The use of fractionated spacecraft systems in on-orbit spacecraft assembly has the potential to provide benefits to both the defense and civil space community. To this end, much research must be conducted to develop and prove the requisite technologies to achieve these benefits. This thesis contributes to that effort by presenting the design and system integration, operating procedures and software development for a prototype three Degrees-Of-Freedom (DOF) Spacecraft Simulator. This simulator will be used in the Proximity Operations Simulator Facility, as part of the Naval Postgraduate School's Spacecraft Robotics Laboratory, to simulate autonomous guidance, navigation and control (GNC) for spacecraft proximity operations and assembly within the framework of the Autonomous Multi-Agent Physically Interacting Spacecraft project. The new spacecraft simulator includes several key enhancements over the previous Autonomous Docking and Spacecraft Servicing Simulator (AUDASS) developed in 2005 including a smaller and more agile structure, reduced air consumption and a Miniature Single-Gimbaled Control-Moment-Gyroscope (MSGCMG) to provide the necessary torque about the rotation axis. The MSGCMG in the simulator is a low-cost, low-mass, easily controlled momentum exchange device with a high torque to required power ratio. Furthermore, it provides the vehicle with high slew-rate capability, a key measure of performance in on orbit spacecraft assembly. Simulation and experimental results are presented for the prototype AMPHIS vehicle, demonstrating a potential slew-rate of 4.8 deg/s for a 30 degree rest-to-rest maneuver. The ultimate goal of this thesis is to provide the design specifications, combined with the necessary documentation and software development, for the prototype vehicle of the testbed for the AMPHIS project. The work conducted in fabricating the prototype vehicle will enable rapid fabrication of two additional vehicles which will provide an essential hardware-in-the-loop capability for experimentation with evolving control algorithms, sensors and mating mechanisms to be used for autonomous spacecraft assembly.

For robotic systems to realise lifelong autonomy they must be able to navigate accurately in changing environments. In this thesis we describe, implement and validate a new approach to the problem of long-term navigation. To begin, we present our stereo visual odometry system which provides highly accurate pose estimation. Our approach combines several techniques found in existing implementations and a recently published image descriptor that simplifies the solution architecture. The performance and versatility of our system is demonstrated through testing on multiple datasets. Equipped with our visual odometry system, we describe a new approach to the problem of lifelong navigation. We learn a model whose complexity varies naturally in accordance with the variation of scene appearance. As the robot repeatedly traverses its workspace, it accumulates distinct visual experiences that, in concert, implicitly represent the scene variation - each experience captures a visual mode. When operating in a previously visited area, we continually try to localise in these previous experiences while simultaneously running the visual odometry. Failure to localise in a sufficient number of prior experiences indicates an insufficient model of the workspace and instigates the laying down of the live image sequence as a new distinct experience. In this way, over time we can capture the typical temporally varying appearance of an environment and the number of experiences required tends to a constant. Although we focus on vision as a primary sensor, the ideas we present here are equally applicable to other sensor modalities. We demonstrate our approach working on a road vehicle operating over a three month period at different times of day, in different weather and lighting conditions.

Purpose: In oncological orthopedics, navigation systems are limited to use in specialized centers, because specific, expensive, software is necessary. To resolve this problem, we present a technique using general spine navigation software to resect tumors located in different segments. Materials and Methods: This technique requires a primary surgery during which screws are inserted in the segment where the bone tumor is; next, a CT scan of the entire segment is used as a guide in a second surgery where a resection is performed under navigation control. We applied this technique in four selected cases. To evaluate the procedure, we considered resolution obtained, quality of the margin and its control. Results: In all cases, 1 mm resolution was obtained; navigation allowed perfect control of the osteotomies, reaching the minimum wide margin when desired. No complications were reported and all patients were free of disease at follow-up (average 25.5 months). Conclusions: This technique allows any bone segment to be recognized by the navigation system thanks to the introduction of screws as landmarks. The minimum number of screws required is four, but the higher the number of screws, the greater the accuracy and resolution. In our experience, five landmarks, placed distant from one another, is a good compromise. Possible disadvantages include the necessity to perform two surgeries and the need of a major surgical exposure; nevertheless, in our opinion, the advantages of better margin control justify the application of this technique in centers where an intraoperative CT scanner, synchronized with a navigation system or a dedicated software for bone tumor removal were not available.

Le but de cette thèse est de proposer et d’évaluer une représentation 3D de collections non structurées de documents visuels. L'objectif est d’accroître le confort visuel, de réduire le temps de recherche et de favoriser la mémorisation des positions des documents, par rapport aux logiciels actuels de gestion de collections d'images. Avec une représentation sous forme d'un cylindre vertical, nous avons conçu 2 visualisations : les photographies sont plaquées sur la face interne ou externe du cylindre, pour implémenter 2 métaphores courantes d’interaction, respectivement locomotion et manipulation. 3 études expérimentales ont été réalisées, avec la même méthodologie : la réalisation par les participants de 2 tâches courantes de recherche visuelle. La première étude est exploratoire, la seconde compare les 2 vues 3D plus finement et la dernière compare les vues 3D à une vue 2D classique. Dans cette dernière étude, l’usage d’un oculomètre permet de comprendre les relations entre les caractéristiques géométriques des vues et les stratégies visuelles des participants. Les participants préfèrent majoritairement les vues 3D. Les performances n'influent pas sur les préférences ; les caractéristiques géométriques statiques ont principalement été mentionnées pour justifier les préférences. Les stratégies de parcours visuel sont davantage influencées par la caractéristique dynamique des vues (i.e., défilement horizontal), ce qui avantage la vue interne. Enfin, les vues 3D garantissent un meilleur confort visuel, en offrant un meilleur guidage du regard (principalement la vue interne) et en minimisant les saccades désordonnées observées dans la vue 2D. / The goal of this thesis is to propose and to assess a 3D representation of large unstructured collections of images. The purpose is to improve comfort of visual search, to reduce search times and to better support memorization of documents’ locations, compared with standard photo browsers. With a representation as a vertical cylinder, we designed two visualizations: photos are plastered on the inner or the outer side of the cylinder, to implement 2 common interaction metaphors, respectively locomotion and manipulation. 3 experimental studies were performed, with the same methodology: subjects performed 2 common types of visual search tasks. The first study is exploratory, the second one compares both 3D views in-depth and the last one compares 3D views with a standard 2D view. In this study, an eye-tracker helps us to understand relations between geometrical features and visual exploration strategies. 3D views are preferred by a majority of participants. Preferences are not influenced by performances; static visual features were mainly used to justified preferences. Visual exploration strategies are mainly influenced by the common dynamic visual feature (i.e., horizontal scrolling), which is an advantage for the inner view. At last, 3D views ensure a better comfort of visual search, with a better gaze guidance (especially the inner view) and by minimizing uncoordinated saccades which were observed in the 2D view.

No description available.

A robótica móvel é uma área de pesquisa que está obtendo grande atenção da comunidade científica. O desenvolvimento de robôs móveis autônomos, que sejam capazes de interagir com o ambiente, aprender e tomar decisões corretas para que suas tarefas sejam executadas com êxito é o maior desafio em robótica móvel. O desenvolvimento destes sistemas inteligentes e autônomos consiste em uma área de pesquisa multidisciplinar considerada recente e extremamente promissora que envolve; por exemplo, inteligência artificial, aprendizado de máquina, estimação estatística e sistemas embarcados. Dentro desse contexto, esse trabalho aborda o problema de navegação e monitoramento de ambientes utilizando robôs móveis. Dada uma representação do ambiente (mapa topológico) e uma lista com urgências de cada uma das regiões do mapa, o robô deve estimar qual o percurso mais eficiente para monitorar esse ambiente. Uma vez que a urgência de cada região não visitada aumenta com o tempo, o trajeto do robô deve se adaptar a essas alterações. Entre as diversas aplicações práticas desse tipo de algoritmo, destaca-se o desenvolvimento de sistemas de segurança móveis inteligentes / The mobile robotics is a research area that has started to get some serious attention from the scientific community. The development of robots able to interact with the environment - to learn and make correct decisions so their tasks are successfully completed - is the biggest challenge in mobile robotics. The development of these intelligent and autonomous systems consists of a multidisciplinary research area considered recent and very promising that involves, for example, artificial intelligence, machine learning, statistical estimation and embedded systems. Within this context, this paper addresses the problem of navigation in dynamic environments with priorities using mobile robots. Given a representation of the environment (topological map) and a list of priorities for each region of the map, the robot must estimate what is the most efficient way to monitor this environment. As the dynamic priority of each region increases with time since the last visit of the robot, its trajectory must adapt to these changes. This approach is similar to the traveling salesman, but a solution that specifically addresses the problem described in this dissertation was not found in the literature. Among the many practical applications of this type of algorithm, we highlight the development of smart mobile security systems

Le travail présenté dans ce mémoire porte sur l’aide à la navigation pour un fauteuil intelligent. Le but est de réaliser un système de reconnaissance et de suivi de trajets. Ce travail s’inscrit dans le projet VAHM qui utilise des fauteuils électriques du commerce sur lesquels ont été ajoutés une série de capteurs et un calculateur. Le VAHM utilise différents comportements (suivi de direction, suivi de mur, évitement d’obstacles,...) pour contrôler le fauteuil et pour réagir à l’environnement. On étudie, en premier lieu, la possibilité d’utiliser l’enchaînement des comportements du fauteuil pour modéliser et reconnaître un trajet en utilisant les MMC. Cette méthode montre ses limites en ce qui concerne la généralisation du modèle à des trajets non appris. Puis l’utilisation des MMC-MD est alors étudiée en intégrant, en plus des comportements, des données provenant des capteurs. Cependant les essais montrent que la reconnaissance est meilleure en utilisant uniquement les données provenant des capteurs. Ceci qui nous a conduit à n’utiliser que les données provenant des capteurs dans la suite du projet pour réaliser la reconnaissance. L’algorithme CONDENSATION est alors utilisé. Il possède de nombreux avantages dont celui de ne pas avoir besoin de modèle complexe puisque un seul trajet de référence est nécessaire pour réaliser la reconnaissance. Ces modèles de trajets sont ensuite mis dans un réseau bayésien pour faciliter les transitions d’un modèle à l’autre. La dernière partie du mémoire présente les résultats de la mise en œuvre de la reconnaissance en conditions réelles sur le fauteuil VAHM 3 / The work presented in this thesis concerns navigation assistance for a smart wheelchair. We propose a new approach based on this behaviour based structure, aiming at assisting the user in a global way. The intelligent system relies on a modelisation of the most frequently used routes and assists the user when navigating by suggesting the next movement when the route has been recognised. The user will not have to change the direction, thus sparing him/her an action. This work is included in the VAHM project that use commercial electric wheelchair which is enhanced by several sensors and a calculator. The VAHM project is using a multi-agent control system. Different behaviours (direction following, wall following, obstacle avoidance ...) are designed react to the environment. At first, we tried to modelize and recognize behaviour sequences with HMM. This method showed its limits to generalize the recognition of unlearned trajectories. Then we used M-HMM (by adding sensor’s data) to achieve recognition. However, our tests show that the recognition is better by using only sensor’s data. These observations have conducted us to use only sensor’s data for the recognition. The CONDENSATION algorithm had been tested to perform recognition from sensor's data. This system earn numerous advantages including that it doesn’t need a complex model since only one reference trajectory is necessary to perform the recognition. Then these trajectory models have been put into a Bayesian network to facilitate transitions between models in order to recognize a long trajectory. The last part of this thesis presents recognition’s results in real conditions on the VAHM 3 wheelchair prototype

People in unfamiliar environments often need navigation guidance to reach a destination. Research has found that compared to outdoors, people tend to lose orientation much more easily within complex buildings, such as university buildings and hospitals. This paper proposes a category-based method to generate landmark-based route instructions to support people's wayfinding activities in unfamiliar indoor environments. Compared to other methods relying on detailed instance-level data about the visual, semantic, and structural characteristics of individual spatial objects, the proposed method relies on commonly available data about categories of spatial objects, which exist in most indoor spatial databases. With this, instructions like "Turn right after the second door, and use the elevator to go to the second floor" can be generated for indoor navigation. A case study with a university campus shows that the method is feasible in generating landmark-based route instructions for indoor navigation. More importantly, compared to metric-based instructions (i.e., the benchmark for indoor navigation), the generated landmark-based instructions can help users to unambiguously identify the correct decision point where a change of direction is needed, as well as offer information for the users to confirm that they are on the right way to the destination.

With the rise in aging population, about 6.8 million American residents are depen- dent on mobility devices for their day to day activity. More than 40% of these users have di?culty in moving the mobility device on their own. These numbers serve as a motivation on developing a system than can help in manipulation with simple muscle activity and localize the mobility device in the user's home in case of medical emergencies. This research is aimed at creating a user interface of Elec- tromyographic Sensor, attached to the forearm, incorporated with present smart wheelchairs and a simple localization technique using ducial markers. The main outcome of the research is a simulator of the smart wheelchair to analyze the results of my research.

"This research introduces the structure and elements of the system used to predict the user's interested location. The combination of DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm and GMM (Gaussian Mixture Model) algorithm is used to find locations where the user usually visits. In addition, the testing result of applying other clustering algorithms such as Gaussian Mixture model, Density Based clustering algorithm and K-means clustering algorithm on actual data are also shown as comparison. With having the knowledge of locations where the user usually visits, Discrete Bayesian Network is generated from the user's time-sequence location data. Combining the Bayesian Network, the user's current location and the time when the user left the other locations, the user's interested location can be predicted."