[pt] Este trabalho aborda o problema de compactação de imagens, sem distorção, explorando o algoritmo M-LZA, proposto por Finamore e Nunes [2]. Nomcando de forma diferente os nós da árvore de codificação gerada pelo algoritmo M-LZA [2], o algoritmo V-LZA aqui introduzido busca ser uma alternativa mais eficiente do que seu antecessor com relação ao tamanho em bits, ocupado pela imagem codificada e/ou a rapidez de convergência para uma taxa de compactação mínima. / [en] This work focus on the problem of image compactation in the errorless sense, by means of exploring the M-LZA algorithm proposed by Finamore and Nunes [2]. Naming diferently the nodes of the coding tree fenerated by the M- LZA algorithm, the algorthm V-LZA introduced in this work, seek to became a more efficient alternative than its predecessor concerning the size of the coded image and/or its converging speed to a minimum compactation rate.

Advancement of micro-electro-mechanics and wireless communication have proliferated the deployment of large-scale wireless sensor networks. Due to cost, size and power constraints, at most a few sensor nodes can be equipped with a global positioning system; such nodes (whose positions can be accurately determined) are referred to as anchors. However, one can deter-mine the distance between two nearby sensors using some form of local communication. The problem of computing the positions of the non-anchor nodes from the inter-sensor distances and anchor positions is referred as sensor network localization (SNL). In this dissertation, our aim is to develop an accurate, efficient, and scalable localization algorithm, which can operate both in the presence and absence of anchors. It has been demon-strated in the literature that divide-and-conquer approaches can be used to localize large net-works without compromising the localization accuracy. The core idea with such approaches is to partition the network into overlapping subnetworks, localize each subnetwork using the available distances (and anchor positions), and finally register the subnetworks in a single coordinate system. In this regard, the contributions of this dissertation are as follows: We study the global registration problem and formulate a necessary “rigidity” condition for uniquely recovering the global sensor locations. In particular, we present a method for efficiently testing rigidity, and a heuristic for augmenting the partitioned network to enforce rigidity. We present a mechanism for partitioning the network into smaller subnetworks using cliques. Each clique is efficiently localized using multidimensional scaling. Finally, we use a recently proposed semidefinite program (SDP) to register the localized subnetworks. We develop a scalable ADMM solver for the SDP in question. We present simulation results on random and structured networks to demonstrate the pro-posed methods perform better than state-of-the-art methods in terms of run-time, accuracy, and scalability.

[pt] Este trabalho analisa algumas estratégias para decodificação de sinais TCM. Além disto, propõe-se uma nova estratégia que faz uso de um processamento por percurso sobrevivente na treliça de decodificação para gerar e atualizar, de acordo com um algoritmo do tipo PLL, um conjunto de estimativas de fase, cada uma associada a um estado da treliça, que são utilizadas simultaneamente no processo de decodificação. Expressões úteis para o direcionamento de parâmetros usados no algoritmo de estimação são apresentadas. Estas expressões foram baseadas em uma análise linear simplificada que considerou tanto o desempenho do estimador em estado estacionário quanto no período transiente. Resultados de desempenho obtidos, via simulação, com o uso do método proposto são comparados com resultados obtidos com outras estratégias para recuperação de portadora em sistemas TCM. Estes resultados mostram uma clara superioridade do método proposto para diversos tipos de erros de fase. / [en] This thesis analyses some coherent decoding schemes for TCM signals reception. In addition, a new scheme, which makes use of a per survivor processing, is propoed. In this scheme, each trellis state has an associated phase estimate which is generated by a data-aided PLL-like algorithm, based on the survivor sequence associated to that state. Some useful expressions are developed to help the parameters selection for use on the estimation algorthm. These expressions are based on a simplified linear analysis which are carried out on both transient and stationary cases. We compare the proposed scheme performance to others schemes by means of computer simulation. These results show that the new scheme has better performance in terms of bit-error-rate.

Atualmente, o serviço digital conhecido como Vídeo sob Demanda - Video on Demand (VoD) - está em ascensão e costuma requerer uma quantidade significativa de recursos físicos para a sua implementação. Para reduzir os custos de operacionalização desse serviço, uma das alternativas comumente usada é o emprego de proxies que cacheiam as partes mais importantes do acervo, com o objetivo de atender a demanda para esse conteúdo no lugar do servidor primário do sistema VoD. Nesse contexto, para melhorar a eficiência do proxy, propõe-se neste trabalho um novo algoritmo de cacheamento que explora o posicionamento dos clientes ativos para determinar a densidade de clientes dentro de uma janela de tempo existente em frente de cada trecho de vídeo. Ao cachear os trechos de vídeo com maior densidade em frente a eles, o algoritmo é capaz de alcançar um alto desempenho, em termos de taxa de acertos para as requisições recebidas pelo proxy, durante intervalos de alta carga de trabalho. Para avaliar esta abordagem, o novo algoritmo desenvolvido foi comparado com outros de natureza semelhante, fazendo uso tanto de métricas tradicionais, como a taxa de acertos, como também de métricas físicas, como, por exemplo, o uso de recursos de processamento. Os resultados mostram que o novo algoritmo explora melhor a banda de processamento disponível na arquitetura de base do proxy para obter uma taxa de acertos maior em comparação com os algoritmos usados na análise comparativa. Por fim, para dispor das ferramentas necessárias para construir essa análise, produziu-se uma outra contribuição importante nesse trabalho: a implementação de um simulador de proxies VoD que, até onde se sabe, é o primeiro a possibilitar a avaliação do hardware utilizado para implementar essa aplicação. / Today, Video on Demand (VoD) is a digital service on the rise that requires a lot of resources for its implementation. To reduce the costs of running this service, one of the commonly used alternatives is using proxies that cache the most important portions of the collection in order to meet the demand for this content in place of the primary server of the VoD system. In this context, to improve the efficiency of proxy, we proposed a novel caching algorithm that explores the positioning of the active clients to determine the density of clients inside a time window existing in front of each video chunk. By caching the video chunks with the greater density in front of them, the algorithm is able to achieve high performance, in terms of the hit ratio for the requests received by the proxy, during periods of high workload. To better evaluate our approach, we compare it with others of similar nature, using both traditional metrics like hit rate, as well as physical metrics, such as the use of processing resources. The results show that the new algorithm exploits the processing bandwidth available in the underlying architecture of the proxy for obtaining a larger hit rate in comparison to the other algorithms used in the comparative analysis. Finally, to dispose of the necessary tools to perform this analysis, we produced another important contribution in this work: the implementation of a VoD proxy simulator that, to the best of our knowledge, is the first one to enable the evaluation of the hardware used to implement this application.

The teaching and learning of Euclidean geometry in black South African schools is characterised by teachers' poor knowledge of Euclidean geometry and the use of traditional taaching methods that encourage rote learning. In the light of this, the study investigated the extent to which learners perform in Process-Based Instruction. Process-Based Instruction is charactarised by the gradual transfer of instruction from the teacher to the learner. In Process-Based Instruction learners are expected to communicate thought,form concepts and master theorems by drawing, 'showed with and used hands', talking and writing. Learners' performance is assessed by rubrics and is analysed using graphs. The findings of the empirical investigation revealed that learners of both grades 8 and 9 cannot communicate thought by any media (talking, writing, 'showed with and used hands' and drawing). A few learners formed a few concepts and mastered theorems. On the basis of the findings and the shortcomings of the study recommendations are made. / Educational studies / M. Ed. (Didactics)

Text in English, abstract in English and Afrikaans / The teaching and learning of Euclidean geometry in black South African schools is characterised by teachers' poor knowledge of Euclidean geometry and the use of traditional teaching methods that encourage rote learning. In the light of this, the study investigated the extent to which learners perform in Process-Based Instruction. Process-Based Instruction is characterised by the gradual transfer of instruction from the teacher to the learner. In Process-Based Instruction learners are expected to communicate thought, form concepts and master theorems by drawing, 'showed with and used hands', talking and writing. Learners' performance is assessed by rubrics and is analysed using graphs. The findings of the empirical investigation revealed that learners of both grades 8 and 9 cannot communicate thought by any media (talking, writing 'showed with and used hands' and drawing). A few learners formed a few concepts and mastered theorems. On the basis of the findings and the shortcomings of the study recommendations are made. / Die onderrig van Euclediaanse Meetkunde in swart Suid-Afrikaanse skole word gekenmerk aan onderwysers se swak kennis van die vak, asook die gebruik van onderrigmetodes wat op die leerder se geheue staatmaak. Hierdie studie ondersoek die sukses van leerders se prestasie in proses-gebasseerde ondenig. Proses-gebaseerde onderig word gekenmerk aan die geleidelike oordra van kennis, van die onderwyser na die leerder. Daar word van leerders verwag om gedagtes oor te dra, konsepte (begrippe) te verstaan en stellings onder die knie te kry deur middel van sketse, die gebruik van gebare, verbale verduideliking en skriftelike werk. Leerders se prestasie word geevalueer deur die gebruik van klassifisering in kolomme (rubrics) en word ge-analiseer deur gralieke. Die bevinding van die empieriese ondersoek het getoon dat leerders van sowel graad 8 as graad 9 nie gedagtes kan oordra deur enige van die gemelde metodes nie. 'n Paar leerders het wel enkele begrippe verstaan en kon stellings bemeester. Na aanleiding van die bevindings van die studie en die tekortkominge wat daardeur opgespoor is, word sekere voorstelle gemaak. / Didactics / M. Ed. (Didactics)

The advance of internet applications, such as video streaming, big data and cloud computing, is reshaping the telecommunication and internet industries. Bandwidth demands in datacentres have been boosted by these emerging data-hungry internet applications. Regarding inter- and intra-datacentre communications, fine-grained data need to be exchanged across a large shared memory space. Large-scale high-speed optical switches tend to use a rearrangeably non-blocking architecture as this limits the number of switching elements required. However, this comes at the expense of requiring more sophisticated route selection within the switch and also some forms of time-slotted protocols. The looping algorithm is the classical routing algorithm to set up paths in rearrangeably non-blocking switches. It was born in the electronic switch era, where all links in the switches are equal. It is, therefore, not able to accommodate loss difference between optical paths due to the different length of waveguides and distinct numbers of crossings, and bends, leading to sub-optimal performance. We, therefore, propose an advanced path-selection algorithm based on the looping algorithm that minimises the path-dependent loss. It explores all possible set-ups for a given connection assignment and selects the optimal one. It guarantees that no individual path would have a sufficiently substantial loss, therefore, improve the overall performance of the switch. The performance of the proposed algorithm has been assessed by modelling switches using the VPI simulator. An 8×8 Clos-tree switch demonstrates a 2.7dB decrease in loss and 1.9dB improvement in IPDR with 1.5 dB penalty for the worst case. An 8×8 dilated Beneš shows more than 4 dB loss reduction for the lossiest path and 1.4 dB IPDR improvement for 1 dB power penalty. The improved algorithm can be run once for each switch design and store its output in a compact lookup table, enabling rapid switch reconfiguration. Microelectromechanical systems (MEMS) based optical switches have been fabricated with over 1,000 ports which meet the port count requirements in data centre networks. However, the reconfiguration speed of the MEMS switches is limited to the millisecond to microsecond timescale, which is not sufficient for packet switching in datacentres. Opto-electronic devices, such as Mach-Zehnder Interferometers (MZIs) and semiconductor optical amplifiers (SOAs) with nanosecond response time show the potential to fulfil the requirements of packet switching. However, the scalability of MZI switches is inherently limited by insertion loss and accumulated crosstalk, while the scalability of SOA switches is restricted by accumulated noise and distortion. We, therefore, have proposed a dilated Beneš hybrid MZI-SOA design, where MZIs are implemented as 1×2 or 2×1 low-loss switching elements, minimising crosstalk by using a single input, and where short SOAs are included as gain or absorption units, offering either loss compensation or crosstalk suppression though adding only minimal noise and distortion. A 4×4 device has been fabricated and exhibits a mere 1.3dB loss, an extinction ratio of 47dB, and more than 13dB IPDR for a 0.5dB power penalty. When operating with 10 Gb/s per port, 6pJ/bit energy consumption is demonstrated, delivering 20% reduced energy consumption compared with SOA-based switches. The tolerance of the current control accuracy of this switch is very broad. Within a 5 mA bias current range, the power penalty can be maintained below 0.2 dB for 8 dB IPDR and 12 mA for 10 dB IPDR with a penalty less 0.5 dB. The excellent crosstalk and power penalty performance demonstrated by this chip enable the scalability of this hybrid approach. The performance of 16×16 port dilated Beneš hybrid switch is experimentally assessed by cascading 4×4 switch chips, demonstrating an IPDR of 15 dB at a 1 dB penalty with a 0.6 dB power penalty floor. In terms of switches with port count larger than 16×16, the power penalty performance has been analysed with physical layer simulations fitted with state-of-the-art data. We assess the feasibility of three potential topologies, with different architectural optimisations: dilated Beneš, Beneš and Clos-Beneš. Quantitative analysis for switches with up to 2048 ports is presented, achieving a 1.15dB penalty for a BER of 10-3, compatible with soft-decision forward error correction.

Made available in DSpace on 2014-06-11T19:34:57Z (GMT). No. of bitstreams: 0 Previous issue date: 2012-03-01Bitstream added on 2014-06-13T19:44:41Z : No. of bitstreams: 1 negrao_cg_dr_guara.pdf: 894711 bytes, checksum: d613ce50ad1ab59115fdbfb212a6e245 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Este trabalho propõe a utilização da técnica de modelagem para obtenção das equações do movimento de sistemas dinâmicos, apresentando uma contribuição para o desenvolvimento do algoritmo computacional e implementado através da técnica de grafos de ligação. O objetivo é entender o desempenho do algoritmo desenvolvido através dos grafos de ligações e, assim, avaliar e validar as propostas deste trabalho. O problema da busca para obter uma equação do modelo de sistemas dinâmicos com geração automática em plantas industriais tem levantado a necessidade de desenvolver um algoritmo para sistemas que forneça a equação do movimento do modelo dinâmico. A proposição desta Tese consiste na aplicação da técnica de grafos de ligação para a determinação das equações de estado a partir do algoritmo desenvolvido com um programa computacional para geração automática das equações do movimento na forma de equações de estado. Os algoritmos foram desenvolvidos em ambiente matemático simbólico. Alguns estudos de caso (aplicações) foram utilizados para validação do algoritmo, constituindo estudo de casos sistemáticos, sendo: -Sistema dinâmico viga-bola, sistema mecânico massa-mola-amortecedor e seu análogo o circuito RLC e sistema servo-válvula eletro-hidráulica. No primeiro caso, sistema dinâmico viga-bola (sistema não-linear), o sistema compreende de uma viga fixada ao eixo do motor e uma bola deslizando sobre a viga, onde foi obtida a modelagem via grafos de ligação, utilizando o algoritmo computacional e encontrado o modelo na forma de equação de estado. No segundo caso, sistema mecânico massa-mola-amortecedor e seu análogo circuito RLC, foi obtido o grafo... / This work proposes the use of the technique of modeling for getting the equations of the movement of dynamic linear and non-linear systems, presenting a contribution for the development of the computational algorithm and implemented through the technique of bond graphs. The objective is the understanding of the performance of the algorithm developed by the bond graphs and thus to evaluate and validate the proposals in this work. The problem of the search to obtain an equation of the model of dynamic systems with automatic generation in industrial plants has been lifting the necessity of developing an algorithm for systems which supply the equation of the movement of the dynamic model. The proposition of this Thesis consists in the application of the bond graphs technique for the determination of the equations of state from the algorithm developed with a computational program for automatic generation of the equations of the movement in the form of equations of state. The algorithms were developed in symbolic mathematical environment. Some case studies (applications) were used for validation of the developed algorithm. The first case is a non-linear system, while the second and third thing cases are linear, constituting systematic case study, being: Dynamic system beam-ball, mechanical system mass-spring-damper and its analogue the RLC circuit and electro-hydraulic servo valve system . In the first case, dynamic system beam-ball (non-linear systems), the system consists of a beam attached to the motor shaft and a ball sliding on the beam, where the modeling was obtained by bond graphs, using the computational algorithm and the model was found in the form of equation of state. In the second case, mechanical system mass-spring-damper and its analogue RLC circuit, the validation was carried out... (Complete abstract click eletronic access below)

abstract: Machine learning technology has made a lot of incredible achievements in recent years. It has rivalled or exceeded human performance in many intellectual tasks including image recognition, face detection and the Go game. Many machine learning algorithms require huge amount of computation such as in multiplication of large matrices. As silicon technology has scaled to sub-14nm regime, simply scaling down the device cannot provide enough speed-up any more. New device technologies and system architectures are needed to improve the computing capacity. Designing specific hardware for machine learning is highly in demand. Efforts need to be made on a joint design and optimization of both hardware and algorithm. For machine learning acceleration, traditional SRAM and DRAM based system suffer from low capacity, high latency, and high standby power. Instead, emerging memories, such as Phase Change Random Access Memory (PRAM), Spin-Transfer Torque Magnetic Random Access Memory (STT-MRAM), and Resistive Random Access Memory (RRAM), are promising candidates providing low standby power, high data density, fast access and excellent scalability. This dissertation proposes a hierarchical memory modeling framework and models PRAM and STT-MRAM in four different levels of abstraction. With the proposed models, various simulations are conducted to investigate the performance, optimization, variability, reliability, and scalability. Emerging memory devices such as RRAM can work as a 2-D crosspoint array to speed up the multiplication and accumulation in machine learning algorithms. This dissertation proposes a new parallel programming scheme to achieve in-memory learning with RRAM crosspoint array. The programming circuitry is designed and simulated in TSMC 65nm technology showing 900X speedup for the dictionary learning task compared to the CPU performance. From the algorithm perspective, inspired by the high accuracy and low power of the brain, this dissertation proposes a bio-plausible feedforward inhibition spiking neural network with Spike-Rate-Dependent-Plasticity (SRDP) learning rule. It achieves more than 95% accuracy on the MNIST dataset, which is comparable to the sparse coding algorithm, but requires far fewer number of computations. The role of inhibition in this network is systematically studied and shown to improve the hardware efficiency in learning. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2017

Orientador: Nazem Nascimento / Banca: Mauro Hugo Matias / Banca: José Geraldo Trani Brandão / Banca: Viktor Pastoukhov / Banca: Luiz Heleno Moreira Duque / Resumo: Este trabalho propõe a utilização da técnica de modelagem para obtenção das equações do movimento de sistemas dinâmicos, apresentando uma contribuição para o desenvolvimento do algoritmo computacional e implementado através da técnica de grafos de ligação. O objetivo é entender o desempenho do algoritmo desenvolvido através dos grafos de ligações e, assim, avaliar e validar as propostas deste trabalho. O problema da busca para obter uma equação do modelo de sistemas dinâmicos com geração automática em plantas industriais tem levantado a necessidade de desenvolver um algoritmo para sistemas que forneça a equação do movimento do modelo dinâmico. A proposição desta Tese consiste na aplicação da técnica de grafos de ligação para a determinação das equações de estado a partir do algoritmo desenvolvido com um programa computacional para geração automática das equações do movimento na forma de equações de estado. Os algoritmos foram desenvolvidos em ambiente matemático simbólico. Alguns estudos de caso (aplicações) foram utilizados para validação do algoritmo, constituindo estudo de casos sistemáticos, sendo: -Sistema dinâmico viga-bola, sistema mecânico massa-mola-amortecedor e seu análogo o circuito RLC e sistema servo-válvula eletro-hidráulica. No primeiro caso, sistema dinâmico viga-bola (sistema não-linear), o sistema compreende de uma viga fixada ao eixo do motor e uma bola deslizando sobre a viga, onde foi obtida a modelagem via grafos de ligação, utilizando o algoritmo computacional e encontrado o modelo na forma de equação de estado. No segundo caso, sistema mecânico massa-mola-amortecedor e seu análogo circuito RLC, foi obtido o grafo... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: This work proposes the use of the technique of modeling for getting the equations of the movement of dynamic linear and non-linear systems, presenting a contribution for the development of the computational algorithm and implemented through the technique of bond graphs. The objective is the understanding of the performance of the algorithm developed by the bond graphs and thus to evaluate and validate the proposals in this work. The problem of the search to obtain an equation of the model of dynamic systems with automatic generation in industrial plants has been lifting the necessity of developing an algorithm for systems which supply the equation of the movement of the dynamic model. The proposition of this Thesis consists in the application of the bond graphs technique for the determination of the equations of state from the algorithm developed with a computational program for automatic generation of the equations of the movement in the form of equations of state. The algorithms were developed in symbolic mathematical environment. Some case studies (applications) were used for validation of the developed algorithm. The first case is a non-linear system, while the second and third thing cases are linear, constituting systematic case study, being: Dynamic system beam-ball, mechanical system mass-spring-damper and its analogue the RLC circuit and electro-hydraulic servo valve system . In the first case, dynamic system beam-ball (non-linear systems), the system consists of a beam attached to the motor shaft and a ball sliding on the beam, where the modeling was obtained by bond graphs, using the computational algorithm and the model was found in the form of equation of state. In the second case, mechanical system mass-spring-damper and its analogue RLC circuit, the validation was carried out... (Complete abstract click eletronic access below) / Doutor