Modeling and interactive rendering of urban environments are difficult as it is telling the story of simulating our real and complex world. Urban environment contains various buildings with different kinds of materials and different shapes, modeling such objects with conventional triangle mesh are not reasonable. It's not difficult to create the models, but to store and transmit the enormous data caused by the mesh models. Such amount of data is also very difficult to be put into computer memory for realtime rendering. Besides, roads and trees are also important parts of an urban environment. It is obvious that they share almost the same complexity of modeling and rendering with buildings due to similar problems. Therefore realtime interactive rendering with good visualization quality is very difficult to achieve. In this thesis, we are going to introduce various techniques for different objects in an urban environment to overcome mentioned problems. The first part of this thesis introduces a compact representation of urban building models called Non-uniform Texture, NUT for short. The representation is able to model a building with memory footprint more than two orders of magnitude smaller than that of its mesh counterpart, while impressive details and rendering performance are preserved. The second part introduces a compact representation of urban roads. The representation is also very efficient in data storage and memory consumption. Combing with NUT, the representation also offers great visual quality. The third part introduces an efficient rendering pipeline of trees. The pipeline requires no details to be stored to achieve data efficiency. The pipeline uses almost all functionality of mainstream GPU to provide stunning visual and rendering performance. The final part of this thesis introduces a robust computation of up vector, to solve the problem of camera rotation during video capture. With all the mentioned features, we are going to produce a living urban environment within one system. / published_or_final_version / Computer Science / Doctoral / Doctor of Philosophy

Achieving realistic simulations of terrain vehicles in their work environment does not only require a careful model of the vehicle itself but the vehicle's interactions with the surroundings are equally important. For off-road ground vehicles the terrain will heavily affect the behaviour of the vehicle and thus puts great demands on the terrain model. The purpose of this project has been to develop and evaluate a deformable terrain model, meant to be used in real-time simulations with multi-body dynamics. The proposed approach is a modification of an existing elastoplastic model based on linear elasticity theory and a capped Drucker-Prager model, using it in an adaptive way. The original model can be seen as a system of rigid bodies connected by elastoplastic constraints, representing the terrain. This project investigates if it is possible to create dynamic bodies just when it is absolutely necessary, and store information about possible deformations in a grid. Two methods used for transferring information between the dynamic bodies and the grid have been evaluated; an interpolating approach and a discrete approach. The test results indicate that the interpolating approach is preferable, with better stability to an equal performance cost. However, stability problems still exist that have to be solved if the model should be useful in a commercial product.

Not available / text

Context. Video games are a big entertainment industry. Many video games let players play against or together. Some video games also make it possible for players to play against or together with computer controlled players, called bots. Artificial Intelligence (AI) is used to create bots. Objectives. This thesis aims to implement cooperative behaviors between two bots and determine if the behaviors lead to an increase in win ratio. This means that the bots should be able to cooperate in certain situations, such as when they are attacked or when they are attacking. Methods. The bots win ratio will be tested with a series of quantitative experiments where in each experiment two teaming bots with cooperative behavior will play against two teaming bots without any cooperative behavior. The data will be analyzed with a t-test to determine if the data are statistical significant. Results and Conclusions. The results show that cooperative behavior can increase performance of two teaming Real Time Strategy bots against a non-cooperative team with two bots. However, the performance could either be increased or decreased depending on the situation. In three cases there were an increase in performance and in one the performance was decreased. In three cases there was no difference in performance. This suggests that more research is needed for these cases.

Heartwater is a tick borne disease that affects ruminants and wild animals in Africa south of the Sahara. It is caused by Ehrlichia ruminantium and transmitted by the tick Amblyomma hebraeum. The protocols currently used to detect heartwater take several days to complete. Here, we describe the development of a pCS20 quantitative real-time PCRTaqMan probe assay to detect E. ruminantium in livestock blood and ticks from the field. The assay is based on the conserved pCS20 gene region of E. ruminantium that contains two overlapping genes, rnc and ctaG [Collins, N.E., Liebenberg, J., De Villiers, E.P., Brayton, K.A., Louw, E., Pretorius, A., Faber, F.E., Van Heerden, H., Josemans, A., Van Kleef, M., Steyn, H.C., Van Strijp, M.F., Zweygarth, E., Jongejan, F., Maillard, J.C., Berthier, D., Botha, M., Joubert, F., Corton, C.H., Thomson, N.R., Allsopp, M.T., Allsopp, B.A., 2005. The genome of the heartwater agent Ehrlichia ruminantium contains multiple tandem repeats of actively variable copy number. PNAS 102, 838–843]. The pCS20 quantitative real-time PCRTaqMan probe was compared to the currently used pCS20 PCR and PCR/32P-probe test with regards to sensitivity, specificity and the ability to detect DNA in field samples and in blood from experimentally infected sheep. This investigation showed that the pCS20 quantitative real-time PCRTaqMan probe was the most sensitive assay detecting seven copies of DNA/ml of cell culture. All three assays, however, cross react with Ehrlichia canis and Ehrlichia chaffeensis. The pCS20 real-time PCR detected significantly more positive field samples. Both the PCR and pCS20 real-time PCR could only detect E. ruminantium parasites in the blood of experimentally infected sheep during the febrile reaction. The PCR/32P-probe assay, however, detected the parasite DNA 1 day before and during the febrile reaction. Thus, because this new quantitative pCS20 real-time PCRTaqMan probe assay was the most sensitive and can be performed within 2 h it is an effective assay for epidemiological surveillance and monitoring of infected animals.

published_or_final_version / abstract / toc / Computer Science and Information Systems / Doctoral / Doctor of Philosophy

A PCR Differentiation Method, a hybrid of the US EPA and MOE methods, for quantifying human infectious C. parvum/C. hominis as a group and non-human infectious C. andersoni/C. muris was developed. Primers and probe sets targeting the hsp70 gene were designed for C. andersoni/C. muris; those for C. parvum/C. hominis were obtained from the MOE method. Results showed that C, andersoni/C. muris primers were specific for C. andersoni/C. muris oocysts, while those for C. parvum/hominis primers detected C. parvum/hominis and C. meleagridis. All primers were then used to quantify oocysts from urban and agricultural environmental water samples in Kitchener/Waterloo. Human infectious Giardia lamblia was also incorporated into this study. C. parvum/C. hominis and Giardia lamblia were detected at urban and agricultural areas, whereas C. andersoni/C. muris was only detected at agricultural sites. The PCR Differentiation Method is a reliable method for quantifying Cryptosporidium and Giardia lamblia in environmental water samples. / Best in Science Program of the Ontario Ministry of the Environment, Natural Sciences and Engineering Council (NSERC) of Canada Discovery Grants

Real-time social systems are the fastest growing phenomena on the web, enabling millions of users to generate, share, and consume content on a massive scale. These systems are manifestations of a larger trend toward the global sharing of the real-time interests, affiliations, and activities of everyday users and demand new computational approaches for monitoring, analyzing, and distilling information from the prospective web of real-time content. In this dissertation research, we focus on the real-time social trails that reflect the digital footprints of crowds of real-time web users in response to real-world events or online phenomena. These digital footprints correspond to the artifacts strewn across the real-time web like posting of messages to Twitter or Facebook; the creation, sharing, and viewing of videos on websites like YouTube; and so on. While access to social trails could benefit many domains there is a significant research gap toward discovering, modeling, and leveraging these social trails. Hence, this dissertation research makes three contributions: • The first contribution of this dissertation research is a suite of efficient techniques for discovering non-trivial social trails from large-scale real-time social systems. We first develop a communication-based method using temporal graphs for discovering social trails on a stream of conversations from social messaging systems like instant messages, emails, Twitter directed or @ messages, SMS, etc. and then develop a content-based method using locality sensitive hashing for discovering content based social trails on a stream of text messages like Tweet stream, stream of Facebook messages, YouTube comments, etc. • The second contribution of this dissertation research is a framework for modeling and predicting the spatio-temporal dynamics of social trails. In particular, we develop a probabilistic model that synthesizes two conflicting hypotheses about the nature of online information spread: (i) the spatial influence model, which asserts that social trails propagates to locations that are close by; and (ii) the community affinity influence model, which asserts that social trail prop- agates between locations that are culturally connected, even if they are distant. • The third contribution of this dissertation research is a set of methods for social trail analytics and leveraging social trails for prognostic applications like real-time content recommendation, personalized advertising, and so on. We first analyze geo-spatial social trails of hashtags from Twitter, investigate their spatio-temporal dynamics and then use this analysis to develop a framework for recommending hashtags. Finally, we address the challenge of classifying social trails efficiently on real-time social systems.

The production of bio-ethanol from lignocellulosic material requires a more efficient process to be feasible and compete with products from fossil fuels. There is a need to rapidly and nondestructively be able to determine key components during fermentation. Raman spectroscopy is a technique, which can be used to monitor the fermentation process in real-time and provide information about key components which can be accessed immediately, thus facilitating process control. A continuous system with membrane cell recycling was set up and fermentations were performed using Saccharomyces cerevisiae ATCC 96581. Fermentations were performed to test for optimal dilution rates and operating times, the effect of different sugar concentrations in the media feed, and which position in the system was optimal for Raman data collection. Raman data and aliquot samples for HPLC validation were continuously collected throughout the fermentations. Raman data was analysed with PLS models to obtain component concentrations, for which RMSE was calculated in order to compare to HPLC validation set. Fermentations were performed with synthetic glucose media as well as with poplar hydrolysate. It was shown that the continuous system with membrane cell recycling could achieve a glucose-to-ethanol conversion of between 75-100%. The process could be sufficiently monitored by Raman spectroscopy, and predicted concentrations were within the range of the validation set in most cases. However, the error of prediction varied between the different fermentations.

As multi-core systems are becoming more popular in real time embedded systems, strict timing requirements for accessing shared resources must be met. In particular, a detailed latency analysis for Double Data Rate Dynamic RAM (DDR DRAM) is highly desirable. Several researchers have proposed predictable memory controllers to provide guaranteed memory access latency. However, the performance of such controllers sharply decreases as DDR devices become faster and the width of memory buses is increased. Therefore, a novel and composable approach is proposed that provides improved latency bounds compared to existing works by explicitly modeling the DRAM state. In particular, this new approach scales better with increasing number of cores and memory speed. Benchmark evaluation results show up to a 45% improvement in the worst case task execution time compared to a competing predictable memory controller for a system with 16 cores.