts of error concerning CCITT error burst definition. The proposed model has a fixed number of only seven parameters but its number of states increase with one of those parameters that increase the accuracy but not the complexity. This work adopted techniques of optimization associated to Maximum Likelihood (ML) and Particle Swarm Optimization (PSO) to perform the parameter estimation to the proposed model. The results show that the proposed model achieves accurate characterization of channels with memory from many different sources.

Made available in DSpace on 2014-06-11T19:25:37Z (GMT). No. of bitstreams: 0 Previous issue date: 2010-12-14Bitstream added on 2014-06-13T19:12:34Z : No. of bitstreams: 1 soeiro_cs_me_araca.pdf: 414870 bytes, checksum: 7e39d30ed89b5dd3fa716cdd85d43676 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / O metabolismo oxidativo e apoptose dos neutrófilos de pacientes humanos nefropatas e sua relação com as toxinas urêmicas tem sido, nos últimos anos, amplamente investigados devido sua importância como elemento imunossupressor. Recentemente surgiram evidências de que a uremia causa disfunção neutrofílica em cães nefropatas, porém não se sabe se o acúmulo de compostos nitrogenados, que também ocorrem nas azotemias não renais, igualmente afeta a função dos neutrófilos. O objetivo do presente estudo foi testar ex vivo a hipótese de que plasmas urêmicos e azotêmicos igualmente afetam o metabolismo oxidativo e a apoptose dos neutrófilos de cães. Para tal, neutrófilos de cães sadios foram isolados e incubados com plasma autólogo, plasma de cão azotêmico e urêmico. A produção de superóxido, com e sem o estimulo com PMA, foi estimada pelo método de redução do nitroazul tetrazólio (NBT) e por citometria de fluxo capilar utilizando-se a sonda hidroetidina (HE). A taxa de neutrófilos viáveis, em apoptose inicial e final, foi quantificada por citometria utilizando-se Anexina V-PE e o índice apoptótico mensurado pelo método morfométrico. A produção de superóxido gerada pelos neutrófilos isolados, em ambos os tratamentos (plasma urêmico e azotêmico) apresentou significante redução (p<0,05). Já a apoptose dos neutrófilos de cães sadios foi acelerada, quando incubados com plasmas urêmico e azotêmico. Pode-se concluir que os componentes presentes nos plasmas urêmicos e azotêmicos alteram ex vivo o metabolismo oxidativo e a apoptose dos neutrófilos, fortalecendo a hipótese de que in vivo ambas condições podem comprometer a imunidade inata de cães / The oxidative metabolism and apoptosis of neutrophils from human patients with nephropathy and its relation with uremic toxins has been widely investigated in the last years because of its importance as an immunosuppressive element. Recently evidences suggests that uremia causes neutrophil dysfunction in dogs with renal disease, but it is unclear whether the accumulation of nitrogen compounds, which also occur in non-renal azotemia, can as well affect the role of neutrophils. The objective of this study was to test ex vivo the hypothesis that uremic and azotemic plasma also affects the oxidative metabolism and apoptosis of neutrophils in dogs. To this end, neutrophils from healthy dogs were isolated and incubated with autologous plasma, plasma of azotemic and uremic dog. The production of superoxide, with and without PMA stimulation was estimated by the method of nitroblue tetrazolium reduction (NBT) and by capillary flow cytometry using the hydroethidine probe (HE). The rate of viable, in early and late apoptosis neutrophils was quantified by flow cytometry using Annexin V-PE and theapoptotic index was measured by morphometric method. The production of superoxide generated by isolated neutrophils in both treatments (azotemic and uremic plasma) showed a significant reduction (p <0.05). Neutrophil apoptosis of healthy dogs was accelerated when incubated with uremic and azotemic plasma. In conclusion, the components present in uremic and azotemic plasma change ex vivo the oxidative metabolism and apoptosis of neutrophils, emphasizing the hypothesis that in vivo both conditions can compromise the innate immunity of dogs

abstract: Currently, educational games are designed with the educational content as the primary factor driving the design of the game. While this may seem to be the optimal approach, this design paradigm causes multiple issues. For one, the games themselves are often not engaging as game design principles were put aside in favor of increasing the educational value of the game. The other issue is that the code base of the game is mostly or completely unusable for any other games as the game mechanics are too strongly connected to the educational content being taught. This means that the mechanics are impossible to reuse in future projects without major revisions, and starting over is often more time and cost efficient. This thesis presents the Content Agnostic Game Engineering (CAGE) model for designing educational games. CAGE is a way to separate the educational content from the game mechanics without compromising the educational value of the game. This is done by designing mechanics that can have multiple educational contents layered on top of them which can be switched out at any time. CAGE allows games to be designed with a game design first approach which allows them to maintain higher engagement levels. In addition, since the mechanics are not tied to the educational content several different educational topics can reuse the same set of mechanics without requiring major revisions to the existing code. Results show that CAGE greatly reduces the amount of code needed to make additional versions of educational games, and speeds up the development process. The CAGE model is also shown to not induce high levels of cognitive load, allowing for more in depth topic work than was attempted in this thesis. However, engagement was low and switching the active content does interrupt the game flow considerably. Altering the difficulty of the game in real time in response to the affective state of the player was not shown to increase engagement. Potential causes of the issues with CAGE games and potential fixes are discussed. / Dissertation/Thesis / Doctoral Dissertation Engineering 2017

O conhecimento acerca das vocalizações dos odontocetos é extremamente heterogêneo e concentra-se principalmente nas espécies mantidas em cativeiro. Neste estudo, descrevo quantitativamente parte do repertório vocal dos golfinhos-rotadores, Stenella longirostris, que vivem ao redor do Arquipélago de Fernando de Noronha, subdividido nos seguintes elementos: assobios, grasnidos emitidos na faixa audível e grasnidos emitidos pelos machos durante a corte. Além disso, avaliei a influência tantos dos diferentes sistemas de gravação e análise de dados, quanto da escolha, interpretação e extração de variáveis por diferentes pesquisadores na variabilidade dos assobios e grasnidos emitidos por golfinhosrotadores. Para avaliar a variabilidade dos assobios desta população de golfinhos-rotadores, uma série de análises quantitativas foi conduzida. Nove variáveis foram extraídas de cada contorno de assobio selecionado aleatoriamente. Os valores médios observados foram muito similares àqueles reportados por OSWALD et al. (2003) para golfinhos-rotadores do Oceano Pacífico Tropical Oriental e apresentaram diferença significativa apenas em relação a freqüência máxima dos assobios. Com relação aos grasnidos, amostras foram selecionadas aleatoriamente e destas foram extraídas oito variáveis. Uma série de análises quantitativas foram então conduzidas para avaliar a variabilidade destas emissões. Os valores obtidos para cada uma das variáveis diferiram dos valores previamente reportados para a espécie, mas as razões para esta variação são ainda desconhecidas. Na busca de indícios sobre a função dos grasnidos no comportamento reprodutivo, eu dividi os grasnidos gravados em duas categorias: corte e não-corte. Grasnidos de corte foram aqueles sons gravados na presença de grupos de golfinhos compostos por uma ou duas fêmeas e por vários machos, engajados em comportamento de corte. Grasnidos de não-corte eram todos aqueles gravados em outros comportamentos ou situações. Os resultados obtidos mostraram diferenças significativas entre as duas categorias tanto nas análises univariadas, quanto na Análise Discriminante (AD), pelo menos quanto à duração e ao número de pulsos. Além disso, a AD mostrou que 98% das amostras foram corretamente classificadas nessas categorias. Estes resultados reforçam a idéia de que os grasnidos são usados em contextos sociais, especialmente em situações aflitivas e exaltam a importância da identificação do emissor dos sinais, assim como o uso de equipamentos de gravação de amplo espectro de freqüência. Por fim, realizei uma série de testes para avaliar a metodologia de gravação e análise de sons. Para isso, gravei as emissões utilizando dois sistemas distintos de gravação, ambos com taxa de amostragem de 48 kHz. Além disso, dois pesquisadores diferentes extraíram as variáveis de um mesmo grupo de amostras e os resultados obtidos para cada pesquisador, bem como para cada sistema foram comparados. Os resultados mostraram que, para os assobios, houve diferenças significativas entre as análises de cada pesquisador, quanto à às variáveis freqüência mínima, freqüência final e duração. Ao comparar os resultados obtidos pelos diferentes sistemas de gravação, notei que a duração apresentou diferenças significativas. Quanto aos grasnidos, notei uma diferença significativa no número de amostras de qualidade gravadas por cada um dos sistemas de gravação. Entretanto, as amostras gravadas em ambos os sistemas não apresentaram diferenças significativas em nenhuma das variáveis extraídas, tampouco foram observadas diferenças entre os diferentes pesquisadores. Acredito que o conhecimento gerado aqui contribuirá para uma melhor compreensão dos padrões de variação da espécie e servirá como base para uma série de outros estudos comparativos. / Knowledge on odontocete vocalizations is extremely heterogeneous and centered mainly in those species kept in captivity. In this study, I describe quantitatively part of the vocal repertoire of the spinner dolphin, Stenella longirostris, from Fernando de Noronha Archipelago, off Brazil, subdivided in the following elements: whistles, burst-pulses emitted in the aural range, and burst-pulses emitted by males during courtship. Furthermore, I evaluated the influence of different recording and analysis systems, and the choice, interpretation and extraction of the acoustic variables by different researchers in the variability of whistles and burst-pulses emitted by spinner dolphins. To evaluate the variability of whistles emitted by this population, a series of quantitative analyses was conducted. Nine variables were extracted from each randomly selected whistle contour. The mean values observed were very similar to those reported by OSWALD et al. (2003) for the spinner dolphins from Eastern Tropical Pacific and presented significant differences only in relation to the maximum frequency of the whistles. I extracted eight variables from randomly selected burst-pulsed sound samples. A series of quantitative analyses were conducted to evaluate the variability of these emissions. The values obtained from each of the variables differed from previously data reported for the species, but the reasons to this variation are still unknown. In the search of any signs of burstpulse functions in reproductive behaviors, I divided the recorded burst-pulses in two categories: court and no-court pulses. Court bursts were those sounds recorded in the presence of groups of dolphins composed by one or two females and several males, engaged in courtship behavior. No-court bursts were all those recorded in other behaviors or situations. The results showed significant differences between the two categories in univariate analysis and also in the Discriminant Function Analysis (DFA). The DFA also showed that 98% of the samples were correctly classified in these categories. These results reinforce the idea that burst-pulsed sounds are used in social contexts, specially in afflictive situations, and exalt the importance of identifying the signal emitter, as well as using wide frequency capable recording equipment. Finally, I performed a series of tests to evaluate the methodology of recording and sound analysis. To achieve this, I recorded the sound emissions using two different recording systems, both with sample rate of 48 kHz. Furthermore, two rsearchers extracted the variables of a group of samples, and the results from each researcher and each recording system were compared. Taking into account the whistles, the results showed that there were significant differences between the researchers on duration, minimum and final frequencies recorded. I compared the results obtained by different recording systems and noticed that the sound duration presented significant differences. With respect to burst-pulses, I noticed a significant difference in the number of samples with quality recorded by each of the recording systems. However, the samples recorded in both systems didn\'t show any difference in none of the extracted variables, nor there were observed differences between researchers. I believe that the knowledge produced here will contribute to the comprehension of variation patterns, and will also serve as a basis to further comparative studies.

A key task of managing an inbound call center is in estimating its performance and consequently plan its capacity, which can be considered a complex task since several system variables are stochastic. These issues are highly crucial for certain time-sensitive services, such as emergency call services. Waiting times affect the service quality of call centers in general, but various customers may place different waiting time expectancies depending on the need. Call center managers struggle to find the relationship between these expectations to their strategical, tactical and operational issues. They are assisted by queueing models that approximate the outcome. Simple setups use analytical approximations while a network of multi-skilled agents serving several customer classes is dependent on computer simulations. Regardless of simple or complex setups, models assume that the system components are homogenous, that the components have some parametric distribution, and that they remain the same regardless of the setup. Human resource and marketing research show that such status quo assumptions are not highly reliable. As an example, customer experience is often affected by the skill of the agent, and agents themselves are affected by their workload and duties, which inter alia affect their efficiency. This thesis aim to assist the Swedish emergency call center with a strategical issue, which require detection of some causalities in the set of system components. The overall aim is to design a simulation model, but such model requires a lot of detailed system knowledge, which itself adds to the knowledge gap in the research field. Findings that contribute to the scientific knowledge body include the burst model that addresses some of the non-stationarity of call arrivals, since some rapid rate increments derives from a latent emergency event. Other contributions are the introduction of stochastic agent behavior, which increases the uncertainty in queueing models; and the service time relationship to geographical distance. The latter may involve general evidence on how area-specific understanding and cultural differences affect the quality of service. This is important for organizations that consider off-shoring or outsourcing their call center service. These findings, along with several undiscovered and unknown influencers, are needed in order to design a reliable simulation model. However, the proposed model in this study cannot be rejected, in terms of waiting time replication. This robust model allowed traffic routing strategies to be evaluated and also assisted managers of the emergency call center into a strategical shift in the late 2015. / <p>Vid tidpunkten för framläggningen av avhandlingen var följande delarbeten opublicerade: delarbete 1 och 3 inskickat.</p><p>At the time of the defence the following papers were unpublished: paper 1 and 3 submitted.</p>

Afin d’estimer la capacité de réponse du chêne sessile (Quercus petraea Matt. Liebl.) aux changements climatiques en cours, le potentiel d’adaptation de cette espèce pour le débourrement a été mesuré en populations naturelles. Ces populations sont situées le long d’un gradient altitudinal comprenant 12 populations entre 131 et 1630 m, et d’un gradient latitudinal comprenant 21 populations de l’ensemble de l’aire de répartition. Tout d’abord l’empreinte démographique sur les niveaux de diversité génétique a été estimée sur les populations du gradient altitudinal à partir de marqueurs neutre. Les résultats ont montré que la diversité est maintenue le long du gradient altitudinal grâce notamment à des forts flux de gènes entre populations. La variabilité génétique du débourrement à été mesurée en tests de provenances pour 10 populations du gradient altitudinal. Les résultats ont montré une forte différenciation ainsi qu’une héritabilité élevée du trait. Une variabilité génétique importante est maintenue à l’intérieur des populations et semble indiquer que de multiples pressions de sélection agissent de manière fluctuante et/ou opposée. La diversité de gènes candidats pour le débourrement a été étudiée sur les populations des deux gradients environnementaux. Un niveau de diversité nucléotidique relativement fort et un faible déséquilibre de liaison qui décroit rapidement avec la distance ont été observés. Des signatures de sélections ont été mises en évidence sur un ensemble de gènes candidats. Une étude d’association a été menée entre variabilité du caractère et polymorphisme au sein des gènes candidats sur les populations des deux gradients. Un total de 16 associations significatives a été observé impliquant 10 gènes candidats. Ces résultats suggèrent un potentiel d’adaptation important face aux changements climatiques et offrent des perspectives intéressantes pour la compréhension des processus évolutifs qui régissent l’adaptation du chêne sessile pour le débourrement. / In order to assess the capacity of sessile oak (Quercus petraea Matt. Liebl.) to withstand the ongoing climate changes, we estimated its adaptative potential for bud burst within natural populations. These populations are located along two steep temperature gradients: an altitudinal gradient comprising 12 populations located between 131 m 1630 m, and a latitudinal gradient including 21 populations from the species’ distribution range. First the demographic imprint on the overall genetic diversity was assessed on the altitudinal gradient populations using neutral markers. Results showed that genetic diversity was homogeneously distributed along the gradient and maintained at high altitudes. The genetic variability of bud burst was measured in provenance tests for 10 populations of the altitudinal gradient. We found a high level of genetic differentiation and a high heritability for the trait. A high variability was also observed within populations, indicating that selection pressures may fluctuate in natural conditions. Genetic diversity of candidate genes for bud burst was assessed on populations from both gradients. A high level of nucleotide diversity was observed, and linkage disequilibrium was low. Selective signatures were observed on few candidate genes. An association mapping study was performed between bud burst variability and polymorphism at the candidate genes on populations of both gradients separately. A total of 16 associations involving 10 genes were observed. These results suggest an important adaptive potential of sessile oak for bud burst in the face of climate change and provide interesting perspectives for the comprehension of evolutionary processes controlling bud burst adaptation of sessile oak.

Superbursts (SBs) are large, seemingly spontaneous activity fluctuations often encountered in high density neural networks in vitro. Little effort has been put forth to define and analyze SBs which are paroxysmal bursting discharges. Through qualitative and quantitative means, I have described specific occurrences of superbursting activity. A complex of paroxysmal bursting has been termed a "superburst episode," and each individual SB is a "superburst event" which is comprises a fine burst structure. Quantitative calculations (employing overall spike summations and coefficient of variation (CV) calculations), reveal three distinct phases. Phase 1 is a "build up" phase of increasingly strong, coordinated bursting with an average of a 17.6% ± 13.7 increase in activity from reference. Phase 2, the "paroxysmal" phase, is comprised of massive coordinated bursting with high frequency spike content. Individual spike activity increases by 52.9% ± 14.6. Phase 3 is a "recovery phase" of lower coordination and an average of a 50.1% ± 35.6 decrease in spike production from reference. SBs can be induced and terminated by physical manipulation of the medium. Using a peristaltic pump with a flow rate of 0.4ml/min, superbursting activity ceases approximately 28.3 min after the introduction of flow. Alternatively, upon cessation of medium flow superbursting activity reemerges after approximately 8.5 min. Additionally, this study explored other methods capable of inducing superbursting activity using osmotic shocks. The induction and termination of SBs demonstrates that the cell culture environment plays a major role in generating this phenomenon. The observations that high density multi-layer neuronal networks in culture are more likely to enter paroxysmal bursting also supports the hypothesis that enrichment and depletion layers of metabolites and ionic species are involved in such unusual activity. The dynamic similarity of the SB phenomenon with epileptiform discharges make further quantification on the spike pattern level pertinent and important.

La régulation de la transcription est l‟un des processus cellulaires des plus fondamentaux et constitue la première étape menant à l‟expression protéique. Son altération a des effets sur l‟homéostasie cellulaire et est associée au développement de maladies telles que le cancer. Il est donc crucial de comprendre les règles fondamentales de la fonction cellulaire afin de mieux cibler les traitements pour les maladies. La transcription d‟un gène peut se produire selon l‟un des deux modes fondamentaux de transcription : en continu ou en burst. Le premier est décrit comme un processus aléatoire et stochastique qui suit une distribution de Poisson. À chaque initiation de la transcription, indépendante de la précédente, un seul transcrit est produit. L‟expression en burst se produit lorsque le promoteur est activé pour une courte période de temps pendant laquelle plusieurs transcrits naissants sont produits. Apportant la plus grande variabilité au sein d‟une population isogénique, il est représenté par une distribution bimodale, où une sous-population n‟exprime pas le gène en question, alors que le reste de la population l‟exprime fortement. Les gènes des eucaryotes inférieurs sont pour la plupart exprimés de manière continuelle, alors que les gènes des eucaryotes supérieurs le sont plutôt en burst. Le but de ce projet est d‟étudier comment l‟expression des gènes a évolué et si la transcription aléatoire, ou de Poisson, est une propriété des eucaryotes inférieurs et si ces patrons ont changé avec la complexité des organismes et des génomes. Par la technique de smFISH, nous avons étudié de manière systématique quatre gènes évolutivement conservés (mdn1+, PRP8/spp42+, pol1+ et cdc13+) qui sont continuellement transcrits dans la levure S. cerevisiae. Nous avons observé que le mode d‟expression est gène-et-organisme spécifique puisque prp8 est exprimé de manière continuelle dans la levure S. pombe, alors que les autres gènes seraient plutôt exprimés en légers burst. / Regulating transcription is one of the most fundamental cellular processes and the first step of a long cascade of processes leading to protein expression. Altering transcriptional output often has major effects on cellular homeostasis and is associated with many disease phenotypes, such as cancer. Understanding the fundamental rules governing transcription regulation is therefore instrumental in understanding cellular function as well as in finding disease treatments. Transcription of a gene can occur through two fundamental different modes: “continuous” or “bursting”. Continuous transcription is defined as stochastic process where a promoter is always in its “on” state and each initiation event is independent of the previous. Bursting transcription occurs when a promoter is activated for a short time and multiple mRNAs are produced during the “on” state, followed by long periods of transcription inactivity. It leads to greater variability in an isogenic population, as expression is often bimodal as a sub-population does not express a given gene. Bursting expression is frequently observed in higher eukaryotes, while a continuous pattern seems to be common in lower eukaryotes. The goal of this project is to study how gene expression patterns evolved. We investigate whether Poisson-like transcription is a property of lower eukaryotes and whether transcription patterns have changed when organisms and genomes evolved into more complex systems. Using smFISH, we have systematically determined expression patterns of four evolutionarily conserved genes; mdn1+, PRP8/spp42+, pol1+ and cdc13+, previously shown to be continuously expressed in the yeast S. cerevisiae. Expression was studied in the yeast S. pombe, as an example for another lower eukaryote as well as in human cell-lines. We observe that expression patterns are organism-and-gene specific suggesting that expression patterns have evolved to fulfill gene specific functions.

The remarkable growth in the applications of low power wireless networks (LPWNs) in various disciplines such as health-care, wildlife monitoring, unmanned vehicles and the emerging Internet of Things (IoT) brings along various challenges. Such applications demand the transfer of large amounts of data in short durations. Unlike conventional medium access control protocols, which force each competing node to contend for each packet it transmits, bulk data transmission enables a node to exclusively use a channel for transferring a large amount of data in succession. Bulk data transmission is a technique in which a sender node is granted exclusive access of the channel in order to transmit all the packets accumulated in its buffer. However, there are two problems with this strategy: (1) For how long should bulk data transfer last if there are multiple contending nodes? (2) How should this strategy deal with the significant fluctuation in the quality of a low-power wireless link? Understanding link quality fluctuations in a wireless sensor network is useful for various reasons. For example, nodes can determine when and for how long they should transmit packets, so that they can reduce the packet loss rate and the cost of retransmission (delay as well as power consumption). However, the quality of a link depends on many factors, which cannot be known except in a probabilistic sense. In this dissertation, I propose an efficient burst transmission scheme that measures and models the dynamic link quality fluctuations. Introducing a large empirical study at the beginning of this dissertation leads to a good understanding of the effect of external factors such as the environment (indoor,outdoor), Cross Technology Interference (CTI) and mobility of a sender node causing link quality to fluctuate. The analysis and observations of the empirical study establishes the basis on which the model for link quality estimation is built and designed. Here I propose three approaches to deal with different aspects of link quality fluctuation. (i) Offline approach- long-term characteristics: The offline approach models the link quality fluctuations by taking into account a large set of data. To obtain such a data set, experiments were performed on the site under study for several weeks. It was observed that the link quality fluctuates considerably even in static deployment. Understanding the stable durations, good and bad alike contribute to the efficient transmission of packets. I propose two offline approaches: (i) The first uses the conditional probability distribution function of signal-to-noise (SNR) fluctuation to estimate the expected reliable and unreliable period. (ii) The second uses k-mean clustering to characterise the link quality fluctuations into different states where the relationship between the states is defined by transitional probabilities. The advantages of employing an offline approach is (i) availability of sufficient memory, (ii) low computational cost, and (iii) possible use of a complex algorithm. However, these approaches can not deal with short-term link quality fluctuation. (ii) Online approach- short-term characteristics: Unlike the offline approaches, an online approach models the link quality in real time and deals with short-term link quality fluctuation. However, this approach has some limitations, such as (i) limited memory space to store data, (ii) high computational cost, (iii) and employment of a simple algorithm to estimate the burst size. My proposed online approach uses adaptive history array to estimate the duration of good and bad states from the statistics of incoming acknowledgement packets. (iii) Hybrid approach- long-to-short-term characteristics: A hybrid approach combines both offline and online methods. I also take advantage of both offline and online models in my proposed hybrid approach. My aim is to characterise the long-term link quality fluctuation with statistics that are obtained offline and to employ the statistics of received acknowledgement packets in real-time to deal with short-term link quality fluctuations. The online statistics are used to fine-tune and calibrate the offline model. To evaluate the performance of my proposed approaches, I implement them in TinyOS and deploy them on TelosB sensor nodes. Furthermore, the proposed approaches in this thesis are compared with the state-of-the-art approaches. The thesis concludes by showing that my approaches efficiently model the link quality fluctuation and propose correct burst size to achieve high throughput, reduce transmission delay, and power consumption under different channel conditions.

Nonstationary signals are widely observed in many real-world applications, e.g., radar, sonar, radio astronomy, communication, acoustics, and vibration applications. Joint time-frequency (TF) domain representations provide a time-varying spectrum for their analyses, discrimination, and classifications. Nonstationary signals commonly exhibit sparse occupancy in the TF domain. In this dissertation, we incorporate such sparsity to enable robust TF analysis in impaired observing environments. In practice, missing data samples frequently occur during signal reception due to various reasons, e.g., propagation fading, measurement obstruction, removal of impulsive noise or narrowband interference, and intentional undersampling. Missing data samples in the time domain lend themselves to be missing entries in the instantaneous autocorrelation function (IAF) and induce artifacts in the TF representation (TFR). Compared to random missing samples, a more realistic and more challenging problem is the existence of burst missing data samples. Unlike the effects of random missing samples, which cause the artifacts to be uniformly spread over the entire TF domain, the artifacts due to burst missing samples are highly localized around the true instantaneous frequencies, rendering extremely challenging TF analyses for which many existing methods become ineffective. In this dissertation, our objective is to develop novel signal processing techniques that offer effective TF analysis capability in the presence of burst missing samples. We propose two mutually related methods that recover missing entries in the IAF and reconstruct high-fidelity TFRs, which approach full-data results with negligible performance loss. In the first method, an IAF slice corresponding to the time or lag is converted to a Hankel matrix, and its missing entries are recovered via atomic norm minimization. The second method generalizes this approach to reduce the effects of TF crossterms. It considers an IAF patch, which is reformulated as a low-rank block Hankel matrix, and the annihilating filter-based approach is used to interpolate the IAF and recover the missing entries. Both methods are insensitive to signal magnitude differences. Furthermore, we develop a novel machine learning-based approach that offers crossterm-free TFRs with effective autoterm preservation. The superiority and usefulness of the proposed methods are demonstrated using simulated and real-world signals. / Electrical and Computer Engineering