Background. Particle systems are a staple feature of most modern renderers. There are several technical challenges when it comes to rendering transparent particles. Particle sorting along the view direction is required for proper blending and casting shadows from particles requires non-standard shadow algorithms. A recent technology that could be used to adress these technical challenges is hardware accelerated ray tracing. However there is a lack of performance data gathered from this type of hardware. Objectives. The objective of this thesis is to measure the performance of a prototype that uses hardware accelerated ray tracing to render particles that cast shadows. Methods. A prototype is created and measurements of the ray tracing time are made. The scene used for the benchmark test is a densely packed particle volume of highly transparent particles, resulting in a scene that looks similar to smoke. Particles are sorted along a ray by repeatedly tracing rays against the scene and incrementing the ray origin past the previous intersection point until it has passed all the objects that lie along the ray. Results. Only a small number of particles can be rendered if real time rendering speeds are desired. High quality shadows can be produced in a way that is very simple compared to texture based methods. Conclusions. Future hardware speed ups can improve the rendering speeds but more sophisticated sorting methods are needed to render larger amounts of particles.

The complexity of mobile robots increases as more parts are added to the system. Introducing microcontrollers into a mobile robot abstracts and modularises the system architecture, creating a demand for seamless microcontroller integration. The Robot Operating System (ROS) used by ABB’s new mobile robot, the mobile YuMi prototype (mYuMi), allows standardised robot software libraries and packages to simplify robotic creations. As ABB is porting over from ROS1 to ROS2, the ROS2 compatible Microcontroller Robot Operating System (micro-ROS) will be incorporated into the system to smoothly integrate microcontrollers into mYuMi. In order to display the validity of micro-ROS, this project used tracing and latency measurements with external applications to test the remote communication between mYuMi using ROS2 and microcontrollers using micro-ROS, with three different microcontrollers tested. The communication was evaluated in different scenarios with a test bench, using ping pong communication to get the round-trip time. A reinforcement of the test results was presented by demonstrating the use of micro-ROS live in a prototype developed, where mYuMi controlled a 1D rangefinder and an RC servo motor by utilising two microcontrollers. The results concluded that the micro-ROS delay could be analysed in theory with external applications, equivalent micro-ROS functionality should apply to most microcontrollers, and the test results and prototype displayed the potential of micro-ROS matching ROS2 in terms of delay and stability.

Over the past 20 years, the development in ComputationalElectromagnetics has produced a vast choice of methods based onthe large number of existing mathematical formulations of theMaxwell equations. None of them dominate over the others,instead they complement each other and the choice of methoddepends on the frequency range of the electromagnetic waves.This work is focused on the most popular method in the highfrequency scenario, namely the Geometrical Theory ofDiffraction (GTD). The main advantage of GTD is the ability topredict the electromagnetic field asymptotically in the limitof vanishing wavelength, when other methods, such as the Methodof Moments, become computationally too expensive. The low cost of GTD is due to both the fact that there is noruntime penalty in increasing the frequency and that the raytracing, which GTD is based on, is a geometrical technique. Thecomplexity is then no longer dependent on electrical size ofthe problem but instead on geometrical sub problems which aremanageable. For industrial applications the geometricalstructures, with which the rays interact, are modelled bytrimmed Non-Uniform Rational B-Spline (NURBS) surfaces, themost recent standard used to represent complex free-formgeometries. Due to the introduction of NURBS, the geometrical subproblems tend to be mathematically and numerically cumbersome,but they can be highly simplified by proper Object Orientedprogramming techniques. This allowed us to create a flexiblesoftware package, MIRA: Modular Implementation of Ray Tracingfor Antenna Applications, with an architecture that separatesmathematical algorithms from their implementation details andmodelling. In addition, its design supports hybridisationtechniques in combination with other methods such as Method ofMoment (MoM) and Physical Optics (PO). In a first hybrid application, a triangle-based PO solveruses the shadowing information calculated with the ray tracerpart of MIRA. The occlusion is performed between triangles andtheir facing NURBS surfaces rather than between their facingtriangles, thus reducing the complexity. Then the shadowinginformation is used in an iterative MoM-PO process in order tocover higher frequencies, where the contribution of theshadowing effects, in the hybrid formulation, is believed to bemore significant. Thesis presented at the Royal Institute of Technology ofStockholm in 2003, for the degree of Licentiate in ScientificComputing. / NR 20140805

京都大学 / 新制・論文博士 / 博士(医学) / 乙第13466号 / 論医博第2253号 / 新制||医||1055(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 村川 泰裕, 教授 中島 貴子, 教授 藤田 恭之 / 学位規則第4条第2項該当 / Doctor of Medical Science / Kyoto University / DFAM

The Dlx homeobox genes encode homeodomain transcription factors that are involved in multiple developmental aspects. In the brain, these genes take part in neuronal migration and differentiation, more precisely in the migration and differentiation of GABAergic neurons. Dysfunctions in the GABAergic system can lead to various pathological conditions, where impaired inhibitory function is one of the main causes of several neuropathies characterized by neuronal hyperexcitability. The Dlx genes are organized as bi-gene clusters and highly conserved cis-regulatory elements have been previously characterized to be fundamental for the regulation of Dlx expression in developing embryos of different vertebrates. The activity of these regulatory elements and the Dlx genes has been well studied in developmental stages of mice and zebrafish, but little is known about their activity in the adult brain. The extensive neurogenesis that takes place in the adult zebrafish brain provides an ideal platform for the visualization of mechanisms involving dlx genes during adulthood and their possible involvement in adult neurogenesis. Here we show novel information concerning the expression of dlx1a, dlx2a, dlx5a and dlx6a in the adult zebrafish brain and provide insight into the identity of cells that express dlx. We also demonstrate the involvement of dlx genes in brain regeneration and through lineage tracing, their fate determination in the adult zebrafish brain. Analyses in the adult zebrafish has revealed that all four dlx paralogs are expressed in the forebrain and midbrain throughout adulthood and expression is found in almost all areas presenting continuous proliferation. Most dlx-expressing cells present GABAergic neuronal identity in the adult forebrain where, in some areas they were identified as the Calbindin subtype. In some areas of the midbrain, especially within the hypothalamus, many dlxexpressing cell co-localized with a marker for neural stem cells. However, cells expressing dlx iii genes did not co-localize with markers for proliferating cells or for glia. Investigations during brain regeneration in response to injury in the adult zebrafish brain has revealed that dlx5a expression decreases shortly after lesion and that the dlx5a/6a bi-gene cluster, more specifically, dlx5a, is up regulated during the peak of regeneration response proposing a possible role for dlx during regeneration in adults. Studies of lineage tracing have shown the progeny of dlx1a/2a-expressing cells in adults are located within small clusters in different areas of the adult brain where they seem to become mature neurons. Our observations provide a better understanding about the role of dlx genes during adulthood, further contributing to the general knowledge of the molecular pathways involved in adult neurogenesis and regeneration in the zebrafish adult brain.

京都大学 / 0048 / 新制・論文博士 / 博士(医学) / 乙第13142号 / 論医博第2142号 / 新制||医||1026(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 伊佐 正, 教授 野田 亮, 教授 岩田 想 / 学位規則第4条第2項該当 / Doctor of Medical Science / Kyoto University / DFAM

No description available.

Groundwater flow through fractured karst conduit systems can be complex and difficult to diagnose. This project explores the role of geologic structures that influence the location of recharge points, flow paths, velocities, and discharge locations within Morrell Cave and at the resurgence of Morrell Spring, both of which are located near the city of Bluff City, TN. Understanding of the groundwater sources and flow paths in the Bluff City area will allow future researchers to more readily identify sources of pollution and better resolve local agricultural well drawdown conflicts among residents. The objectives of this project are to: 1) identify the active allogenic recharge sources of Morrell Spring, the largest known spring in the Bluff City area; 2) delineate a springshed for Morrell Spring and; 3) diagnose the structural controls for groundwater flow paths to Morrell Spring. It was found that surface streams flowing across the Sevier Shale on the northern slope of Holston Mountain enter the subsurface karst system through swallets along the Sevier shale and the Jonesboro Limestone contact. Once underground the water flows to the NW following 2 dominant joint sets until it reaches the NE/SW oriented fault line along which Morrell Cave has formed. Upon entering the cave the groundwater flows to the NE to Morrell Spring and into the South Fork Holston River.

Processor address traces are invaluable for characterizing workloads and testing proposed memory hierarchies. Long traces are needed to exercise modern cache designs and produce meaningful results, but are difficult to collect with hardware monitors because microprocessors access memory too frequently for disks or other large storage to keep up. The small, fast buffers of the monitors fill quickly; in order to obtain long contiguous traces, the processor must be stopped while the buffer is emptied. This halting may perturb the traces collected, but this cannot be measured directly, since long uninterrupted traces cannot be collected. We make the case that hardware performance counters, which collect runtime statistics without influencing execution, can be used to measure halting effects. We use the performance counters of the Pentium 4 processor to collect statistics while halting the processor as if traces were being collected. We then compare these results to the statistics obtained from unhalted runs. We present our results in terms of which counters are affected, why, and what this means for trace-collection systems.

No description available.