Listing 1 - 6 of 6 |
Sort by
|
Choose an application
Moderne consumenten beschikken over tal van elektronische toestellen, zoals mobiele telefoons, digitale camera's, GPS, PDA en MP3 spelers. De mogelijkheden van elk van deze toestellen is gedurende de voorbije jaren sterk toegenomen, met zowel een spectaculaire verbetering in het aantal functies, als in de kwaliteit van de geleverde functionaliteit. Het is echter niet evident om de vereiste rekenkracht te bieden die nodig is om deze trend mogelijk te maken, binnen de beperkingen van een batterijgevoed draagbaar toestel. Hierdoor is het bij het ontwerp van de draagbare toestellen van de toekomst nodig om het volledige systeem te optimaliseren. De ingebedde processoren moeten de juiste balans houden tussen flexibiliteit, energie-effciëntie en rekenkracht. Hierbij zal een ontwerper het energieverbruik minimaliseren (zover als nodig is), voor de vereiste rekenkracht, met een voldoende flexibiliteit. Hierbij moet de invloed van het ontwerp van een component op andere delen van het systeem in rekening worden gebracht, wat leidt tot een zoektocht naar de Pareto-optimale punten in een multidimensionale exploratieruimte.De focus van dit onderzoek, binnen het globale ontwerp van batterijgevoede ingebedde systemen, ligt op de energiebewuste architectuurexploratie van domeinspecifieke processordatapaden tijdens de vroegere fases van het ontwerp en de co-optimalisatie van de datapadarchitectuur met de bijbehorende afbeeldings- of compilatietechnieken. Door het uitvoeren van een gedetailleerde energieschatting voor een volledig ingebed systeem worden de flessenhalzen geïdentificeerd, zowel voor het energieverbruik als voor de performantie. Tegelijkertijd wordt een duidelijk overzicht gegeven van de onderlinge relaties tussen de belangrijkste componenten. Op basis van deze kennis worden architectuurwijzigingen en bijbehorende compilatietechnieken voorgesteld, met een duidelijke focus op het processordatapad. De belangrijkste bijdragen van deze thesis zijn ondermeer een energie-, performantie- en verbindingsgevoelige exploratie voor ingebedde processoren (toegepast op hoge granulariteit herconfigureerbare hardware), een softwaremethode voor het exploiteren van heterogeen dataparallelisme, een techniek voor een contextafhankelijke vereenvoudiging van vermenigvuldigingen met een constante factor, inclusief de wisselwerking met de nauwkeurigheidsvereisten van de applicatie.
Choose an application
Modern consumers carry many electronic devices, like a mobile phone, digital camera, GPS, PDA and an MP3 player. The functionality of each of these devices has gone through an important evolution over recent years, with a steep increase in both the number of features as in the quality of the services that they provide. However, providing the required compute power to support (an uncompromised combination of) all this functionality is highly non-trivial. Designing processors that meet the demanding requirements of future mobile devices requires the optimization of the embedded system in general and of the embedded processors in particular, as they should strike the correct balance between flexibility, energy efficiency and performance. In general, a designer will try to minimize the energy consumption (as far as needed) for a given performance, with a sufficient flexibility. However, achieving this goal is already complex when looking at the processor in isolation, but, in reality, the processor is a single component in a more complex system. In order to design such complex system successfully, critical decisions during the design of each individual component should take into account effect on the other parts, with a clear goal to move to a global Pareto optimum in the complete multi-dimensional exploration space. In the complex, global design of battery-operated embedded systems, the focus of Ultra-Low Energy Domain-Specific Instruction-Set Processors is on the energy-aware architecture exploration of domain-specific instruction-set processors and the co-optimization of the datapath architecture, foreground memory, and instruction memory organisation with a link to the required mapping techniques or compiler steps at the early stages of the design. By performing an extensive energy breakdown experiment for a complete embedded platform, both energy and performance bottlenecks have been identified, together with the important relations between the different components. Based on this knowledge, architecture extensions are proposed for all the bottlenecks.
Electronics --- Electrical engineering --- Computer architecture. Operating systems --- microprocessoren --- architectuur (informatica) --- elektrische circuits
Choose an application
Modern consumers carry many electronic devices, like a mobile phone, digital camera, GPS, PDA and an MP3 player. The functionality of each of these devices has gone through an important evolution over recent years, with a steep increase in both the number of features as in the quality of the services that they provide. However, providing the required compute power to support (an uncompromised combination of) all this functionality is highly non-trivial. Designing processors that meet the demanding requirements of future mobile devices requires the optimization of the embedded system in general and of the embedded processors in particular, as they should strike the correct balance between flexibility, energy efficiency and performance. In general, a designer will try to minimize the energy consumption (as far as needed) for a given performance, with a sufficient flexibility. However, achieving this goal is already complex when looking at the processor in isolation, but, in reality, the processor is a single component in a more complex system. In order to design such complex system successfully, critical decisions during the design of each individual component should take into account effect on the other parts, with a clear goal to move to a global Pareto optimum in the complete multi-dimensional exploration space. In the complex, global design of battery-operated embedded systems, the focus of Ultra-Low Energy Domain-Specific Instruction-Set Processors is on the energy-aware architecture exploration of domain-specific instruction-set processors and the co-optimization of the datapath architecture, foreground memory, and instruction memory organisation with a link to the required mapping techniques or compiler steps at the early stages of the design. By performing an extensive energy breakdown experiment for a complete embedded platform, both energy and performance bottlenecks have been identified, together with the important relations between the different components. Based on this knowledge, architecture extensions are proposed for all the bottlenecks.
Electronics --- Electrical engineering --- Computer architecture. Operating systems --- microprocessoren --- architectuur (informatica) --- elektrische circuits
Choose an application
Choose an application
Spectral imaging is a technology that combines photography andspectroscopy by sampling the electromagnetic spectrum for each point in the scene. This spectrum can be used as a ``fingerprint'' to identify different materials and analyze their properties.Recent developments have enabled the integration of thin-film filters onto pixels of an image sensor, making it possible for pixels to sample specific wavelengths.
Furthermore, the filter-on-chip integration makes spectral cameras robust, compact, and lightweight enough to be used in various applications including remote sensing, industrial quality control, healthcare and precision agriculture.However, to enable mass production and widespread adoption across industries, it will be essential to minimize costly and time-consuming expert support for camera calibration.One of the key concerns in calibration is the angle-dependent transmittance of thin-film filters.When illuminated by focused light, the filters might select other wavelengths than originally intended, causing the measured spectra to be smoothed and spectrally shifted. This can strongly impact application performance and hence also limits the selection of compatible lenses.Hardware solutions include the use of filter materials that are less angle sensitive, expensive telecentric lenses, or the development of lens-specific filter designs.These solutions cannot always be implemented due to a lack of compatible materials or budgetary constraints.Instead, we propose a software correction approach that enables cost-efficient use of the same sensor with most commercially available non-telecentric lenses.In this regard, this work has two major contributions.The first contribution consists of new and insightful analytical models that can be used to predict the transmittance of thin-film filters illuminated by camera lenses. The models cover arbitrarily tilted circular apertures and lenses with vignetting.Using analytical approximations predicated on classical thin-film theory, spectrally shifted measurements can be corrected in a practical way. As a result, it does not matter anymore which lens is used as long as the shift can be predicted and other distortions are negligible.The second contribution is the derivation of a novel thin-film filter model which takes into account the finite width of integrated thin-film filters. This ``tiny filter'' model predicts significant deviations from classical thin-film theory which assumes an infinite filter width.The predicted deviations are observed in multiple spectral cameras and result in the formulation of new limitations for filter miniaturization. Furthermore, the model might be used to build software tools to efficiently design tiny filters or perhaps avoid unnecessary investments.In conclusion, the developed software corrections expand the selection of compatible lenses and simplify the calibration process as long as classical thin-film theory applies.When this is no longer the case, no correction methods currently exist. The development of such methods is therefore suggested as a topic for further investigations.
Choose an application
Hyperspectral imaging increases the capabilities of traditional machine vision by extending the information content from three broad bands (RGB) to a spectrum of multiple narrow bands beyond the visible domain. This provides a combination of spectral and spatial information, which increases the potential for applications with respect to traditional color imaging or point spectroscopy.While hyperspectral imaging is a technology that has already shown high potential in a wide range of application domains, its adoption by Industry has been slow so far. This has been attributed to the high camera cost on one hand and processing expertise required for the large amounts of data generated on the other hand. In this sense, recent hyperspectral technology developments are trying to bridge this gap by creating more affordable cameras that can better meet industrial needs. Typically, the development of more industrially suited cameras is done at the expense of either a lower number of bands or lower spatial resolution, which may in turn reduce their discrimination performance with respect to high-end research equipment.To explore these trade-offs, a system-wide exploration was performed of hyperspectral imaging based on cameras, which target industrial needs. To this end, multiple system parameters such as wavelength range, camera hardware, illumination system or data analysis methods were varied for some specific applications.First, system level optimization was explored by using the wavelength range as a key system parameter to reduce camera hardware cost for a textile sorting application. In this application, it is shown that a suboptimal wavelength range may still be able to meet the discrimination requirements, while substantially reducing the hardware cost.Next, the focus was shifted to a case of seed mix ingredient discrimination and quantification. The added value of data preprocessing and the integration of spatial information with the spectral information is demonstrated to increase the system performance and reach the application targets. Further, it is demonstrated that the illumination system is a key parameter in hyperspectral imaging applications, in particular with snapshot cameras. The presented results show how illumination can have a relevant impact on the performance (up to 10% increase in classification accuracy) by achieving a more balanced spectral and spatial illumination.Finally, different system parameters such as camera hardware, illumination system and data analysis methods are evaluated together. In terms of data processing, the impact of pre- and post-processing methods are explored, while pixel-based analysis is compared to a more joint spatial-spectral image analysis based on convolutional neural networks. It is demonstrated that the joint evaluation of all these system parameters allows to make the best choices to meet the application requirements and increased the mean classification accuracy by up to 25%. Moreover, it allows to explore varied system configurations that offer different performance-cost-speed tradeoffs.To conclude this dissertation, some guidelines for system level optimization and parameter selection are proposed from the application characteristics and requirements. This paves the way for a broader industrial adoption of hyperspectral imaging technology.
Listing 1 - 6 of 6 |
Sort by
|