Thesis (PhD (Electrical and Electronic Engineering))--University of Stellenbosch. / ENGLISH ABSTRACT: This dissertation presents a synthesis method for broadband high-power absorptive microwave lters by cascading a wide stop-band wa e-iron lter with an absorptive harmonic pad. The classical image impedance synthesis methods for wa e-iron lters are updated to allow for non-uniform boss patterns, which enable control over both the stop-band attenuation and pass-band re ection of the lter. By optimising an accurate circuit model equivalent, computationally intensive numerical EM optimisation are avoided. The nonuniform wa e-iron lter achieves the same electrical speci cation as similar lters in literature, but in a smaller form factor. The prototype presented displays less than -21 dB in-band re ection over 8.5 - 10.5 GHz, with stop-band attenuation in excess of 50 dB over the harmonic bands 17 - 31.5 GHz and 30 dB over the 34 - 42 GHz. The prototype is designed to handling 4 kW peak power incident in the transmitted band, and is 130 mm in length. Minimal full-wave tuning is required post-synthesis, and good agreement is found between synthesised and measured responses. Additionally, a completely novel oblique wa e-iron boss pattern is proposed. For the absorptive harmonic pad, transversal broadwall slots in rectangular waveguide, coupling to an absorptive auxiliary guide, are investigated in the presence of standing wave surface current distributions. An accurate circuit model description of the cascaded structure is developed, and optimised to provide a required level of input match in the presence of an arbitrary re ective lter. Using numerical port parameter data of the wa eiron lter, a harmonic pad is developed that provides -12.5 dB input re ection match across the band 17 - 21 GHz with up to 1 kW peak incident power handling capability, yet is only 33 mm in length. Again, good agreement is found between synthesised and measured responses of the cascaded structure. / AFRIKAANSE OPSOMMING: Hierdie proefskrif stel 'n sintesetegniek voor vir wyeband, ho edrywing absorberende mikrogol lters deur 'n kaskade kombinasie van 'n verlieslose wafelyster lter met 'n wye stopband, en 'n absorberende harmoniekdemper. Die klassieke sintesemetodes vir wafelyster lters word aangepas om nie-uniforme tandpatrone toe te laat, wat beheer oor beide die lter se stopband attenuasie en deurlaatband weerkaatsing moontlik maak. Deur die optimering van 'n akkurate stroombaanmodel van die lter kan berekeningsintensiewe numeriese EM optimering vermy word. Die nieuniforme wafelyster lter behaal dieselfde spesi kasies as soortgelyke gepubliseerde lters, maar is meer kompak. Die vervaardigingsprototipe handhaaf minder as -21 dB intreeweerkaatsing oor die deurlaatband van 8.5 - 10.5 GHz, asook attenuasie van meer as 50 dB oor die stopband 17 - 31.5 GHz en 30 dB oor 34 - 42 GHz. Die prototipe is ontwerp om 4 kW intree kruindrywing te hanteer in die deurlaatband, en is 130 mm lank. Die metode vereis minimale verstellings tydens volgolf simulasie, en die meetresultate stem goed ooreen met die gesintetiseerde gedrag. 'n Nuwe skuinstandpatroon word ook voorgestel vir wafelyster lters. Vir die harmoniekdemper word transversale bre ewandgleuwe in reghoekige gol eier wat koppel na 'n absorberende newegol eier ondersoek in die teenwoordigheid van staandegolfpatrone in oppervlakstroom. 'n Akkurate stroombaanmodel van 'n gleufkaskade word ontwikkel, en geoptimeer om 'n vereiste intreeweerkaatsing te bewerkstellig in samewerking met 'n arbitr^ere weerkaatsende lter. Deur gebruik te maak van poortparameterdata van die wafelyster lter word 'n 33 mm lange hamoniekdemper ontwikkel wat 'n maksimum intreeweerkaatsing van -12.5 dB oor die band 17 - 21 GHz handhaaf vir kruindrywingsvlakke van tot 1 kW. Goeie ooreenstemming tussen gesintetiseerde en gemete resultate word weereens waargeneem vir die volledige saamgestelde struktuur.

Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2009. / This thesis details the development and comparison of two linear control systems that performhover control for a vertical take-off and landing unmanned aerial vehicle. A non-linear mathematical model of the aircraft dynamics is developed. A classical successive loop closure control approach is presented, which applies static gains to the decoupled model around hover. A variable gain approach is presented using optimal control, which linearises the aircraftmodel around its state at fixed time steps. Simulation performance and robustness results are examined for both systems. Different aspects of both controller design processes and results are compared, including navigational performance, robustness and ease of use.

Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2007. / The finite element tearing and interconnect(FETI) domain decomposition(DD) method is investigated in terms of the 2D transverse electric(TEz) finite element method(FEM). The FETI is for the first time rigorously derived using the weighted residual framework from which important insights are gained. The FETI is used in a novel way to implement a total-/scattered field decomposition and is shown to give excellent results. The FETI is newly formulated for the time domain(FETI-TD), its feasibility is tested and it is further formulated and tested for implementation on a distributed computer architecture.

Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2009. / ENGLISH ABSTRACT: Sonars are used extensively for underwater sensing and recent advances in forward-looking imaging sonar have made this type of sonar an appropriate choice for use on Autonomous Underwater Vehicles. The images received from these sonar do however, tend to be noisy and when used in shallow water contain strong bottom reflections that obscure returns from actual targets. The focus of this work was the investigation and development of post-processing techniques to enable the successful use of the sonar images for automated navigation. The use of standard image processing techniques for noise reduction and background estimation, were evaluated on sonar images with varying amounts of noise, as well as on a set of images taken from an AUV in a harbour. The use of multiple background removal and noise reduction techniques on a single image was also investigated. To this end a performance measure was developed, based on the dynamic range found in the image and the uniformity of returned targets. This provided a means to quantitatively compare sets of post-processing techniques and identify the “optimal” processing. The resultant images showed great improvement in the visibility of target areas and the proposed techniques can significantly improve the chances of correct target extraction. / AFRIKAANSE OPSOMMING: Sonars word algemeen gebruik as onderwater sensors. Onlangse ontwikkelings in vooruit-kykende sonars, maak hierdie tipe sonar ’n goeie keuse vir die gebruik op ’n Outomatiese Onderwater Voertuig. Die beelde wat ontvang word vanaf hierdie sonar neig om egter raserig te wees, en wanneer dit in vlak water gebruik word toon dit sterk bodemrefleksies, wat die weerkaatsings van regte teikens verduister. Die fokus van die werk was die ondersoek en ontwikkeling van naverwerkings tegnieke, wat die sonar beelde bruikbaar maak vir outomatiese navigasie. Die gebruik van standaard beeldverwerkingstegnieke vir ruis-onderdrukking en agtergrond beraming, is geëvalueer aan die hand van sonar beelde met verskillende hoeveelhede ruis, asook aan die hand van ’n stel beelde wat in ’n hawe geneem is. Verdere ondersoek is ingestel na die gebruik van meer as een agtergrond beramings en ruis onderdrukking tegniek op ’n enkele beeld. Hierdie het gelei tot die ontwikkeling van ’n maatstaf vir werkverrigting van toegepaste tegnieke. Hierdie maatstaf gee ’n kwantitatiewe waardering van die verbetering op die oorspronklike beeld, en is gebaseer op die verbetering in dinamiese bereik in die beeld en die uniformiteit van die teiken se weerkaatsing. Hierdie maatstaf is gebruik vir die vergelyking van verskeie tegnieke, en identifisering van die “optimale” verwerking. Die verwerkte beelde het ’n groot verbetering getoon in die sigbaarheid van teikens, en die voorgestelde tegnieke kan ’n betekenisvolle bedrae lewer tot die suksesvolle identifisering van obstruksies.

Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2007. / The diagonal covariance Gaussian Probability Density Function (PDF) has been a very popular choice as the base PDF for Automatic Speech Recognition (ASR) systems. The only choices thus far have been between the spherical, diagonal and full covariance Gaussian PDFs. These classic methods have been used for some time, but no single document could be found that contains a comparative study on these methods in the use of Pattern Recognition (PR). There also is a gap between the complexity and speed of the diagonal and full covariance Gaussian implementations. The performance differences in accuracy, speed and size between these two methods differ drastically. There is a need to find one or more models that cover this area between these two classic methods. The objectives of this thesis are to evaluate three new PDF types that fit into the area between the diagonal and full covariance Gaussian implementations to broaden the choices for ASR, to document a comparative study on the three classic methods and the newly implemented methods (from previous work) and to construct a test system to evaluate these methods on phoneme recognition. The three classic density functions are examined and issues regarding the theory, implementation and usefulness of each are discussed. A visual example of each is given to show the impact of assumptions made by each (if any). The three newly implemented PDFs are the Sparse-, Probabilistic Principal Component Analysis- (PPCA) and Factor Analysis (FA) covariance Gaussian PDFs. The theory, implementation and practical usefulness are shown and discussed. Again visual examples are provided to show the difference in modelling methodologies. The construction of a test system using two speech corpora is shown and includes issues involving signal processing, PR and evaluation of the results. The NTIMIT and AST speech corpora were used in initialisation and training the test system. The usage of the system to evaluate the PDFs discussed in this work is explained. The testing results of the three new methods confirmed that they indeed fill the gap between the diagonal and full covariance Gaussians. In our tests the newly implemented methods produced a relative improvement in error rate over a similar implemented diagonal covariance Gaussian of 0.3–4%, but took 35–78% longer to evaluate. When compared relative to the full covariance Gaussian the error rates were 18–22% worse, but the evaluation times were 61–70% faster. When all the methods were scaled to approximately the same accuracy, all the above methods were 29–143% slower than the diagonal covariance Gaussian (excluding the spherical covariance method).

This thesis relates to the design, implementation and evaluation of statistical face recognition techniques. In particular, the use of Hidden Markov Models in various forms is investigated as a recognition tool and critically evaluated. Current face recognition techniques are very dependent on issues like background noise, lighting and position of key features (ie. the eyes, lips etc.). Using an approach which specifically uses an embedded Hidden Markov Model along with spectral domain feature extraction techniques, shows that these dependencies may be lessened while high recognition rates are maintained.

Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2009. / ENGLISH ABSTRACT: Energy-efficient (EE) lighting projects form a substantial percentage of Demand Side Management (DSM) initiatives. These largely entail the exchange of one lighting technology for another more energy-efficient lighting technology. The DSM process typically involves a proposal from an Energy Services Company (ESCO) to retrofit an existing lighting technology with another on the property of a third party, the client. For scoping purposes, ESCOs perform energy savings calculations based on information obtained from the datasheets of the relevant lighting technologies. Such datasheet specifications rarely incorporate the effects of supply voltage fluctuations on energy consumption, which can impact on the accuracy of the savings calculations. Furthermore, modern EE lighting technologies such as Compact Fluorescent lamps (CFLs) employ power electronic circuitry that can in principle give rise to Quality of Supply (QoS) problems such as harmonic distortion. The usage profiles of artificial light fittings targeted in DSM interventions represent another important factor in determining the savings impacts of such projects. There is currently limited information on methodologies for obtaining such usage profiles. In practice, the scoping and impact verification of EE lighting projects are conducted using project-specific applications and spreadsheets that are time-consuming and error-prone. In view of the above-mentioned considerations, this investigation aims to address the lack of voltage-dependent energy consumption data and QoS impacts by conducting a laboratory investigation for all relevant lighting technologies, namely incandescent lamps, CFLs, tubular fluorescent lamps and high intensity discharge lamps. Appropriate mathematical models for the voltage-dependent energy consumption characteristics of these light technologies are derived from the measurements. The supply current harmonic distortion associated with the various lamp types are investigated, particularly with regard to neutral current loading caused by zero-sequence harmonics. Methodologies for obtaining accurate and reliable light usage data using commercially available data loggers are reviewed. A database structure is subsequently designed and implemented to store the information relevant for impact assessment, including the mathematical models of energy consumption, supply voltage profiles and light usage profiles. Finally, an Integrated Software Program (ISP) is developed to implement a methodology for assessing the savings impacts of practical EE lighting projects, using the database as the main input source. The ISP is tested by implementing a real case study. It is shown that the ISP yields accurate results for the case study considered in the evaluation. / AFRIKAANSE OPSOMMING: Energiedoeltreffende (ED) beligtingsprojekte vorm ‟n wesenlike persentasie van vraagkantbestuur (VKB) inisiatiewe. Dit het grootliks te doen met die vervanging van een beligtingstegnologie met ‟n ander meer energiedoeltreffende beligtingstegnologie. Die VKB proses behels normaalweg ‟n voorstel van Energie Dienste Maatskappy (EDM) om ‟n bestaande beligtingstegnologie te vervang met ‟n ander op die perseel van ‟n derde party, die kliënt. EDMs doen energiebesparingsberekeninge op grond van tegniese inligting wat vanaf die datablaaie van die betrokke beligtingstegnologieë verkry word. Hierdie datablad spesifikasies maak selde voorsiening vir die uitwerking van toevoerspanningfluktuasies op energieverbruik, wat die akkuraatheid van die besparingsberekeninge kan beïnvloed. Moderne ED beligtingstegnologieë soos kompakte fluoresseerlampe maak verder gebruik van drywingselektronika stroombane wat in beginsel kan lei tot kwaliteit van toevoer (KVT) probleme soos harmoniese distorsie. Die gebruiksprofiele van kunsmatige lig verteenwoordig nog ‟n belangrike faktor wat die besparingsimpakte van VKB projekte bepaal. Daar is tans beperkte informasie oor die metodologie om sulke gebruiksprofiele te verkry. In die praktyk word die verifiëring van die impak van ED beligtingsprojekte gedoen deur gebruik te maak van projekspesifieke programme en sigblaaie wat tydrowend is en geneig is om te lei tot foute. In die lig van die bogenoemde oorwegings, streef hierdie ondersoek om die tekort aan spanningsafhanklike energieverbruiksdata en KVT impakte te aan te spreek deur „n laboratorium ondersoek uit te voer vir al die relevante beligtingstegnologieë, naamlik filament lampe, kompakte fluoresseerlampe, buisvormige fluoresseerlampe en hoë-intensiteit ontladingslampe. Gepaste wiskundige modelle vir die spanningsafhanklikeenergieverbruik eienskappe van hierdie beligtingstegnologieë word vanuit die metings afgelei. Die harmoniese vervorming van die toevoerstroom van die verskillende beligtingstegnologieë word ondersoek, veral met verwysing tot neutraalstroombelasting wat veroorsaak word deur zero volgorde harmoniese ordes. Metodologieë vir die verkryging van akkurate en betroubare ligverbruikprofiele deur die gebruik van komersieel beskikbare dataversamelaars is nagegaan. ͗n Databasis struktuur is vervolgens ontwerp en geïmplementeer om die toepaslike inligting vir bepaling van die impakte te stoor, insluitend die wiskundige modelle vir energieverbruik, toevoerspanning-en ligverbruikprofiele. ‟n Geïntegreerdesagtewareprogram (GSP) is ontwerp om die metodologie vir die bepaling van besparingsimpakte van praktiese ED beligtingsprojekte te implimenteer, deur gebruik te maak die databasis as die hoofbron van insette. Die GSP is getoets deur ‟n werklike gevallestudie te implimenteer. Daar is bewys dat die GSP akkurate resultate lewer vir die gevallestudie wat in die evaluering gebruik is.

Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2008. / A communications protocol is designed for real time control and data handling for a Nanosatellite application. The communication protocol is based on the Controller Area Network (CAN) technology. The protocol handles different message types such as time synchronization, telecommand messages, telemetry acquisition, unsolicited telemetry messages, large file transfers and debug messages. The design of the protocol entails finding a suitable target microcontroller in which the protocol implementation is demonstrated. This requires consideration of a number of development factors such as cost, complexity, availability, reliability and operational environment (space). The AVR AT90CAN128 microcontroller was chosen as a target microcontroller as it gave most of the required factors mentioned above. The protocol implementation involves developing low level software drivers, the middleware and the application programs to demonstrate handling of each supported message. In the implementation the media access scheme and low layer communication is provided by the CAN low level kernel (physical and data link layers). The protocol performance was evaluated by measuring the software response latencies, the bus throughputs and the software efficiencies. Power consumption due to CAN communication was also measured. System reliability was tested by loading the CAN bus with extreme communication traffic and letting the system run for a long time. The observation was that messages were handled consistently.

Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2009. / This thesis describes the successful development of an autopilot for an unmanned radio controlled helicopter. It presents a non-linear helicopter model. An adaptive linearised model is derived and used to design a controller. The adaptive full state controller is tested in various ways, including two aerobatic manoeuvres. A number of analyses are performed on the controller, including its robustness to parameter changes, noisy estimates, wind and processing power. The controller is compared with a non-adaptive counterpart, which leads to the conception, design and analysis of a much improved control structure. Practical flight test results are presented and analysed. In some instances available literature was reworked and re-derived to produce a genericmodel-controller package that can easily be adapted for helicopters of any make, model and size.

Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2006. / The open-ended coaxial probe is revisited as a broadband measurement system for general high frequency permittivity measurements. Three coaxial probes were developed that are suited for the measurement of both liquids and solids. The components of a permittivity measurement system were investigated and improvements were made to the coaxial probe where needed. This includes the development of a full wave code with great calculation time improvements without sacrificing accuracy. This code allows measurements to be performed in a high frequency laboratory and the permittivity extracted without any mentionable delay. A capacitance model that better describes the impedance of an open-ended coaxial line is also suggested that can be used for real-time permittivity extraction over a limited frequency range. Calibration formed a vital part of the project and great time was spent developing a TRL and a SOLT calibration set for the coaxial probe geometry. The combination of the TRL and SOLT standards also allows measurement of the residual errors after calibration and is used in an uncertainty analysis of the extracted permittivity. Well known materials such as PTFE, PVC, methanol and water were measured to test the probes. The measured dielectric constants are all within 3% of values quoted in literature. The loss term of the samples are also in good agreement with the expected values.