Crowd simulations are attracting increasing attention from both academia and the industry field and are implemented across a vast range of applications, from scientific demonstrations to video games and films. As such, the demand for greater realism in their aesthetics and the amount of agents involved is always growing. A successful crowd simulation must simulate large numbers of pedestrians' behaviours as realistically as possible in real-time. The thesis looks at two important aspects of crowd simulation and real-time animation. First, this thesis introduces a new data structure called Extended Oriented Bounding Box (EOBB) and related methods for fast collision detection and obstacle avoidance in the simulation of crowd motion in virtual environments. The EOBB is extended to contain a region whose size is defined based on the instantaneous velocity vector, thus allowing a bounding volume representation of both geometry and motion. Such a representation is also found to be highly effective in motion planning using the location of vertices of bounding boxes in the immediate neighbourhood of the current crowd member. Second, we present a detailed analysis of the effectiveness of spatial subdivision data structures, specifically for large-scale crowd simulation. For large-scale crowd simulation, computational time for collision detection is huge, and many studies use spatial partitioning data structure to reduce the computational time, depicting their strengths and weaknesses, but few compare multiple methods in an effort to present the best solution. This thesis attempts to address this by implementing and comparing four popular spatial partitioning data structures with the EOBB.

The first step in many mining operations is blasting, and the purpose of blasting is to fragment the rock mass in the most efficient way for that mine site and the material end use. Over time, new developments to any industry occur, and design and implementation of traditional techniques have to change as a consequence. Possibly the greatest improvement in blasting in recent years is that of electronic detonators. The improvements related to safety and increased fragmentation have been invaluable. There has been ongoing debate within the explosives industry regarding two possible theories for this. Shorter timing delays that allow interaction between adjacent shock waves or detonation waves, or the increase in accuracy associated with electronic detonators. Results exist on the improved accuracy of electronic detonators over that of electric or non-electric, but data on the relationship between the collision of dynamic stress waves and fragmentation is less understood. Publications stating that the area of greatest fragmentation will occur between points of detonation where shock waves collide exist, but experimental data to prove this fact is lacking. This dissertation looks extensively at the head on collision of shock (in the rock mass) and detonation (in the detonation column) waves with relation to fragmentation through a number of small scale tests in concrete. Timing is a vital tool for this collision to occur and is the variable utilized for the studies. Small scale tests in solid masonry blocks, 15 x 7⅞ x 7⅞ inches in size, investigated shock and detonation wave collisions with instantaneous detonation. Blocks were wrapped in geotextile fabric and a wire mesh to contain the fragments so that in situ tensile crack formations could be analyzed. Detonating cord was used as the explosive with no stemming to maintain the shock pressure but reduce the gas pressure phase of the fragmentation cycle. Model simulations of these blocks in ANSYS Autodyn looked at the stress and pressure wave patterns and corresponding damage contours for a direct comparison with the experimental investigation. Detonation wave collision in a single blast hole was found to positively influence the fragmentation and throw of the material. Mean fragment size decreased compared to tests with no detonation wave collision. Area of greatest throw occurred at the point of detonation collision where a buildup of gas pressure exited the block from one location. Head on collision of shock waves did not positively influence the muck pile. Largest fragments were located at the point of shock collision. The lack of particle velocity with relation to shock collision in previous literature could be attributed to the increased particle size here. Directional particle velocities could actually increase the strength and density of the rock at this location, decreasing the degree of fragmentation rather than increasing it.

Quadrupole ion trap(QIT) mass spectrometry has become one of the most widelyused tools in the analysis of the structure of small molecules. The motion of the ionsstored in the quadrupole ion trap is extremely important. This ion motion within thequadrupole ion trap is controlled by several factors including the m/z ratio and thecollisional cross section of the ion. Investigation of ion motion within the QIT has thepotential to elucidate a new way to separate ions based on these factors. DC tomographyexperiments allow for the trajectory of the ion motion to be measured withoutmodifications to the ion trap. The ability to use DC tomography for separation ofisomeric ions on a commercial GC/MS system was investigated.Investigation of the mass range within the ion trap is necessary for the analysis ofa wide range of molecules. The ability of the quadrupole ion trap to perform MS/MSanalyses can provide insight into the structural information of many compounds.However, there exists a low mass cut-off (LMC) within the quadrupole ion trap and thusinformation about the low m/z fragments from a parent ion is lost. Schwartz and coworkerspresented a new technique labeled pulsed q dissociation (PQD) at the 53rdAnnual ASMS Conference in San Antonio TX in 2005. PQD eliminates the LMC byperforming CID at a qz of 0.4 but, then immediately lowering the q level before the massscan in a linear ion trap. By operating the quadrupole ion trap in this same manner, lowm/z product ions can be detected. This technique and elucidation of the energetic processcontained within PQD were explored further using a modified commercial quadrupoleion trap and the results discussed in this work.

Nous nous intéressons aux simulations mécaniques pour des applications de synthèse d'images, particulièrement en temps réel, comme les jeux vidéos ou les simulateurs d'apprentissage de gestes techniques. La simulation physique interactive a fait des progrès remarquables ces dernières années, mais n'a pas encore atteint un niveau de maturité suffisant pour être applicable par des non-spécialistes à une grande variété de modèles. Nous présentons ici des contributions permettant de lever un certain nombre de verrous technologiques. La modélisation d'objets souples repose généralement sur une décomposition en cellules déformables. Nous avons proposé des éléments finis tétraédriques rapides et robustes, et généralisé à des grilles hexaédriques simplifiant grandement les problèmes de maillage et de contrôle de la résolution. La détection et la réponse aux collisions posent des problèmes géométriquement complexes et mécaniquement discontinus, qui causent bien souvent les principaux goulots d'étranglement des simulations. Nous avons proposé des solutions au problème du contrôle du temps de calcul, de la diversité des modèles géométriques des surfaces de contact, ainsi que de la robustesse des réponses et de leur calcul accéléré sur GPU. Une grande attention a été portée aux architectures logicielles permettant d'accroître la modularité et les performances des simulateurs. Nous avons proposé un graphe de scène mécanique permettant de hiérarchiser les modèles et décomposer les algorithmes en modules indépendants. Noua avons également exploré la distribution des simulations sur architectures parallèles, rendue nécessaire par l'évolution des matériels.

Dans cette thèse expérimentale, on s'intéresse aux collisions de gouttes mettant en jeu des interfaces asymétriques, soit deux gouttes constituées de liquide différent ou des gouttes de taille différente et recouvertes (ou non) d'une couche de particules hydrophobes. Dans une première partie, on étudie les collisions de gouttes de liquide immiscible. L'asymétrie de tels systèmes repose alors sur le contraste des propriétés des deux liquides : la tension de surface, la viscosité et la densité. Le résultat de ces collisions est une encapsulation totale d'un liquide par un autre ou une encapsulation suivie d'une fragmentation. On s'attache à décrire les régimes observés et à établir des lois permettant de prédire les limites de fragmentation de l'objet obtenu. La seconde partie est consacrée aux interfaces couvertes de particules hydrophobes. Pour ces systèmes, l'asymétrie réside à la fois dans la présence des particules sur une interface et pas sur l'autre et dans le contraste de taille entre les objets étudiés. Ainsi, on considère l'impact entre une petite goutte (recouverte ou non de particules) et une très grosse goutte (recouverte ou non de particules). On caractérise tout d'abord les propriétés mécaniques de ces interfaces via la propagation d'ondes de surface, notamment en terme de tension de surface effective et de module de courbure. Puis, on sonde, dans différentes situations d'impact, la robustesse de ces objets afin d'évaluer la capacité de ces couches particulaires à prévenir la coalescence

This work develops a one dimensional, slab geometry, multigroup collision probability code named Oklo which solves both criticality calculations and fixed source problems. The code uses the classical collision probabilities approach where the first flight collision probabilities are calculated analytically for void, reflected and periodic boundary conditions. The code has been verified against analytical criticality benchmark test sets from Los Alamos National Laboratory, which have been used to verify MCNP amongst other codes. The results from the code show a good agreement with the benchmark test sets for the critical systems presented in this report. The results from the code also match the infinite multiplication factors k and average scalar flux ratios for infinite multiplicative systems from the benchmark test sets. The criticality results and the fixed source results from the Oklo code have been compared with criticality results and fixed source results from a discrete ordinates code and the results for both types of problems show a good agreement with the results from the discrete ordinates code as we increase the N for the discreet ordinates code. / Thesis (M.Sc. Engineering Sciences (Nuclear Engineering))--North-West University, Potchefstroom Campus, 2012.

This work develops a one dimensional, slab geometry, multigroup collision probability code named Oklo which solves both criticality calculations and fixed source problems. The code uses the classical collision probabilities approach where the first flight collision probabilities are calculated analytically for void, reflected and periodic boundary conditions. The code has been verified against analytical criticality benchmark test sets from Los Alamos National Laboratory, which have been used to verify MCNP amongst other codes. The results from the code show a good agreement with the benchmark test sets for the critical systems presented in this report. The results from the code also match the infinite multiplication factors k and average scalar flux ratios for infinite multiplicative systems from the benchmark test sets. The criticality results and the fixed source results from the Oklo code have been compared with criticality results and fixed source results from a discrete ordinates code and the results for both types of problems show a good agreement with the results from the discrete ordinates code as we increase the N for the discreet ordinates code. / Thesis (M.Sc. Engineering Sciences (Nuclear Engineering))--North-West University, Potchefstroom Campus, 2012.

Older adults are at an increased risk for motor-vehicle collisions (MVCs) once distance driven is considered, a finding that is partly attributed to a decline in attention related processes associated with age. MVCs typically occur in highly specific areas, suggesting a role of the complexity of the driving environment contributing to the occurrence of MVCs. The goal of this thesis was to explore the attentional demands of simulated driving events of varying complexity among young, mature and older drivers. In the present studies, attentional demand associated with driving was assessed through the peripheral detection task (PDT), a method in which a stimulus unrelated to the driving task is presented and drivers manually respond immediately upon its detection; latency to respond is recorded. The complexity of the driving environment was operationalized in terms of vehicle handling and of information processing elements. In the first study, inexperienced drivers completed a series simulated driving scenarios that varied according to their information processing and vehicle handling demands. The results showed a reduction in PDT performance at intersections where information processing is increased as well as when handling maneuvers behind a lead vehicle were required. Building on these findings, the second study employed the identical protocol as the first but examined differences in attentional demand between mid-aged and older drivers. The results indicated that when information processing demands were increased through the addition of traffic, and buildings, all participants exhibited greater workload regardless of age. The third study presented young, mid-aged, and older drivers with a simulated driving assessment course and administered several cognitive tasks. The results of the third study supported the hypothesis in that complex driving situations elicited greater attentional demand among drivers of all ages. Older adults showed greater attentional demand in comparison to young and mid-aged adults even after controlling for baseline response time. Older drivers also scored poorer on a global measure of driving safety. The results of this thesis highlight the roles of intrinsic and extrinsic factors involved in safe driving and are discussed in terms of appropriate interventions to improve road safety.

This thesis presents a 3D reconstruction approach to the detection of static obstacles from a single rear view parking camera. Corner features are tracked to estimate the vehicle’s motion and to perform multiview triangulation in order to reconstruct the scene. We model the camera motion as planar motion and use the knowledge of the camera pose to efficiently solve motion parameters. Based on the observed motion, we selected snapshots from which the scene is reconstructed. These snapshots guarantee a sufficient baseline between the images and result in more robust scene modeling. Multiview triangulation of a feature is performed only if the feature obeys the epipolar constraint. Triangulated features are semantically labelled according to their 3D location. Obstacle features are spatially clustered to reduce false detections. Finally, the distance to the nearest obstacle cluster is reported to the driver.

The electric charge distribution of molecules such as H₂ and D₂ is inversion-symmetric so that permanent dipole moments do not exist: such molecules are infrared-inactive. It is therefore interesting that gaseous, liquid, and solid hydrogen and its isotopes actually absorb infrared radiation, for example if gas densities are sufficiently high. The observed absorption arises from electric dipole moments induced by intermolecular interactions. It is of a supermolecular origin, due to binary (or higher-order) molecular complexes that may be transient (i.e., in a collisional encounter) or relatively stable (van-der-Waals molecule). Interaction-induced electric dipoles arise from the same mechanisms that generate the intermolecular forces: exchange forces, dispersion forces, and multipolar induction. Recently the induced dipole and potential energy surfaces of H₂ pairs have been obtained by advanced quantum-chemical calculations. Interaction-induced absorption, more commonly called collision-induced absorption (CIA), by H₂ pairs is an important opacity source in the atmospheres of various types of planets and cool stars, such as late stars, low-mass stars, brown dwarfs, certain white dwarfs, etc., and therefore of special astronomical interest. The emission spectra of cool white dwarf stars differ significantly from the expected blackbody spectra of their cores, mainly due to collision-induced absorption by collisional complexes of hydrogen and helium in the stellar atmospheres. Before proceeding to the frequencies and temperatures of interest it is good to check the new potential energy surface and induced dipole surface in all possible ways by comparison with existing isotopic laboratory measurements. Furthermore, the new potential energy surface is directly compared with previously available, well established intermolecular potential energy surfaces. The electric charge distributions of deuterium and hydrogen are very similar. The new potential energy and induced dipole surfaces were originally obtained to facilitate the computation of the collision-induced absorption of hydrogen. However, by replacing the rotovibrational wavefunctions of H₂ with those of D₂ the surfaces can also be used to calculate the collision-induced absorption of deuterium pairs, thereby probing them further. At the temperature of 298K existing measurements of the collision-induced absorption of D₂--D₂ gas are compared with our quantum scattering calculations in the D₂ fundamental band (approximately 2,500cm⁻¹ to 4,500cm⁻¹). Furthermore, measurements of the collision-induced absorption of deuterium (D₂) in the D₂ first overtone band (about 5,250cm⁻¹ to 7,250cm⁻¹) at 201K are reported. These measurements are compared with ab initio calculations of the absorption spectra. Close agreement of measured and calculated spectra is seen.