M.Sc. (Pharmaceutical Affairs), Faculty of Health Sciences, University of the Witwatersrand, 2009

LAKE SURFACE TEMPERATURE DETERMINED BY SATELLITE IMAGES Abstract The aim of this master thesis is to create a method to determine surface temperature of water in lake Osoyoos (LSWT - Lake Surface Water Temperature). It is necessary to know this temperature in order to make easier decisions in a paradigm of global warming processes. Thanks to LSWT it is possible to make a timeline visualisation of temperature development in years. The first chapter deals with a physical definition of how to get thermal information - the basis of remote thermal sensing is measuring of radiance intensity and it's transformation to brightness temperature. This radiation is not only influenced by atmospheric properties but also by local climatic and meteorological conditions, but it is also influenced by the relief. The thesis seeks to create a method for calculating surface temperature of lake Osoyoos located at Canadian-American border. This calculation is based on data provided by ASTER using a split-window method. The method works with the differences in bands of sensing in order to remove atmospheric influences. This thesis contains several charts, graphs and pictures. The final result of this paper is a map of distribution of lake surface water temperatures. Keywords: Thermal remote sensing, land surface temperature, ASTER, TIR

In this thesis, we examine the performance of different forecasting methods. We use dataof monthly house prices from the larger Stockholm area and the municipality of Uppsalabetween 2005 and early 2019 as the time series to be forecast. Firstly, we compare theperformance of two machine learning methods, the Long Short-Term Memory, and theSupport Vector Machine methods. The two methods forecasts are compared, and themodel with the lowest forecasting error measured by three metrics is chosen to be comparedwith a classic seasonal ARIMA model. We find that the Long Short-Term Memorymethod is the better performing machine learning method for a twelve-month forecast,but that it still does not forecast as well as the ARIMA model for the same forecast period.

The hot box technique is an experimental method to achieve the U-value of elements in stationary conditions; however, it is not always possible to work in stationary conditions in real world. This thesis consisted of evaluating the heat transfer of a window of a historical building with a unique hot box set-up. The window had a low emissivity plastic film to improve thermal efficiency, and the hot box was unique because the outside temperature could not be controlled. The applicability of the hot box technique to dynamic conditions was assessed using COMSOL Multiphysics 5.3. COMSOL Multiphysics is a finite element method solver software with a heat transfer module. Two heat transfer simulations were conducted in 2D based on the indoor and outdoor temperature when the hot box was in operation. First, a stationary study was carried when the outdoor temperature remained stable for 1 day. Then, the study was extended to a transient study to analyze in detail the effect of the external temperature fluctuations for 5 days. The results indicate that a cautious approach should be taken when applying the hot box technique under transient conditions, but that stationary conditions could not be achieved during one day. Nevertheless, the reliability of the simulation solution could have improved more.

Em todo serviço de hemoterapia, a questão da segurança transfusional gera grande preocupação, visto que não existe transfusão isenta de risco de contaminação. Este trabalho tem como objetivo sugerir ao Centro de Hematologia e Hemoterapia do estado de Santa Catarina (HEMOSC), uma proposta de programa de uso de plasma de quarentena visando a segurança transfusional. O plasma de quarentena é aquele originado de um doador de repetição, que apresentou todos os testes sorológicos para doenças hemotransmissíveis não reagentes em doação anterior, até que o doador realize uma nova doação, este hemocomponente é mantido em quarentena, evitando o uso do mesmo em período de janela imunológica. Se todos os testes laboratoriais exigidos pela legislação vigente mantiverem-se não reativos após a nova doação, a unidade de plasma é liberada para uso. Sugere-se adequações no sistema de informática, referentes a triagem clínica e processamento do sangue, especificações para armazenamento e congelamento de plasma. Após análise da produção de plasma da hemorrede catarinense nos anos de 2012 e 2013, foi possível constatar que a instituição apresenta grande potencial para implantação da proposta, sendo necessário adequações na área física dos setores de processamento da hemorrede / Throughout the hemotherapy service, the question of transfusion safety generates great concern, whereas there is no free risk transfusion of contamination. This work aims to suggest to the Center of Hematology of the state of Santa Catarina (Hemosc) a proposal for use of quarantine plasma program to transfusion safety. The quarantine plasma that is originated from a donor replay, which showed all the non-reactive serologic tests for diseases transmissible by blood previous donation until the donor make a new donation, the blood component is kept in quarantine, avoiding the use of it in window periods. If all laboratory tests required by law are kept non-reactive after the new donation, the plasma unit is released for use. To implement the program for this product, adjustments in the computer system, related to recruitment, clinical screening, processing of blood, and specifications for storage and freezing of plasma is needed. After analysis of plasma production in Santa Catarina in 2012 and 2013, we acknowledge that the institution has a great potential for deployment of the proposal, being necessary adjustments in the physical area of the sectors of processing of the HEMOSC

by Raymond Cheuk-kuen Chen. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1994. / Includes bibliographical references (leaves 157-160). / Chapter 1. --- INTRODUCTION --- p.8 / Chapter 1.1. --- Windowing System --- p.8 / Chapter 1.2. --- Fonts --- p.10 / Chapter 1.2.1. --- Bitmap Fonts --- p.11 / Chapter 1.2.2. --- Outline Fonts --- p.12 / Chapter 1.3. --- Different font support models --- p.15 / Chapter 1.3.1. --- Supported by applications --- p.15 / Chapter 1.3.2. --- Supported by windowing system --- p.17 / Chapter 1.3.'3. --- Supported by a dedicated server --- p.19 / Chapter 1.4. --- Issues of Chinese Font Support --- p.20 / Chapter 2. --- OVERVIEW OF X WINDOW SYSTEM --- p.22 / Chapter 2.1. --- Introduction --- p.22 / Chapter 2.2. --- Architecture --- p.23 / Chapter 2.3. --- Font Management in the X Window System --- p.23 / Chapter 2.3.1. --- Before X Version 11 Release5 --- p.24 / Chapter 2.3.2. --- In X Version 11 Release5 --- p.25 / Chapter 2.3.3. --- Portable Compiled Format --- p.25 / Chapter 2.3.4. --- Font Server --- p.26 / Chapter 2.3.5. --- Font Management Library --- p.28 / Chapter 2.4. --- Internal Code --- p.29 / Chapter 3. --- CHINESE FONT SERVER --- p.30 / Chapter 3.1. --- Motivation --- p.30 / Chapter 3.2. --- Font Server Architecture --- p.31 / Chapter 3.2.1. --- Device Independent Font Server layer(DIFS) --- p.32 / Chapter 3.2.2. --- Operating System layer(OS) --- p.32 / Chapter 3.2.3. --- Font Management Library(FML) --- p.33 / Chapter 3.2.4. --- Font Path Element --- p.34 / Chapter 3.2.5. --- Font File Renderer --- p.35 / Chapter 3.2.6. --- Font server Renderer --- p.36 / Chapter 3.3. --- Implementation of Chinese Font Server --- p.36 / Chapter 3.3.1. --- Font data and code set --- p.36 / Chapter 3.3.2. --- Registering a new font reader --- p.38 / Chapter 3.3.3. --- Font specific functions --- p.42 / Chapter 3.3.4. --- Load-All Scheme --- p.43 / Chapter 3.3.5. --- Demand-Loading Scheme --- p.44 / Chapter 3.3.6. --- Embedding of font rasterizer --- p.44 / Chapter 3.4. --- Test Results --- p.45 / Chapter 3.4.1. --- X Application Tests --- p.45 / Chapter 3.4.2. --- Demand-Loading Test --- p.49 / Chapter 3.5. --- Some Remarks --- p.53 / Chapter 4. --- OVERVIEW OF PRINTING SYSTEM --- p.54 / Chapter 4.1. --- Motivation --- p.54 / Chapter 4.2. --- Design Considerations --- p.56 / Chapter 4.2.1. --- Modification of the X server --- p.56 / Chapter 4.2.2. --- Embed the printing system into the font server --- p.57 / Chapter 4.2.3. --- Distributed Architecture --- p.58 / Chapter 4.3. --- System Architecture --- p.60 / Chapter 4.4. --- Printer Server --- p.61 / Chapter 4.5. --- Font Server --- p.63 / Chapter 4.6. --- Printing Services Protocols --- p.63 / Chapter 4.7. --- X Window System Server --- p.65 / Chapter 4.8. --- Printer Server Library --- p.65 / Chapter 4.9. --- Client Applications --- p.65 / Chapter 5. --- DESIGN AND IMPLEMENTATION OF A PRINTER SERVER --- p.67 / Chapter 5.1. --- Objects identification --- p.67 / Chapter 5.1.1. --- Dispatcher (dispatcher) --- p.68 / Chapter 5.1.2. --- Communication Channel (ComChannel) --- p.68 / Chapter 5.1.3. --- Font Cache Manager (FnCache) --- p.69 / Chapter 5.1.4. --- PrnFont (PrnFont) --- p.69 / Chapter 5.1.5. --- Per-Font Cache (CacheStruct) 一- --- p.70 / Chapter 5.1.6. --- Font Server (FnServer) --- p.71 / Chapter 5.1.7. --- Client Manager (LRUList) --- p.71 / Chapter 5.1.8. --- Client Record (ClientRec) --- p.71 / Chapter 5.1.9. --- Printer Driver (PrnDriver) --- p.71 / Chapter 5.1.10. --- Down Loaded Font Table (DownLoadedFont) --- p.72 / Chapter 5.1.11. --- Request Header (reqHeader) --- p.72 / Chapter 5.1.12. --- Generic Reply(replyGeneric) --- p.74 / Chapter 5.2. --- Objects Organization --- p.74 / Chapter 5.2.1. --- Server Control Subsystem --- p.75 / Chapter 5.2.2. --- Client Management Subsystem --- p.78 / Chapter 5.2.3. --- Request Handling Subsystem --- p.84 / Chapter 5.2.4. --- Font Managing Subsystem --- p.86 / Chapter 6. --- SAMPLE PRINTER DRIVER --- p.94 / Chapter 6.1. --- Printer Control Languages --- p.94 / Chapter 6.1.1. --- Structure of PCL Command --- p.95 / Chapter 6.1.2. --- PCL Command Example --- p.97 / Chapter 6.2. --- Printer Font Resources --- p.98 / Chapter 6.3. --- Traditional Font Handling Methods in a Printer Driver --- p.99 / Chapter 6.4. --- Soft Font Creation in PCL Printer --- p.101 / Chapter 6.4.1. --- Font ID number --- p.102 / Chapter 6.4.2. --- Font Descriptor --- p.102 / Chapter 6.4.3. --- Character Code - --- p.104 / Chapter 6.4.4. --- Character Descriptor --- p.105 / Chapter 6.4.5. --- Character Bitmap Data --- p.107 / Chapter 6.5. --- New font downloading schemes for double-byte fonts --- p.107 / Chapter 6.5.1. --- Terminology --- p.108 / Chapter 6.5.2. --- Underlying Concepts of Algorithm One --- p.109 / Chapter 6.5.3. --- Algorithm One --- p.111 / Chapter 6.5.3.1. --- Code Mapping --- p.112 / Chapter 6.5.3.2. --- Example --- p.114 / Chapter 6.5.3.3. --- Memory Consideration --- p.115 / Chapter 6.5.4. --- Algorithm Two --- p.117 / Chapter 7. --- EXPERIMENT RESULTS AND DISCUSSIONS --- p.121 / Chapter 7.1. --- Cache Test --- p.121 / Chapter 7.2. --- Printer Driver Test --- p.125 / Chapter 7.2.1. --- Testing with 10 points font --- p.126 / Chapter 7.2.2. --- Testing with 12 points font --- p.129 / Chapter 7.2.3. --- Testing with 15 points font --- p.131 / Chapter 7.2.4. --- Testing with 18 points font --- p.134 / Chapter 7.3. --- Time Measurement --- p.136 / Chapter 7.4. --- Discussion --- p.139 / Chapter 7.5. --- Further Improvement --- p.143 / Chapter 8. --- CONCLUSIONS --- p.145 / APPENDIX A. PRINTER DRIVER CLASS --- p.147 / APPENDIX B. SAMPLE OUTPUT --- p.149 / REFERENCES --- p.157

"This thesis describes the development of an optimized software-defined-radio implementation of a distributed beamforming system and presents experimental results for two-source and three- source wired-channel and acoustic-channel distributed beamforming using the time-slotted round-trip carrier synchronization protocol. The frequency and phase synthesizer used in this system is based on an optimized ``hybrid' phase locked loop (PLL) with averaging window which is shown to have high frequency estimation accuracy and consistency. For the wired-channel experiments, each source node was implemented by a TMS320C6713DSK while for the acoustic experiments, each source node in the system was built using commercial off-the-shelf parts including TMS320C6713DSK, microphone, speaker, audio amplifier, and battery. The source node functionality including phase locked loops and the logic associated with the time-slotted round-trip carrier synchronization protocol was realized through real-time software independently running on each source node's C6713 digital signal processor. Experimental results for two-source and three-source realizations of the wired-channel and acoustic-channel distributed beamforming system are presented. The results show that near-ideal beamforming performance can be consistently achieved at acoustic wavelengths equivalent to common radio frequency wavelengths."

IEEE 802.11ah is designed to support battery powered devices that are required to serve for several years in the Internet of Things networks. The Restricted Access Window (RAW) has been introduced in IEEE 802.11ah to address the scalability of thousands of densely deployed devices. As the RAW sizes entail the consumed energy to support the transmitting devices in the network, hence the control mechanism for RAW should be carefully devised for improving the overall energy e ciency of IEEE 802.11ah. This thesis presents a two-stage adaptive RAW scheme for IEEE 802.11ah to optimise the energy efficiency of massive channel access and transmission in the uplink communications for highly dense networks. The proposed scheme adaptively controls the RAW sizes and device transmission access by taking into account the number of devices per RAW, retransmission mechanism, harvested-energy and prioritised access. The scheme has four completely novel control blocks: RAW size control that adaptively adjusts the RAW sizes according to different number of devices and application types in the networks. RAW retransmission control that improves the channel utilisation by retransmitting the collided packets at the subsequent slot in the same RAW. Harvested-energy powered access control that adjusts the RAW sizes with the consideration of the uncertain amount of harvested-energy in each device and channel conditions. Priority-aware channel access control that reduces the collisions of high-priority packets in the time-critical networks. The performance of the proposed controls is evaluated in Matlab under different net work scenarios. Simulation results show that the proposed controls improve the network performances in terms of energy efficiency, packet delivery ratio and delay as compared to the existing window control.

A umidade é um dos principais agentes causadores de manifestações patológicas no ambiente construído, e, a chuva dirigida, é uma das principais fontes de umidade que afetam a durabilidade das fachadas das edificações. Desta forma, métodos vêm sendo desenvolvidos como forma de avaliação, entre eles se encontram os experimentais que buscam avaliar este fenômeno a partir da realização de ensaios. A pesquisa é motivada pela constatação da quase inexistência de estudos experimentais de quantificação de chuva dirigida, bem como pela identificação da problemática relacionada ao manchamento e infiltração de água na área próxima às janelas. Sendo assim, o objetivo principal desta pesquisa é identificar, com base em testes de chuva dirigida, a influência das características de projeção, com variações de 50 mm e 120 mm, de inclinação, com valores de 6% e 20%, e engaste lateral do peitoril, sem engaste e com engates de 100 mm, no molhamento da área de fachada localizada próxima às janelas. Este trabalho foi desenvolvido a partir da avaliação de nove modelos com dimensões 1,20m x 2,40m, sendo um modelo de referência sem peitoril, e outros oito englobando as diferentes configurações de projeção, inclinação e engastes laterais dos peitoris, em granito. As técnicas empregadas nas avaliações dos modelos foram a coleta de água da chuva nos paramentos dos modelos, registros de imagens termográficas por infravermelho, chuva colorida para identificação dos fluxos e registros fotográficos digitais A chuva simulada apresentava vazão total de 6 l/m, com diâmetros de gota que variavam de 0,86 mm – 1,45 mm com interceptação de velocidade de vento de 3,2 m/s, e, os testes foram desenvolvidos em quatro ciclos de 15 minutos de duração cada. Os resultados demonstraram que a utilização de prolongamentos laterais nos peitoris promoveram melhor desempenho dos modelos com relação ao descolamento da água da chuva das paredes, quando em comparação àqueles que não possuem esta configuração. Considerando a inclinação dos peitoris identificou-se que o valor de 6% permite, mesmo com a existência de um lacrimal na face inferior do peitoril, a percolação da água da chuva pela porção inferior deste elemento, com um gotejamento vertical. Nas configurações com 20% de inclinação do peitoril não se verifica este comportamento, contudo, o gotejamento ocorre no sentido da parede do modelo, direcionando parte do fluxo da chuva para a área localizada abaixo do peitoril, condicionando valores mais elevados de coleta de chuva. A projeção de 50 mm protege de forma menos eficiente as paredes dos modelos do molhamento pela chuva dirigida, gerando áreas protegidas menores e maiores valores de água coletada quando comparadas à projeção de 120 mm. / Moisture is one of the leading causes of pathological manifestations in buildings. Wind driven rain, is one of the main sources that affect the buildings facades durability. Thus, with the mean to evaluate this phenomenon, some experimental methods has been evaluated. The inexistence of experimental studies for wind driven rain quantification, motivated this research. Besides, this research aims to identify the staining and water seepage in the windows area problems. Thus, the main objective of this research is to identify, based on wind driven rain tests, the influence of geometrical characteristics of sills, in the wetness area located near windows. For this, there were developed sills with projection variations of 50 mm and 120 mm, inclinations of 6 % and 20 %, and sill without side extensions of window and 100 mm extensions. The experimental program contemplate the evaluation of nine wall models, with dimensions 1.20 m x 2.40 m, where a window were fixed and granite sills were installed, with the geometrical variations proposed. Simulated driving rain had total flow 6 l/m, with drop diameters ranging from 0.86 mm - 1.45 mm, with wind speed intercept of 3.2 m /s, and the tests were carried out in four cycles of 15 minutes each. There were collected rainwater on some controlled points of the walls models, besides infrared thermography images, colorful rain for identification of flows and digital photography. With the final results it was possible to conclude that the use of lateral extensions in windows sills promote better performance in the models, in relation to detachment of rainwater to the walls, when compared to sills without lateral extension. About sills inclination, it was found that the 6 % value results in a better behavior compared with 20% inclination. Finally, the windows sills with 120 mm protects more efficiently the models walls.

Objetivos: Descrever as dimensões e a posição da crista fenestra no interior do nicho ósseo da janela redonda, assim como identificar sua presença através da tomografia computadorizada (TC) de alta resolução. Métodos: Uma amostra de 10 ossos temporais humanos adultos foi dissecada para o estudo microscópico do nicho ósseo da janela redonda. Resultados: Em todos os ossos estudados, a membrana da janela redonda foi totalmente visualizada somente após a remoção dos bordos do nicho ósseo. A crista fenestra revelou-se uma afilada saliência óssea situada sempre nos bordos anterior e inferior do nicho da janela redonda. Sua área variou entre 0,28 e 0,80 mm2 (média: 0,51 mm2). A proporção da área ocupada pela crista fenestra no lúmen da janela redonda variou entre 23 e 50% (média: 36%). Foi observada uma tendência de que nichos estreitos alberguem cristas de menores dimensões (coeficiente de Spearman: 0,491). Em todos os casos, a avaliação radiológica não foi capaz de definir a presença da crista fenestra. Conclusões: A crista fenestra ocupa uma variável porém significativa área no lúmen do nicho ósseo da janela redonda. Sua presença representa um expressivo obstáculo para o acesso à rampa timpânica. A TC de alta resolução não acrescenta informações pré-operatórias relevantes acerca da sua presença e das suas dimensões. / Objectives: To describe the dimensions and position of the crista fenestra within the round window niche and determine its presence by means of highresolution computed tomography (CT). Methods: A series of 10 adult human temporal bones were dissected for microscopic study of the round window niche. Results: In all specimens, the round window membrane was fully visualized only after complete removal of its bony overhangs. The crista fenestra was identified as a sharp bony crest located in the anterior and inferior borders of the niche, and its area ranged from 0.28 to 0.80 mm2 (mean, 0.51 mm2). The proportion of the area occupied by the crista fenestra in the whole circumference of the round window ranged from 23 to 50% (mean, 36%). Narrower niches tended to have smaller crests (Spearman coefficient: 0.491). In all cases, radiological assessment was unable to determine the presence of the crista fenestra. Conclusion: The crista fenestra occupies a variable but expressive area within the bony round window niche. Its presence is an important obstacle to adequate access to the scala tympani. High-resolution CT scan provides no additional preoperative information with regard to the presence of the crista fenestra or its linear dimensions.