GNSS is often inaccurate and satellite signals are not always available, which results in ambiguous situations. In order to reduce their negative effects on train-borne localization, this work proposes an approach for the detection of tracks, turnouts, and branching directions solely from 2d lidar sensor measurements. The experimental evaluation shows highly correct and complete results. In summary, these detections are sufficient to reduce ambiguity problems in train-borne localization.

Mobile robotic systems need to perceive their surroundings in order to act independently. In this work a perception framework is developed which interprets the data of a binocular camera in order to transform it into a compact, expressive model of the environment. This model enables a mobile system to move in a targeted way and interact with its surroundings. It is shown how the developed methods also provide a solid basis for technical assistive aids for visually impaired people.

Künftige Fahrerassistenzsysteme und hochautomatisierte Fahrfunktionen basieren auf digitalen Karten. Um den hohen Anforderungen an die Korrektheit und Aktualität der enthaltenen Informationen gerecht zu werden, stellt die vorliegende Arbeit neuartige automatisierte Verfahren zur Extraktion aktueller Karteninformationen aus Flottendaten vor. Der Fokus liegt dabei auf der Inferenz von statischen Kreuzungsinformationen mittels maschinellen Lernens und statistischer Verfahren. The next generation of driver assistance systems and highly automated driving functions are based on digital maps. In order to meet the high requirements on the correctness and up-to-dateness of this information, this work presents new automated methods to extract up-to-date map information from fleet data. The focus is on the inference of static intersection information from fleet data through machine learning and statistical methods.

In der vorliegenden Arbeit wird eine Genauigkeitsuntersuchung von Normalenvektoren aus binokularen Kamerasystemen vorgestellt, die eine Modellierung der Orientierungsunsicherheit der Normalenvektoren aus wenigen Parametern ermöglicht. Die Ergebnisse der Genauigkeitsuntersuchung werden verwendet, um zwei Methoden zur Umfelderfassung aus Orientierungsinformationen auszulegen: eine Freibereichsschätzung für Fahrzeuge sowie eine Treppendetektion für Menschen mit eingeschränkter Sehfähigkeit. The presented work introduces an accuracy analysis of normal vectors from binocular camera systems, which allows to model the surface orientation uncertainties by considering few parameters. The results of the accuracy analysis are used to design two methods for environment perception: a free space estimation for vehicles as well as the staircase detection for visually impaired persons.

In dieser Arbeit wird die Trajektorienplanung für ein automatisiertes Fahrsystem behandelt. Es werden sowohl kontinuierliche, lokal konvergierende, als auch diskrete, global optimale Verfahren beschrieben. Es wird analytisch und im Experiment gezeigt, dass beide Ansätze für das automatisierte Fahren geeignet sind, wobei jeweils unterschiedliche Fahrsituationen abgedeckt werden. This work describes trajectory planning for autonomous driving. It treats continuous methods, which converge to a local optimum, as well as combinatorial methods, which are globally optimal. It is shown by analysis and experimentally, that both approaches are suitable for autonomous driving, but cover different driving situations.

Multi-camera systems are being deployed in a variety of vehicles and mobile robots today. To eliminate the need for cost and labor intensive maintenance and calibration, continuous self-calibration is highly desirable. In this book we present such an approach for self-calibration of multi-Camera systems for vehicle surround sensing. In an extensive evaluation we assess our algorithm quantitatively using real-world data.