ESTG - Mestrado em Engenharia Eletrotécnica - Telecomunicações
Permanent URI for this collection
Browse
Browsing ESTG - Mestrado em Engenharia Eletrotécnica - Telecomunicações by Title
Now showing 1 - 10 of 41
Results Per Page
Sort Options
- ADDITIVE MANUFACTURING FOR ANTENNA TECHNOLOGIESPublication . Carvalho, Saúl dos Santos; Caldeirinha, Rafael Ferreira da Silva; Reis, João Ricardo VitorinoThis dissertation explores the use of Additive Manufacturing (AM) technology, commonly known as Three-dimensional (3D) printing, for fabricating antennas, focusing on both static 3D and dynamic Four-dimensional (4D) printing techniques. It evaluates the cost-effectiveness and performance of these antennas compared to the ones fabricated using traditional methods (e.g. commercially available antennas). The study begins with the characterisation of dielectric properties of 3D printed materials through various extraction methods, which is crucial for accurate simulation and performance prediction. With this research several antennas were 3D printed and analysed, including pyramidal horn antennas, which were fully dielectric, with metallised polymers, with metal composites, and fully metallic, operating at K-Band. Additive manufacturing techniques, such as Laser Powder Bed Fusion (LPBF), were employed, with findings indicating that the resultant full metal horn antennas offered the best performance in terms of gain, impedance matching, and bandwidth. Additionally, the study examines 3D printed C-Band microstrip rectangular patch antennas. Results show that while 3D printed microstrip patch antennas can achieve reasonable efficiency and gain, they still lag behind traditional laminate-based antennas. A novel aspect of this research is the design of an X-Band pneumatic-deployable 4D petal horn antenna, which demonstrates potential for real-time adaptability through computational fluid dynamics and electromagnetic simulations. The dissertation concludes that 3D printing technology offers significant advantages for antenna fabrication, particularly in rapid prototyping and customisation. The emergence of 4D printing presents new possibilities for dynamic, adaptable antennas, though further research is needed to address material and fabrication challenges to fully harness its potential.
- Asymmetric 3D video coding based on regions of perceptual relevancePublication . Pinto, Luís Filipe Henriques; Assunção, Pedro António Amado deThis dissertation presents a study and experimental research on asymmetric coding of stereoscopic video. A review on 3D technologies, video formats and coding is rst presented and then particular emphasis is given to asymmetric coding of 3D content and performance evaluation methods, based on subjective measures, of methods using asymmetric coding. The research objective was de ned to be an extension of the current concept of asymmetric coding for stereo video. To achieve this objective the rst step consists in de ning regions in the spatial dimension of auxiliary view with di erent perceptual relevance within the stereo pair, which are identi ed by a binary mask. Then these regions are encoded with better quality (lower quantisation) for the most relevant ones and worse quality (higher quantisation) for the those with lower perceptual relevance. The actual estimation of the relevance of a given region is based on a measure of disparity according to the absolute di erence between views. To allow encoding of a stereo sequence using this method, a reference H.264/MVC encoder (JM) has been modi ed to allow additional con guration parameters and inputs. The nal encoder is still standard compliant. In order to show the viability of the method subjective assessment tests were performed over a wide range of objective qualities of the auxiliary view. The results of these tests allow us to prove 3 main goals. First, it is shown that the proposed method can be more e cient than traditional asymmetric coding when encoding stereo video at higher qualities/rates. The method can also be used to extend the threshold at which uniform asymmetric coding methods start to have an impact on the subjective quality perceived by the observers. Finally the issue of eye dominance is addressed. Results from stereo still images displayed over a short period of time showed it has little or no impact on the proposed method.
- Automotive 4D Radar Imaging with Field of ViewPublication . Jesus, Samuel Viegas Reigota de; Caldeirinha, Rafael Ferreira da Silva; Reis, João Ricardo VitorinoRadio Detection And Ranging (Radar) technology has been around for many decades. Initially developed for military purposes, radar technology has now grown into a larger set of applications. It is used in aviation, to detect other aircrafts and give accurate altitude readings, in marine navigation, to prevent collision with other ships and monitor ship movements in busy waters, in weather forecast, to monitor precipitation and wind, in factories, to enable manufacturing robots to perceive their environment and navigate/position themselves according to it, and even in health care, to monitor breathing and blood pressure, to name a few. In recent years, improvements in integrated circuitry and Radio Frequency (RF) electronics have made radar-on-chip possible, significantly reducing production costs and therefore enabling a wider adoption of radar technology, now commonly used in many civil applications. In the automotive industry for instance, Millimeter Wave (mmWave) radar sensors are being implemented in everyday vehicles for purposes such as Collision Warning System (CWS), Blind Spot Detection (BSD), Autonomous Cruise Control (ACC), Parking Assistance Systems (PAS) and many more, thus being integrated in systems such as Advanced Driver Assistance Systems (ADAS) and Autonomous Driving Systems (ADS). These systems comprise a suite of electronic automotive technologies designed to assist drivers in various driving scenarios, enhance safety, and improve the overall driving experience by providing real-time data through a combination of software algorithms and sensors, including cameras, ultrasonic sensors, Light Detection And Ranging (Lidar) and radar. Using real-time data from each component, ADAS and ADS can perceive the environment, plan a response, and control the vehicle’s mechanical systems to execute the intended action. As the market pushes for higher levels of vehicle autonomy, both ADAS and ADS are becoming increasingly more complex and sophisticated, requiring improvements in radar technology to meet the evolving demands of these advanced systems. Therefore, this dissertation presents the work on the development of a 4 Dimensions (4D) Multiple-Input-Multiple-Output (MIMO) automotive imaging radar with a 360◦ Field-of- View (FoV). After a brief description of the proposed architecture, using three mmWave 4D imaging radar sensors (AWR1843BOOST) by Texas Instruments (TI), disposed in a equilateral triangular prism shape, this dissertation presents the first results on the implementation of a real-time and high resolution radar sensor with a 360◦ FoV. To conclude the performance evaluation of the presented 360◦ radar, experimental results obtained in a controlled environment, i.e. inside an anechoic chamber, are presented to validate the proposed concept. A mmWave Compact Antenna Test Range (CATR) anechoic chamber was also designed and built in order to properly calibrate the radar. Successful target detection is achieved by multi-point point-cloud. The proposed radar architecture may be seen as an opportunity to augment and improve ADAS and ADS.
- Avaliação económica do impacto da precisão em previsão de consumosPublication . Ferreira, Tiago Paulo; Neves, Luís Miguel Pires; Sousa, João Miguel Charrua deCom as recentes preocupações em reduzir consumos, emissões poluentes e custos associados à produção de energia elétrica, o papel das empresas operadoras tem-se tornado de extrema importância no ponto de vista da gestão e utilização dos recursos de forma otimizada, para fazer face à crescente procura de energia. Contudo, para que seja possível diminuir custos operacionais é necessário conhecer previamente o consumo de energia elétrica para uma determinada hora, semana ou mês. Com base nos dados obtidos, será possível colocar em funcionamento as unidades de produção de energia mais económicas, tornando clara a importância da utilização de modelos e técnicas que permitam reduzir os erros associados à previsão dos consumos. O principal objetivo deste trabalho passa pelo estudo do impacto económico originado pelos erros de previsão, através da aplicação de métodos e técnicas de previsão de consumos. Estas técnicas permitem reduzir o erro associado à diferença entre a energia prevista e a energia que foi efetivamente consumida. Como retorno irá possibilitar uma melhoria das condições de operacionalidade e funcionamento do sistema elétrico.Foram utilizados três métodos, o primeiro baseado em modelos estatísticos clássicos, a regressão linear múltipla, e dois métodos baseados em inteligência artificial, as redes neuronais artificiais e as máquinas de suporte vetorial. A amostra utilizada é relativa à Subestação de Andrinos, com saída para Pombal, constituída por uma linha de 30kV na zona de Leiria. Os resultados obtidos para uma previsão a uma semana são bastante promissores e apelativos, principalmente ao nível da combinação de métodos de previsão com um erro médio absoluto percentual (MAPE) de 5,95%. Para uma utilização isolada de cada método, os erros MAPE oscilam entre 6,05% e 7,71%. Em termos de avaliação económica, as penalizações variam entre 8667,7€ e 10424€ para cada método isoladamente e de 8532,2€ pela combinação dos vários métodos.
- CLASSIFICAÇÃO DE CENAS ACÚSTICAS EM DISPOSITIVOS COM CONSTRANGIMENTOS COMPUTACIONAIS UTILIZANDO INTELIGÊNCIA ARTIFICIALPublication . Anastácio, Ricardo Ribeiro; Bento, Luís Manuel CondeAo longo desta dissertação é apresentada a pesquisa e desenvolvimento de um modelo neuronal de arquitetura Convolutional Neural Network (CNN) para classificar dez ambientes acústicos distintos, destinado a dispositivos baixa complexidade, i.e. limite 128K em número de parâmetros e máximo de 30 MMAC (milhões de Multiply- ACcumulate (MAC)) por inferência. É utilizada a pesquisa de hiperparâmetros (hypertuning), para conseguir retirar o máximo de informação pertinente dos recursos de áudio e também para otimizar a arquitetura do modelo. São utilizadas técnicas de aumento de dados, suavização de rótulos e paragem de treino antecipada para melhorar a generalização do modelo, melhorando a resposta na presença de novos dados não utilizados para treino. São propostas ainda três abordagens com o objetivo de aumentar o campo de aprendizagem e melhorar a diferenciação entre classes. Estas abordagens combinam métodos de otimização e aprendizagem, como agrupamento de modelos ensemble, separação de classes ou aprendizagem um contra todos (OvA). Foi ainda aplicada a técnica de destilação de conhecimento (KD), que permitiu reduzir a sua complexidade do modelo, esta técnica acabou por funcionar também como regularizador diminuindo o sobreajuste. As abordagens propostas foram validadas através da participação no desafio “Task 1, Low-Complexity Acoustic Scene Classification 2022” proposto pela comunidade internacional DCASE, conseguindo obter o 4.º lugar na classificação de equipas num universo de 19 equipas, e 11.º lugar perante 48 modelos em avaliação. O modelo submetido que obteve melhores resultados é designado AI4EDGE_4 é um ensemble de dez modelos OvA, utilizado como “professor” num processo de destilação de conhecimento num “aluno” de arquitetura TBM2. O modelo base fornecido em DCASE obteve uma exatidão de ACC=44.2% e uma perda de LOSS=1.532, o modelo AI4EDGE_4 obteve melhor desempenho i.e. uma exatidão de AC=51.6% e uma perda de LOSS=1.330.
- Condução autónoma – localização de um veículo com base em odometria, IMU e GPSPublication . Gil, Baptiste Ferreira; Neves, Carlos Fernando Couceiro de Sousa; Castro, Hugo Filipe Costelha deA Condução de um veículo automóvel tem-se tornado cada vez mais automatizada ao longo dos anos. Este processo tem-se vindo a desenvolver com o intuito de facilitar cada vez mais a condução dos veículos, de prevenir acidentes rodoviários ou ainda de ser utilizado em operações militares. Este trabalho tem como principal foco o estudo da aplicação de dois dos principais métodos de cálculo da localização probabilística à localização de um veículo automóvel autónomo: o Filtro Estendido de Kalman e o Filtro de Partículas. Foram implementadas várias versões de ambos os filtros, quer em simulação, quer num veículo real, tendo por base, em termos de sensores, a odometria do robô, uma unidade inercial e um GPS. Este trabalho surge na continuação de vários projetos anteriores, com foco no desenvolvimento de um veículo autónomo de baixo custo, acrescentando-lhe a funcionalidade de se localizar. Neste sentido, o trabalho desenvolvido servirá de base para a navegação autónoma em trabalhos futuros.
- Contributions to HEVC Prediction for Medical Image CompressionPublication . Guarda, André Filipe Rodrigues; Rodrigues, Nuno Miguel Morais; Faria, Sérgio Manuel MacielMedical imaging technology and applications are continuously evolving, dealing with images of increasing spatial and temporal resolutions, which allow easier and more accurate medical diagnosis. However, this increase in resolution demands a growing amount of data to be stored and transmitted. Despite the high coding efficiency achieved by the most recent image and video coding standards in lossy compression, they are not well suited for quality-critical medical image compression where either near-lossless or lossless coding is required. In this dissertation, two different approaches to improve lossless coding of volumetric medical images, such as Magnetic Resonance and Computed Tomography, were studied and implemented using the latest standard High Efficiency Video Encoder (HEVC). In a first approach, the use of geometric transformations to perform inter-slice prediction was investigated. For the second approach, a pixel-wise prediction technique, based on Least-Squares prediction, that exploits inter-slice redundancy was proposed to extend the current HEVC lossless tools. Experimental results show a bitrate reduction between 45% and 49%, when compared with DICOM recommended encoders, and 13.7% when compared with standard HEVC.
- Design and test of RF shielding for beam impedance reduction in CERN acceleratorsPublication . Santos, Miguel Valente dos; Fernandes, Telmo Rui Carvalhinho CunhaCERN, the European Organization for Nuclear Research, is one of the world’s largest laboratories for high-energy particle physics. At CERN, particle accelerators and detectors are used to study the basic constituents of matter. CERN’s flagship accelerator is the Large Hadron Collider (LHC), the most powerful particle accelerator built up to now. However, it would not work without its accelerator injector chain. This chain starts with LINAC4, followed by PSB, PS, and SPS, which inject into the LHC. Relatively new particle accelerators like the LHC were designed from scratch with a limited overall beam coupling impedance budget in mind, but this was not the case for the older accelerators at CERN, designed and built decades ago and originally intended for much lower beam intensities. Currently, one of the limiting factors to increase the beam intensities is instabilities caused by the beam coupling impedance of the older accelerators in the chain, such as the PS. The PS vacuum manifolds were identified as one source of instabilities due to a high beam coupling impedance. The current impedance contribution of the manifolds was simulated and measured. Several possible shielding solutions, aimed at reducing the beam coupling impedance, were studied and simulated, giving promising results. Routine RF measurements of other components of the accelerator were performed, either to complete the accelerator impedance contribution database or to assess whether or not the installed shielding is effective. During one of these measurements, it was found that a spare PS septa magnet had defective RF fingers, and the shielding was not effective.
- Disparity compensation using geometric transformsPublication . Monteiro, Ricardo Jorge Santos; Rodrigues, Nuno Miguel Morais; Faria, Sérgio Manuel Maciel deThis dissertation describes the research and development of some techniques to enhance the disparity compensation in 3D video compression algorithms. Disparity compensation is usually performed using a block matching technique between views, disregarding the various levels of disparity present for objects at different depths in the scene. An alternative coding scheme is proposed, taking advantage of the cameras setup information and the object’s depth in the scene, to compensate more complex spatial distortions, being able to improve disparity compensation even with convergent cameras. In order to perform a more accurate disparity compensation, the reference picture list is enriched with additional geometrically transformed images, for the most relevant object’s levels of depth in the scene, resulting from projections of one view to another. This scheme can be implemented in any state-of-the-art video codec, as H.264/AVC or HEVC, in order to improve the disparity matching accuracy between views. Experimental results, using MV-HEVC extension, show the efficiency of the proposed method for coding stereo video, presenting bitrate savings up to 2.87%, for convergent camera sequences, and 1.52% for parallel camera sequences. Also a method to choose the geometrically transformed inter view reference pictures was developed, in order to reduce unnecessary overhead for unused reference pictures. By selecting and adding to the reference picture list, only the most useful pictures, all results improved, presenting bitrate savings up to 3.06% for convergent camera sequences, and 2% for parallel camera sequences.
- ELECTROMAGNETIC LOG SHIELD FOR ROADS IDE APPLICATIONPublication . Tribovane, Bruno Alves; Caldeirinha, Rafael Ferreira da SilvaIn this dissertation, a highly sustainable barrier made of circular section timber of low industrial processing, is presented. This wooden structure allows the control of electromagnetic (EM) spectrum by fine-tuning the barrier’s intrinsic parameters to shape shielding effectiveness (SE) in order to improve either EM shielding or transparency. An algorithm is developed in MATLAB to design and generate the desired structural geometry within a simulation environment (CST MW), with the implementation of surface roughness and bending on the trunks. Simulations are performed at the frequency range of 0.3 to 10 GHz with the parametrisations such as the spacing between centre of poles, radius and moisture content (MC). The proposed real scale barrier prototype is assembled in the anechoic chamber and measurements are then performed at different transmitter (TX) and receiver (RX) antenna angles with both vertical and horizontal polarisation to better understand this structure’s impact on radio propagation. From this work, it was demonstrated that the wooden barrier’s simulation and measurement results present a good agreement and how the shielding efficiency can be tailored by optimising the radius and spacing at frequencies of commercial interest.