A distributed virtual environment (DVE) is a shared virtual environment where multiple users at their workstations interact with each other. Some of these systems may support a large number of users, e. g. , massive multi-player online games, and these users may be geographically distributed. An important performance measure in a DVE system is the delay for an update of a user's state (e. g. , his position in the virtual environment) to arrive at the workstations of those users who are affected by the update. This update delay often has a stringent requirement (e. g. , less than 100 ms) in order to ensure interactivity among users. <br /><br /> In designing a DVE system, an important issue is how well the system scales as the number of users increases. In terms of scalability, a promising system architecture is a two-level hierarchical architecture. At the lower level, multiple service facilities (or basic systems) are deployed; each basic system interacts with its assigned users. At the higher level, the various basic systems ensure that their copies of the virtual environment are as consistent as possible. Although this architecture is believed to have good properties with respect to scalability, not much is known about its performance characteristics. <br /><br /> This thesis is concerned with the performance characteristics of the two-level hierarchical architecture. We first investigate the issue of scalability. We obtain analytic results on the workload experienced by the various basic systems as a function of the number of users. Our results provide valuable insights into the scalability of the architecture. We also propose a novel technique to achieve weak consistency among copies of the virtual environment at the various basic systems. Simulation results on the consistency/scalability tradeoff are presented. <br /><br /> We next study the update delay in the two-level hierarchical architecture. The update delay has two main components, namely the delay at the basic system (or server delay) and the network delay. For the server delay, we use a network of queues model where each basic system may have one or more processors. We develop an approximation method to obtain results for the distribution of server delay. Comparisons with simulation show that our approximation method yields accurate results. We also measure the time to process an update on an existing online game server. Our approximate results are then used to characterize the 95th-percentile of the server delay, using the measurement data as input. <br /><br /> As to the network delay, we develop a general network model and obtain analytic results for the network delay distribution. Numerical examples are presented to show the conditions under which geographical distribution of basic systems will lead to an improvement in the network delay. We also develop an efficient heuristic algorithm that can be used to determine the best locations for the basic systems in a network.

As the presence of digital game-based learning increases in United States classrooms, understanding their impact on achievement is critical. Digital games for learning offer many potential benefits, including reducing the number of students trapped in a remediation cycle, a contributor to college dropout. Despite the recognized potential of game-based learning, few researchers have explored the relationships between specific patterns of behaviors and types of digital game-based learning environments. The underlying theory for this study was patterns of gamer behaviors may predict in-game behaviors. Archival, third-party data regarding The Lost Function - Episode 1: Sum of the Forgotten Minds by Advanced Training & Learning Technology, LLC was used in this study. Using 4 case groups at the high school and college levels (n=114), self-reported levels of the 3 patterns of gamer behaviors, gender, and age-band were analyzed using multiple regression to determine relationships to time-on-task in a game-based highly interactive virtual environment, designed for mathematics remediation. While the results were inconclusive, this study supported the existing literature regarding gender differences and the lack of mutual exclusivity in behavior typing. Recommendations include additional research in how the statements used in the 3-factor model may be adjusted to allow for a broader population of game players. The social change implication is that further understanding of the relationship between learner traits and digital learning environment may assist educators that employ digital game-based learning a way to better align learners to the most appropriate digital learning environment, thereby increases their chances at success.

This dissertation investigated the effects of spatialized auditory cues on the development of situation awareness for teams. Based on extant research, it was hypothesized that 3-D spatialized auditory cues can be utilized by teams to develop knowledge about team member location in addition to supporting the usage of team behaviors for developing and maintaining situation awareness. Accordingly, the study examined how situation awareness would be differentially influenced by varying the type of auditory cues incorporated into virtual environment (VE) team training scenarios within the context of a MOUT team task. In general, the results of this study provided partial support for the beneficial effects of 3-D audio cues in facilitating the development of situation awareness and reducing workload. Implications are discussed in the context of design guidance for VE training systems.

Context-sensitive cognitive rehabilitation aims to address the specific deficits of patients by taking into account the unique strengths and weaknesses of each brain-injured individual. However, this approach requires customized assessments and trainings that are difficult to validate, time-consuming or simply unavailable for daily clinical use. Given the currently struggling economy and an increasing number of patients with brain injuries, a feasible and efficient solution for this individualized rehabilitation concept is needed. This dissertation addresses the development and evaluation of a VE-based training and assessment for context-sensitive cognitive rehabilitation. The proposed application is designed to closely resemble real-world places that are relevant to each individual neurological patient. Despite such an ecologically valid approach to rehabilitation, the application also integrates traditional process-specific tasks that offer potential for standardization and collection of normative data across patient populations. Three cognitive tasks (navigation, orientation, spatial memory) have been identified for use in individualized VEs. In three experimental trials the feasibility and validity of the technological implementation and theoretical foundation of these tasks has been assessed. In a fourth trial one of the tasks has been used for the rehabilitation of a brain-injured patient. Based on the results of these studies a workflow for the rapid development of VEs has been established which allows a VR developer to provide clinicians with individualized cognitive tasks. In addition, promising results for the clinical use and validation of the proposed system form the basis for future randomized controlled clinical trials. In conclusion, this dissertation elaborates how context-sensitive and process-specific rehabilitation approaches each offer a unique perspective on cognitive rehabilitation and how combining both through the means of VR technology may offer new opportunities to further this clinical discipline.

Cognitive Map Theory (O’Keefe & Nadel, 1978) posits that spatial behavior can reflect locale or taxon strategies. Only locale strategies depend on cognitive maps, and learning recruited by these strategies is unlike associative learning (e.g., Rescorla & Wagner, 1972; Mackintosh, 1975), which is prevalent in the taxon system. Associative learning phenomena like the blocking effect (Kamin, 1969) should therefore not occur during acquisition of cognitive maps. Contrary to this prediction, blocking effects have been demonstrated in spatial learning (e.g., Biegler & Morris, 1999; Chamizo, Sterio, & Mackintosh, 1985; Hamilton & Sutherland, 1999), and have been generally interpreted as evidence against cognitive map theory. Here we provide evidence suggesting that taxon and not locale strategies were promoted in these experiments, and we ask which factors determine whether taxon or locale strategies control spatial behavior in a computer-implementation of a widely used spatial task (Morris Water Maze; Morris, 1981). We isolated two factors relevant for spatial strategy choice in human spatial learning that are both related to the individual’s preexisting knowledge, namely conceptual knowledge about the distal cues, and knowledge about the task affordances. The blocking effect was used as an index for locale or taxon learning. We found that taxon strategies were more likely for abstract distal cues, while concrete cues promoted locale strategies – blocking was present for the former, but not the latter. When subjects were aware that the distal cues predicted locations, locale, and not taxon strategies were recruited, such that blocking was not observed. Spatial strategy choice appears to be largely driven by interindividual differences, and can therefore not be easily predicted a priori. Our findings cannot be explained by associative learning theories, but provide strong support for cognitive map theory and the position that multiple behavioral systems exist in the brain.

Approved for public release; distribution in unlimited. / This thesis examines the effectiveness of the Indoor Simulated Marksmanship Trainer (ISMT) as a tool to train shooters in the fundamentals of marksmanship. Key concepts explored in the research are verification of skills transfer resulting from practice and the predictive value of simulated performance to proficiency at real task performance. There was no statistical difference in the scores of recruits trained in the ISMT versus a control group that was not trained in the ISMT. Scores on simulated firing were not a strong predictor of live fire performance. In a second experiment subjects were evaluated on their proficiency and improvement during un-coached practice at the task of simulated precision fire on a target at a simulated known distance of 300 yards from the shooters. After comparable amounts of practice in the ISMT, subjects who had not previously received formal marksmanship training failed to demonstrate levels of proficiency comparable to those subjects who had previously received formal marksmanship training in the military. Consequently, the research found no evidence to suggest the ISMT qualifies as a black box training apparatus capable of imparting skill through practice without the added presence of expert instruction or an existing knowledge of marksmanship techniques. / Major, United States Marine Corps

Alzheimer's disease (AD) is neurodegenerative disorder that causes memory loss and cognitive dysfunction. It affects one in five people over the age of 80 and is distressing for both sufferers and their families. A transitional stage between normal ageing and dementia including AD is termed a mild cognitive impairment (MCI). Recent studies have shown that people with MCI may convert to AD over time although not all MCI cases progress to AD. Much research is now focussing on early detection of AD and diagnosing an MCI that will progress to AD to allow prompt treatment and disease management before the neurons degenerate to a stage beyond repair. Hence, the ability to obtain a method of identifying MCI is of great importance. Virtual reality plays an important role in healthcare and offers opportunities for detection of MCI. There are various studies that have focused on detection of early AD using virtual environments, although results remain limited. One significant drawback of these studies has been their limited capacity to incorporate levels of difficulty to challenge users' capability. Furthermore, at best, these studies have only been able to discriminate between early AD and healthy elderly with about 80% of overall accuracy. As a result, a novel virtual simulation called Virtual Reality for Early Detection of Alzheimer's Disease (VREAD) was developed. VREAD is a quick, easy and friendly tool that aims to investigate cognitive functioning in a group of healthy elderly participants and those with MCI. It focuses on the task of following a route, since Topographical Disorientation (TD) is common in AD. An investigation was set up with two cohorts: non-elderly and elderly participants. The findings with regard to the non-elderly are important as they represent a first step towards implementation with elderly people. The results with elderly participants indicate that this simulation based assessment could provide a method for the detection of MCI since significant correlations between the virtual simulation and existing neuropsychological tests were found. In addition, the results proved that VREAD is comparable with well-known neuropsychological tests, such as Cambridge Neuropsychological Automated Test Battery, Paired Associate Learning (CANTAB PAL) and Graded Naming Test (GNT). Furthermore, analysis through the use of machine learning techniques with regard to the prediction of MCI also obtained encouraging results. This novel simulation was able to predict with about 90% overall accuracy using weighting function proposed to discriminate between MCI and healthy elderly.

Visual object detection is a popular computer vision task that has been intensively investigated using deep learning on real data. However, data from virtual environments have not received the same attention. A virtual environment enables generating data for locations that are not easily reachable for data collection, e.g. aerial environments. In this thesis, we study the problem of object detection in virtual environments, more specifically an aerial virtual environment. We use a simulator, to generate a synthetic data set of 16 different types of vehicles captured from an airplane. To study the performance of existing methods in virtual environments, we train and evaluate two state-of-the-art detectors on the generated data set. Experiments show that both detectors, You Only Look Once version 3 (YOLOv3) and Single Shot MultiBox Detector (SSD), reach similar performance quality as previously presented in the literature on real data sets. In addition, we investigate different fusion techniques between detectors which were trained on two different subsets of the dataset, in this case a subset which has cars with fixed colors and a dataset which has cars with varying colors. Experiments show that it is possible to train multiple instances of the detector on different subsets of the data set, and combine these detectors in order to boost the performance.

La réalité virtuelle, jusqu’alors réservée à des professionnels, poursuit son développement avec la mise sur le marché du grand public, de périphériques (e.g., Oculus) et d’applications (e.g., Star Wars : Jedi Challenges) dit « immersifs ». Pour autant, la littérature ne recense que peu de modèles et d’outils faisant référence à l’expérience utilisateur ou User eXperience (UX) en Environnement Virtuel Immersif (EVI), et les quelques modèles et outils existants ne considèrent pas l’ensemble des composants de l’UX. Cette thèse vise à proposer et valider un modèle holistique d’UX adapté aux EVI (objectif 1), à construire et valider un questionnaire d’UX opérationnalisant le modèle (objectif 2) et à utiliser le questionnaire validé dans le cadre d’une entreprise internationale, éditrice d’applications immersives dans le domaine ludo-éducatif (objectif 3). Le modèle « Immersive Virtual Experience Model » (IVEM) que nous proposons est multidimensionnel. En effet, il regroupe non seulement les facteurs intrinsèques à l’EVI (i.e., largeur du champ visuel, fréquence d’image, niveau d’interactivité, feedback du contenu 3D) et les facteurs propres à l’utilisateur (i.e., expérience précédente) impactant l’UX, mais aussi les composants caractérisant l’UX référencés dans la littérature (i.e., présence, immersion, engagement, flow, utilisabilité, compétences, émotions, conséquences de l’expérience, jugement, adoption de la technologie). Le modèle IVEM validé via cinq études expérimentales confirme les effets des facteurs sur les composants déjà identifiés dans la littérature et fait apparaître de nouvelles relations entre les composants jusqu’alors non démontrées. Notre questionnaire « Immersive Virtual Experience Questionnaire » (IVEQ) opérationnalise les dix composants de notre modèle, à travers des items provenant de questionnaires validés et répertoriés dans la littérature. Le questionnaire IVEQ validé à travers une étude expérimentale conduite sur 152 participants, est composé de 68 items : 9 items de présence, 3 items d’engagement, 5 items d’immersion, 10 items de flow, 11 items d’émotion, 6 items de compétence, 9 items de jugement, 8 items relatifs aux conséquences de l’expérience et 7 items concernent l’adoption de la technologie. Le questionnaire validé a été expérimenté chez un éditeur de solutions immersives, l’entreprise EON Reality, pour assister la conception de l’application ludo-éducative King Tut VR2, qui mobilise deux périphériques de réalité virtuelle : le casque VRONE et la salle immersive iCube. / Virtual Reality has until recently been mostly dedicated to experts. It is now pursuing its development towards the general public with the launch of new devices (e.g., Oculus) and immersive applications (e.g., Star Wars: Jedi Challenges). However, only few User eXperience (UX) models and tools are identified for Immersive Virtual Environments (IVE) in the scientific literature. And these few existing models and tools do not consider all the relevant UX components. This thesis aims to propose and validate a holistic UX model adjusted to IVE (objective 1), to build and validate a UX questionnaire that operationalizes the model (objective 2) and to use the validated questionnaire in an international company context, specifically an immersive applications editor in the field of edutainment (objective 3). The « Immersive Virtual Experience Model » (IVEM) model that we propose is multidimensional. Indeed, it combines not only IVE intrinsic factors (i.e., field of view, framerate, interactive level, 3D content feedback) and user intrinsic factors (i.e., previous experience) but also components from the scientific literature that characterizes the UX (i.e., presence, immersion, engagement, flow, usability, skills, emotions, experience consequences, judgement, technology adoption). The IVEM model was validated via five experimental studies. It confirms the effects of factors on components already identifies in the scientific literature as well as revealing new relations between components that were not demonstrated until then. Our « Immersive Virtual Experience Questionnaire » (IVEQ) questionnaire operationalizes the ten components of our model through items originating from validated and reported questionnaires in the scientific literature. The IVEQ questionnaire was validated through an experimental study conducted on 152 participants. It now comprises 68 items: 9 items for presence, 3 items for engagement, 5 items for immersion, 10 items for flow, 11 items for emotion, 6 items for skill, 9 items for judgement, 8 items related to experience consequences and 7 items refers to technology adoption. The validated questionnaire was experimented with an immersive solutions editor, specifically the EON Reality company, in order to assist the design of the edutainment application King Tut VR2 that can be used via two virtual reality devices: the VRONE headset and the immersive room iCube.

La communication haptique interpersonnelle est communément utilisée pour se saluer, se féliciter, ou de manière indirecte pour manipuler des objets lourds. Pourtant, la modalité haptique est absente des systèmes de communication à distance usuels. Les progrès récents de l'informatique et des réseaux de communication ont permis l'émergence d'environnements virtuels collaboratifs offrant un retour haptique aux utilisateurs. Les environnements virtuels collaboratifs constituent une solution prometteuse pour faciliter le travail collaboratif à distance. Les contextes applicatifs sont nombreux : entrainement militaire, assemblage mécanique, manipulation moléculaire, conception assistée par ordinateur.Toutefois, les utilisateurs d'environnements virtuels collaboratifs sont confrontés à deux types de problématiques : celles liées aux environnements virtuels 3D et celles liées à la collaboration à distance. En effet, il est difficile pour un utilisateur peu expérimenté d'interagir avec un environnement virtuel 3D. Se déplacer et manipuler des objets dans ce type d'environnement nécessite une bonne représentation mentale de l'espace.De plus, la collaboration à distance limite les possibilités et la qualité de la communication. Les informations dont disposent les utilisateurs sur leurs partenaires sont donc réduites. Dans ce contexte, la coordination des actions entre les utilisateurs devient plus complexe.L'objectif de ces travaux de thèse est donc de proposer des méthodes d'interaction visio-haptique pour faciliter la coordination et augmenter l'attention mutuelle entre les utilisateurs. Via l'amélioration de la coordination et de l'attention inter-utilisateurs, une augmentation des performances et de l'efficacité du travail collaboratif est escomptée.Trois types d'interactions fondamentales sont abordées dans un contexte collaboratif : la coordination des sélections, le paramétrage des sélections et la manipulation collaborative d'objet. Pour chacune des tâches d'interactions précédemment citées, des méthodes d'interaction ainsi que des outils de coordination et de communication ont été proposé et évalué.Ces travaux de thèse présentent les points forts mais aussi les défauts des méthodes d'interaction collaboratives fortement couplées. Les résultats obtenus montrent également l'importance de la communication haptique dans la collaboration à distance. / Haptic interpersonal communication is commonly used to salute, to congratulate, or indirectly to manipulate heavy objects. However, the haptic feedback is missing from the common communication systems. Recent developments in telecommunications and computer capabilities have made possible the use of haptic feedback in collaborative virtual environments. Collaborative virtual environments are a promising solution to facilitate remote work collaboration. They have many applications like: military training, engineering design, e-learning or video games.Nevertheless, users of collaborative virtual environments have to face several issues. First, interacting with a virtual environment is not a straightforward task. Navigate and manipulate objects in a virtual environment requires a good mental representation of space. Secondly, remote collaboration limits the communication quality because even the best video conference system can't support all the richness of a face to face communication. Information available on the partners are reduced thus coordinate actions between users become a complex task.The aim of this thesis is to propose visual haptic interactions to simplify coordination and increase mutual awareness between the users. Through the improvement of coordination and mutual awareness an increase of team performance and work efficiency is expected.Three kind of fundamental 3D interactions are addressed in collaborative context: the coordination of selections, the setting of selections and the manipulation of objects. For each of these three of fundamental interactions, several methods of interaction and tools of coordination are proposed and assessed.The findings of this thesis highlight the strengths and the weaknesses of the closely coupled method of interactions. The results of the experiments also confirm the importance of haptic communication in remote collaboration.