Dissertação (mestrado) - Universidade Federal de Santa Catarina, Centro Tecnológico. Programa de Pós-Graduação em Engenharia Elétrica. / Made available in DSpace on 2013-06-25T22:28:40Z (GMT). No. of bitstreams: 1 313502.pdf: 2548956 bytes, checksum: 09e3d7288119541a1642c771da8cd879 (MD5) / Um dos problemas que envolvem as soluções para a mobilidade de robôs móveis terrestres é estimar a posição do robô com precisão juntamente com a exploração do ambiente, mapeando-o corretamente (SLAM - Simultaneous Localization and Mapping - Localização e Mapeamento Simultâneo). Embora vários algoritmos tenham sido desenvolvidos nos últimos anos, exigindo uma carga de cálculo computacional cada vez maior dos robôs,, estes estão susceptíveis a um mau desempenho quando os sensores apresentam ruídos, quando há problemas nos atuadores, variáveis não modeladas ou em virtude de algum imprevisto momentâneo no ambiente. A proposta deste trabalho é programar um SLAM para robôs móveis interligando-o a uma combinação de sensores inerciais com sensores de odometria através de uma técnica de fusão de sensores conhecida como filtro de Kalman Estendido, para reduzir a incerteza na estimação da posição e melhorar o desempenho do SLAM. Por consequência, o custo computacional é reduzido. O trabalho foi estruturado iniciando por uma revisão a respeito dos conceitos básicos de sensoriamento, a fim de contextualizar o problema e apresentar as nomenclaturas e termos utilizados. A seguir foram abordadas as técnicas de fusão de dados, as representações do robô e do ambiente, as técnicas de mapeamento e exploração e as diversas técnicas de navegação que podem ser utilizadas, para ambientes conhecidos epara ambientes desconhecidos. Essas informações são importantes para um melhor entendimento do problema, de como representá-lo e de como se pode avaliar os resultados obtidos. Na sequência é apresentado o SLAM, destacando as principais técnicas e em detalhes o Grid Based FastSLAM. É demonstrado através de simulações que quanto maior as incertezas sobre a posição do robô, um número maior de partículas é necessário para manter a qualidade do mapa gerado, e como cada partícula possui um mapa associado a si, o custo computacional é consideravelmente aumentado. Outro aspecto analisado foi o impacto na escolha da covariância associada à transição de estados, propondo a utilização da covariância inerente ao cálculo da fusão de sensores como parâmetro de refinamento no SLAM.<br> / Abstract : One of the problems in solutions involving land mobile robots is the estimation of the robot position with precision and at the same time, explore the environment and mapping it correctly (SLAM - Simultaneous Localization and Mapping ). Several algorithms were developed in the last years, demanding large computational resources in robots and even so, these may have a bad performance in cases of sensors having noises, problems in actuators, not modeled variables or when there is something in the environment that wasn't expected. This dissertation proposal is to program a SLAM algorithm for mobile robots and connect it with a sensor data fusion, between odometry and inertial sensors, using the Extended Kalman Filter, achieving a reduction of the position uncertainty and improving the SLAM performance, also reducing the need of computational resources. This work begins with a revision of concepts of robot sensors, needed to understand later algorithms and nomenclatures. In the following items it is described the sensor fusion techniques, the robot localization problem, the map and robot representation alternatives, and the navigation problems for explored and non-explored environments. These information are important for a better understanding of the problem, on how represent it and how to evaluate the obtained results. After this introduction, it's described some SLAM algorithms, featuring in details the Grid Based FastSLAM. It's demonstrated by simulations that as high uncertainty about robot position, as large are the number of particles needed to maintain the generated map quality. This implies in a large computational cost, thus improving the uncertainty with sensor data fusion makes the robot work with less particles. It is also showed that choosing the right covariance in robot transition model is very important and finding a way to connect the covariance of sensor data fusion with SLAM can improve performance even more.

CoordenaÃÃo de AperfeiÃoamento de Pessoal de NÃvel Superior / CoordenaÃÃo de AperfeÃoamento de Pessoal de NÃvel Superior / Cette dissertation a pour but de faire des recherches sur la performance dans la cantoria du Nordeste brÃsilien et sur le slam, genre poÃtique aussi populaire, mais dâapparition plus rÃcente, crÃà aux Ãtats-Unis dans les annÃes 1980 et de forte influence aujourdâhui en France. Certaines formes de la littÃrature orale populaire du Nordeste ont des caractÃristiques trÃs proches de celles du slam, dont on peut citer lâimportance du corps et de son Ãmanation, la voix, ce qui est dÃterminant pour lâattraction et fidÃlisation des spectateurs, soit aux festivals de repente du BrÃsil, soit aux tournois de slam de France. Entre les lignes gÃnÃrales qui permettent le rapprochement des deux genres, on peut observer le jeu interactif avec le public, qui se prÃsente comme une piÃce fondamentale dans la production poÃtique ; les aspects relatifs aux formes du texte, comme la mÃtrique des vers, les rimes et le rythme ; le caractÃre ritualiste de ses formes ; entre autres ÃlÃments de significative discussion. Le slam prÃsente aussi quelques diffÃrences par rapport à la cantoria, comme le caractÃre primordialement urbain et la prÃsence dÃterminante de lâÃcriture. Ces facteurs exposent des particularitÃs intÃressantes aux deux pratiques et composent une partie des prÃoccupations que dÃveloppe cette dissertation. Tandis que le slameur, en gÃnÃral, est scolarisÃ, formà dans un milieu urbain et compte nÃcessairement sur lâÃcriture au moment de sa production, le traditionnel cantador brÃsilien nâen a gÃnÃralement pas besoin pour sa crÃation poÃtique : ce poÃte crÃe en rÃcitant, il mÃmorise, reproduit, modifie en fonction de la situation, improvise des repentes, joue avec le public et avec son rival de cantoria, enfin, il formule et articule sa composition au moment mÃme de la crÃation ou de la prÃsentation. Lâanalyse dÃveloppÃe dans ce travail part de la notion de performance traitÃe par Paul Zumthor. Pour cela, nous proposons de prÃsenter les deux genres parallÃlement, tout en considÃrant leurs indices dâoralità et en cherchant des analogies en ce qui concerne lâaspect circonstanciel de connivence avec le public, le caractÃre scÃnique de ces pratiques, la mesure des vers, les thÃmatiques, le public-cible et les particuliaritÃs trouvÃes dans les deux contextes sociaux. / A presente dissertaÃÃo intenciona investigar a performance na cantoria nordestina brasileira e no slam, gÃnero poÃtico tambÃm popular, mas de surgimento mais recente, criado nos Estados Unidos na dÃcada de 1980 e de forte influÃncia hoje na FranÃa. Algumas formas da literatura oral popular do Nordeste brasileiro tÃm caracterÃsticas muito prÃximas Ãquelas do slam, entre as quais, ressalta-se a importÃncia do corpo e de sua emanaÃÃo, a voz, o que à determinante para a atraÃÃo e fidelizaÃÃo dos espectadores, seja nos festivais de repente do Brasil, seja nos torneios de slam na FranÃa. Entre as linhas gerais que permitem a aproximaÃÃo dos dois gÃneros, podem-se ressaltar o jogo interativo com o pÃblico, que se mostra peÃa fundamental na produÃÃo poÃtica; os aspectos relativos Ãs formas do texto, como a mÃtrica dos versos, as rimas e o ritmo; o teor ritualÃstico das duas formas; entre outros elementos de relevante discussÃo. O slam encontra tambÃm algumas diferenÃas em relaÃÃo à cantoria, como o teor primordialmente urbano e a presenÃa determinante da escrita. Esses fatores expÃem peculiaridades interessantes nas duas prÃticas e compÃem parte das preocupaÃÃes com que se depara esta dissertaÃÃo. Enquanto o slameur, aquele que produz o slam, costuma ter escolaridade, ser formado em um meio urbano e contar necessariamente com a escrita no momento da produÃÃo do slam, o tradicional cantador nordestino brasileiro nÃo costuma precisar dela para a sua criaÃÃo poÃtica: esse poeta cria recitando, memoriza, reproduz, modifica dependendo da situaÃÃo, improvisa repentes, joga com o pÃblico e com seu companheiro de cantoria, enfim, molda e articula sua composiÃÃo no momento mesmo da criaÃÃo ou da apresentaÃÃo. A anÃlise a ser desenvolvida parte da noÃÃo de performance utilizada por Paul Zumthor. Para tanto, propÃe-se apresentar os dois gÃneros paralelamente, levando em consideraÃÃo os seus Ãndices de oralidade e buscando analogias no que concerne ao aspecto circunstancial e coincidente com a presenÃa do pÃblico, ao teor cÃnico dessas prÃticas, à mÃtrica dos versos, Ãs temÃticas, ao pÃblico-alvo e Ãs peculiaridades encontradas em ambos os contextos sociais.

Magíster en Ciencias de la Ingeniería, Mención Eléctrica. Ingeniero Civil Eléctrico / Este trabajo presenta un sistema de Localización y Mapeo Simultáneo (SLAM) de robots humanoides basado en la fusión del sistema ORB-SLAM con información propioceptiva. El objetivo es desarrollar una metodología que permita fusionar ambas informaciones de manera consistente y obtener una auto-localización métrica validada con datos de un robot real. De modo de lograr este objetivo, se propuso una solución que contempla tres objetivos específicos: (1) desarrollar una plataforma de hardware y software que extienda las capacidades computacionales limitadas del robot NAO; (2) proponer una metodología teórica y práctica que permita fusionar la información de odometría propioceptiva del robot dentro de los procesos ejecutados por el sistema ORB-SLAM y (3) desarrollar un sistema de estimación de odometría netamente propioceptivo para el robot NAO. Para la primer aparte se desarrolló una plataforma novedosa, llamada NAO Backpack, que se sujeta al NAO como una mochila y permite cargar un computador de tamaño reducido con mayor poder de cómputo que el NAO. El desarrollo además involucró el software que realiza la comunicación entre ambos con el popular framework de robótica ROS (Robot Operating System). Para la segunda aparte se desarrolló una completa metodología que interviene varias etapas del sistema ORB-SLAM, incluyendo un thread adicional que fusiona la información propioceptiva del robot y genera mediciones de odometría que son propagadas al resto del sistema. Adicionalmente se presentó una solución para obtener la escala típicamente desconocida en sistemas de SLAM visual, se modificaron diversas etapas de las optimizaciones realizadas por el sistema original y se presentaron estrategias para mantener una hipótesis de auto-localización continua incluso en casos de fallas de la localización visual. La última parte contempló el desarrollo de un sistema de odometría basado en un two-state implicit filtering (TSIF), que fusionaba sensores internos como unidades inerciales y encoders, además de considerar otros supuestos sobre la cinemática del robot. El sistema fue validado con datos reales registrados utilizando un preciso sistema de captura de movimiento OptiTrack en ambientes cerrados, así como también en pruebas cualitativas desarrolladas en laboratorios del Departamento de Ingeniería Eléctrica. Los resultados demuestran la efectividad del enfoque presentado para localizar el robot de manera consistente y robusta en lugares desconocidos mediante la fusión de información visual y odométrica.

Tesis para optar al grado de Magíster en Ciencias de la Ingeniería, Mención Eléctrica / La extracción de características es una tarea crítica en la localización y mapeo simultáneo o Simultaneous Localization and Mapping (SLAM) basado en características, que es uno de los problemas más importantes de la comunidad robótica. Un algoritmo que resuelve SLAM utilizando características basadas en LiDAR es el algoritmo LiDAR Odometry and Mapping (LOAM). Este algoritmo se considera actualmente como el mejor algoritmo SLAM según el Benchmark KITTI. El algoritmo LOAM resuelve el problema de SLAM a través de un enfoque de emparejamiento de características y su algoritmo de extracción de características detecta las características clasifican los puntos de una nube de puntos como planos o agudos. Esta clasificación resulta de una ecuación que define el nivel de suavidad para cada punto. Sin embargo, esta ecuación no considera el ruido de rango del sensor. Por lo tanto, si el ruido de rango del LiDAR es alto, el extractor de características de LOAM podría confundir los puntos planos y agudos, lo que provocaría que la tarea de emparejamiento de características falle. Esta tesis propone el reemplazo del algoritmo de extracción de características del LOAM original por el algoritmo Curvature Scale Space (CSS). La elección de este algoritmo se realizó después de estudiar varios extractores de características en la literatura. El algoritmo CSS puede mejorar potencialmente la tarea de extracción de características en entornos ruidosos debido a sus diversos niveles de suavizado Gaussiano. La sustitución del extractor de características original de LOAM por el algoritmo CSS se logró mediante la adaptación del algoritmo CSS al Velodyne VLP-16 3D LiDAR. El extractor de características de LOAM y el extractor de características de CSS se probaron y compararon con datos reales y simulados, incluido el dataset KITTI utilizando las métricas Optimal Sub-Pattern Assignment (OSPA) y Absolute Trajectory Error (ATE). Para todos estos datasets, el rendimiento de extracción de características de CSS fue mejor que el del algoritmo LOAM en términos de métricas OSPA y ATE.

This master thesis is about modifying a certain monocular visual SLAM algorithm to address some of its limitations. The SLAM algorithm is not robust to quick camera motions and input images in which there are few visible features. A second camera and an inertial measurement unit was added to the hardware. Then a method for selecting the appropriate camera for tracking depending on the estimated number of features was implemented to solve the issue of few features. Experiments and results show that this method works well for slow motions. A gyrometer threshold method along with a motion model to solve the issue of quick motions was implemented and reviewed in this thesis. / Detta examensarbete handlar om att ta itu med några begränsningar som en viss monokulär visuell SLAM-algoritm har. SLAM-algoritmen är inte robust mot snabba kamerarörelser och indatabilder som innehåller få karaktärsdrag. Genom att introducera en extra kamera, en accelerometer och en gyrometer, behandlas dessa problem i denna rapport. En metod för att välja kamera att hämta indatabilder från, baserat på det skattade antalet karaktärsdrag i respektive bild implementerades. Denna metod är tänkt att lösa problemet då indatabilder har få karaktärsdrag. Experiment visar att denna metod fungerar bra för långsamma rörelser. En metod som jämför gyrometerdata med ett tröskelvärde tillsammans med en rörelsemodell implementerades för att lösa problemen vid snabb rörelse. Dessa metoder undersöks och diskuteras i rapporten.

Sveriges skogsindustri och kommunala reningsverk producerar i dagsläget cirka 600 000 ton torrsubstrat slam och hur det hanteras inom de olika branscherna är olika. Det här examensarbetet kartlägger hur slamsituationen ser ut inom dessa branscher med avseende på slamhantering. Problem- och målformulering för arbetet är att med en driftfallstudie undersöka hur skjuvspänningen (τ) och viskositeten (υ) ändras och hur det påverkar det totala effektbehovet för en slampump. En livslängdsanalys för tre olika slampumpar har även gjorts för att få en uppfattning om vilka faktorer som är de stora utgiftsposterna under en hel livstid för en slampump där åtta olika driftfall studeras för varje slampump.   Sveriges skogsindustri och VA-bransch producerar stora mängder slam och hanteras olika på grund av infrastruktur, geografiskt läge och certifikat, till exempel REVAQ. Inom skogsindustrin är det vanligast att förbränna det slam som produceras för kvittblivning och VA-branschen återför bildat slam till jordbruket för att ta tillvara på växtnäringen i slam. Slam inom VA-branschen behandlas i rötkammare för utvinning av biogas, stabilisera och få ett mer homogent slam och för att reducera slamvolymen.   Slam är generellt en icke-newtonsk fluid med skjuvförtunnande egenskaper som vid högre skjuvningshastigheter minskar sin viskositet. Slam kan även vara newtonsk vid låga koncentrationer. En minskad viskositet innebär ett lägre tryckfall i rörledningar och teoretiskt sett kan effektbehovet för pumpenheten minskas.   Driftfallstudien genomfördes med vatten som referens och stärkelse som testfluid som är en icke-newtonsk fluid. Driftfallstudien gjordes på Innventias egen flödeslinga som bestod utav en centrifugalpump med patenterad N-teknik från Xylem. Kalibrering av samtliga tryckgivare gjordes innan testerna och insamling av data gjordes med LabView. Kompletterande UVP-mätningar (Ultrasonic Velocity Profiler) togs för att ta fram hastighetsprofiler med stärkelse för att bekräfta inom vilket område för skjuvningshastigheten flödet går från laminärt till turbulent flöde.   Resultat från driftfallstudien visar att det inte finns några indikationer på ändring av effektbehov för pumpenhet vid höga skjuvningshastigheter. UVP-mätningar tillsammans med framtagen flytkurva visar att det går från laminärt till turbulent flöde vid ungefär 100-300 [s-1]. Störst vinst av minskad viskositet är vid skjuvningshastigheten 200 [s-1], för högre skjuvningshastighet är viskositeten konstant. Kostnadsanalysen visar att under en livstid för en slampump blir energikostnaderna mycket påtagliga. Ett exempel för slampump N3301 visar att 90 % av den totala livslängdskostnaden är energikostnader, maximal effekt för slampump N3301 är 70 kW. / Sweden's forestry industry and municipal wastewater treatment plants produces in the current situation around 600 000 ton of total solids and how it is handled in various sectors are different. This thesis surveys how sludge situation is for these industries with focus on sludge management. Problem and goal setting for the thesis is that with a case study examine how shear stress (τ) and the viscosity (υ) change and how this affecting the total energy demands for sludge pump unit. A lifetime analysis for three sludge pumps have also been made to obtain a perception of what factors are the major items of high costs in a lifecycle for a sludge pump, where eight different operating modes are studied for each sludge pump.   Sweden's forest industry and water and wastewater industry produces large amounts of sludge and treated differently because of infrastructure, geographic location and certificates, for example REVAQ. The forest industry is it most common to incinerate the sludge for disposal and for wastewater industry returns sludge for agriculture to take advantage of plant nutrients. Sludge in municipal wastewater treatment plants digesting sludge for production of biogas, stabilize and obtain a more homogeneous sludge and to reduce the sludge volume.   Sludge is generally a non-Newtonian fluid with shear-thinning properties that for higher shear rates decrease its viscosity. Sludge can also be Newtonian at low concentrations. A reduced viscosity means a lower pressure drop in pipes and theoretically the energy demand for a sludge pump can be reduced.   A Case study was conducted with water as a reference and starch for test fluid, which is a non-Newtonian fluid. The case study was done at Innventia´s flow loop consisted of a centrifugal pump with patented N-technology from Xylem. Calibration of all pressure sensors was made before the tests and data collection was done with LabView. Additional UVP (Ultrasonic Velocity Profiler) measurements were made to generate velocity profiles of starch to confirm the field in which the shear flow goes from laminar to turbulent flow.   Results from operations case study show that there are no indications of change in energy demand for the pump unit at high shear rates. UVP measurements together with flow curve shows that it goes from laminar to turbulent flow at 100-300 [s-1]. Biggest change of reduced viscosity is for shear rate up to 200 [s-1], for higher shear rates the viscosity is constant. Cost analysis shows that over a lifetime for sludge pump is energy cost intensive. An example for sludge pump N3301 shows that 90 % of total lifetime costs is for energy only, maximum effect for sludge pump N3301 is 70 kW.

Die dreidimensionale Kartierung der Umgebung spielt speziell in der Robotik eine große Rolle und ist Grundlage für nahezu alle Aufgaben, die eine nicht rein reaktive Interaktion mit dieser Umgebung darstellen. Die vorliegende Arbeit beschreibt den Weg zu solchen Karten. Angefangen bei der reinen (2D-)Lokalisierung eines mobilen Roboters, als erster, fundamentaler Schritt in Richtung autonomer Exploration und Kartierung, beschreibt der Text die Registrierung von Scans zur automatischen, effizienten Generierung von 3D-Karten und gleichzeitiger Lokalisierung in sechs Freiheitsgraden (SLAM-Problem). Es folgen Lösungsstrategien für den Umgang mit akkumulierten Fehlern gerade bei großen explorierten Gebieten: Eine GraphSLAM-Variante liefert global konsistente Karten, optional unterstützt durch eine echtzeitfähige Heuristik zur online-Schleifenoptimierung. Den Abschluss bildet ein alternativer Lokalisierungsansatz zur 3D-Kartierung mittels kooperativ agierenden Robotern.

The objective of the work described in this thesis is the development of a configuration estimation scheme for quasi soft-bodied robots, with the end goal being accurate soft-robot proprioception to enable robotic manipulation of unknown external loads. The first chapter introduces the problem and state-of-the-art modeling methods. The second chapter presents work we did building on previous research on a novel sensing scheme for soft-bodied robotic configuration estimation. The third chapter discusses the development of a geometric shape-sensing model based on overlapping tendon-length measurements. This model is geometrically exact for a body composed of two constant curvature segments in bending. The next chapter discusses our implemention of this model on a real world system and tested in two and three dimensions. We estimated the shape of a 215 mm long robotic segment with less then 3 mm of median error for a set of 50 configurations causing motion in two dimensions and approximately 27 mm of median translational error in three dimensions. The final chapter draws conclusions and proposes future work to allow for full robotic proprioception.

Yes / While Minimally Invasive Surgery (MIS) offers considerable benefits to patients, it also imposes big challenges on a surgeon's performance due to well-known issues and restrictions associated with the field of view (FOV), hand-eye misalignment and disorientation, as well as the lack of stereoscopic depth perception in monocular endoscopy. Augmented Reality (AR) technology can help to overcome these limitations by augmenting the real scene with annotations, labels, tumour measurements or even a 3D reconstruction of anatomy structures at the target surgical locations. However, previous research attempts of using AR technology in monocular MIS surgical scenes have been mainly focused on the information overlay without addressing correct spatial calibrations, which could lead to incorrect localization of annotations and labels, and inaccurate depth cues and tumour measurements. In this paper, we present a novel intra-operative dense surface reconstruction framework that is capable of providing geometry information from only monocular MIS videos for geometry-aware AR applications such as site measurements and depth cues. We address a number of compelling issues in augmenting a scene for a monocular MIS environment, such as drifting and inaccurate planar mapping. Methods A state-of-the-art Simultaneous Localization And Mapping (SLAM) algorithm used in robotics has been extended to deal with monocular MIS surgical scenes for reliable endoscopic camera tracking and salient point mapping. A robust global 3D surface reconstruction framework has been developed for building a dense surface using only unorganized sparse point clouds extracted from the SLAM. The 3D surface reconstruction framework employs the Moving Least Squares (MLS) smoothing algorithm and the Poisson surface reconstruction framework for real time processing of the point clouds data set. Finally, the 3D geometric information of the surgical scene allows better understanding and accurate placement AR augmentations based on a robust 3D calibration. Results We demonstrate the clinical relevance of our proposed system through two examples: a) measurement of the surface; b) depth cues in monocular endoscopy. The performance and accuracy evaluations of the proposed framework consist of two steps. First, we have created a computer-generated endoscopy simulation video to quantify the accuracy of the camera tracking by comparing the results of the video camera tracking with the recorded ground-truth camera trajectories. The accuracy of the surface reconstruction is assessed by evaluating the Root Mean Square Distance (RMSD) of surface vertices of the reconstructed mesh with that of the ground truth 3D models. An error of 1.24mm for the camera trajectories has been obtained and the RMSD for surface reconstruction is 2.54mm, which compare favourably with previous approaches. Second, \textit{in vivo} laparoscopic videos are used to examine the quality of accurate AR based annotation and measurement, and the creation of depth cues. These results show the potential promise of our geometry-aware AR technology to be used in MIS surgical scenes. Conclusions The results show that the new framework is robust and accurate in dealing with challenging situations such as the rapid endoscopy camera movements in monocular MIS scenes. Both camera tracking and surface reconstruction based on a sparse point cloud are effective and operated in real-time. This demonstrates the potential of our algorithm for accurate AR localization and depth augmentation with geometric cues and correct surface measurements in MIS with monocular endoscopes.

Die Begutachtung, der Handel und das Sammeln von Kunstgegenständen findet nicht ausschließlich analog statt. Die Firma 4ARTechnologies entwickelt Softwarelösungen für das digitale Kollektionsmanagement physischer und digitaler Kunst. Mittels Applikationen auf mobilen Endgeräten können Nutzer ihre Gemälde registrieren, authentifizieren und periodisch präzise Zustandsberichte erstellen. Die Erstellung von Zustandsberichten führt jedoch aufgrund von menschlichen Limitierungen zu Problemen in der Handhabung der Applikation und soll mithilfe eines mobilen Roboters automatisiert werden. Das Ziel dieser Arbeit ist die Entwicklung einer Navigation für einen mobilen Roboter. Diese soll folgendes Problem lösen: Lokalisierung eines Gemäldes, kollisionsfreie Annäherung und horizontal-mittige Positionierung davor. Zielplattform dieser Software ist das mobile Betriebssystem iOS. Für die Lösung wurden Verfahren der Navigation mobiler Roboter und der computergestützten Erkennung von Bildern untersucht. Die Navigationssoftware nutzt zur Zielfindung das Feature-Matching aus der OpenCV-Bibliothek. Für die Schätzung der eigenen Position werden relative Lokalisierungverfahren wie Posenverfolgung und Odometrie eingesetzt. Die Abbildung der Umgebung sowie der Bewegungsverlauf des Roboters werden auf einer topologischen Karte dargestellt. Mittels implementiertem BUG3-Algorithmus werden Hindernisse umfahren.:1. Einleitung 1.1. Problembeschreibung und thematische Abgrenzung 1.2. Aufbau Roboter 1.3. Randbedingungen und Anforderungen 2. Theoretische Grundlagen 2.1. Robotik 2.1.1. Mobile Robotik 2.2. Navigation 2.2.1. Lokalisierung 2.2.2. Kartierung 2.2.3. SLAM 2.2.4. Pfadfindung 2.2.5. Augmented Reality 2.3. Computer Vision 2.3.1. OpenCV 2.3.2. Vorlagen Erkennung 2.3.3. Template-basiertes Matching 2.3.4. Feature-basiertes Matching 3. Praktische Umsetzung 3.1. Programmablauf der Navigation 3.1.1. Verbindung mit dem Roboter 3.1.2. Initiale Exploration 3.1.3. Lokalisation und Annäherung 3.1.4. Kollisionsvermeidung 3.1.5. Zielanfahrt und Positionierung 4. Tests 4.1. Störfaktoren 5. Fazit und Ausblick 5.1. Fazit 5.2. Ausblick / The appraisal, trading and collecting of art objects does not only take place analogously. The company 4ARTechnologies develops software solutions for the digital collection management of physical and digital art. Using applications on mobile devices, users can register and authenticate their paintings and periodically create precise condition reports. The creation of condition reports leads to problems in handling the application due to human limitations and should be automated with the help of a mobile robot. The goal of this work is the development of a navigation system for a mobile robot. This should solve the following problem: Localization of a painting and the collision-free arrival and horizontal-center position in front of it. The target platform of this software is the mobile operating system iOS. Several methods, including the navigation of mobile robots and the computer-aided recognition of images were examined for the solution. The navigation software uses feature matching from the Open-CV library to find the destination. Relative localization methods such as pose tracking and odometry are used to estimate the robots own position. The environment and the movement of the robot are shown in a topological map. Obstacles are bypassed using the implemented BUG3 algorithm.:1. Einleitung 1.1. Problembeschreibung und thematische Abgrenzung 1.2. Aufbau Roboter 1.3. Randbedingungen und Anforderungen 2. Theoretische Grundlagen 2.1. Robotik 2.1.1. Mobile Robotik 2.2. Navigation 2.2.1. Lokalisierung 2.2.2. Kartierung 2.2.3. SLAM 2.2.4. Pfadfindung 2.2.5. Augmented Reality 2.3. Computer Vision 2.3.1. OpenCV 2.3.2. Vorlagen Erkennung 2.3.3. Template-basiertes Matching 2.3.4. Feature-basiertes Matching 3. Praktische Umsetzung 3.1. Programmablauf der Navigation 3.1.1. Verbindung mit dem Roboter 3.1.2. Initiale Exploration 3.1.3. Lokalisation und Annäherung 3.1.4. Kollisionsvermeidung 3.1.5. Zielanfahrt und Positionierung 4. Tests 4.1. Störfaktoren 5. Fazit und Ausblick 5.1. Fazit 5.2. Ausblick