The functioning of intelligent autonomous systems requires constant situation awareness and cognition analysis. Thus, it needs a memory structure that contains a description of the surrounding environment (world model) and serves as a central information hub. This book presents a row of theoretical and experimental results in the field of world modeling. This includes areas of dynamic and prior knowledge modeling, information fusion, management and qualitative/quantitative information analysis.

This book is a collection of proceedings of the talks given at the 2016 annual joint workshop of Fraunhofer IOSB and the Vision and Fusion Laboratory (IES) by the doctoral students of both institutions. The topics of individual contributions range from computer vision, optical metrology, and machine learning to data fusion and human-machine interaction.

Frühwarnsysteme dienen zur möglichst frühzeitigen Information über eine sich anbahnende oder auftretende Gefahr, um Personen und Organisationen die Möglichkeit zu geben entsprechend darauf reagieren zu können. Die Konzeption eines Frühwarnsystems stellt komplexe Herausforderungen an die Systemarchitekten, hierzu liefert die vorliegende Arbeit ein Rahmenwerk für die Architektur zukünftiger Frühwarnsysteme. Early warning systems are supposed to deliver information about an emerging threat in order to allow persons and organizations to react accordingly. The design of an early warning system presents complex challenges to the system architects. For this, the present work provides a framework for the architecture of future early warning systems. Particular attention is paid to solve various architectural problems by means of semantic technologies and the automation of workflows.

In this work, the Uncertainty Quantification (UQ) approaches combined systematically to analyze and identify systems. The generalized Polynomial Chaos (gPC) expansion is applied to reduce the computational effort. The framework using gPC based on Bayesian UQ proposed in this work is capable of analyzing the system systematically and reducing the disagreement between the model predictions and the measurements of the real processes to fulfill user defined performance criteria.

Der technische Aufbau von Sicherheitstechnologie hat einen entscheidenden Einfluss auf ihre Akzeptanz. Um diese besser zu verstehen wird ein Akzeptanzmodell für die intelligente Videoüberwachung erstellt. Getrieben durch ein ausgewähltes Szenario, der datenschutzfreundlichen Sturzerkennung, werden akzeptanzsteigernde Technologien entwickelt, in einen Prototyp integriert und evaluiert. Die Arbeit vereint theoretische Überlegungen zur Akzeptanz von Videoüberwachung mit deren gezielten Entwicklung. The technical design of security systems has a high impact on its acceptance. To understand the influencing factors, an acceptance model for smart video surveillance is developed. To improve acceptance different technical components are designed and implemented. Lastly they are integrated into a prototype for fall detection in hospital environments and evaluated for the acceptance. The book combined theoretical work on understanding acceptance with practical work into developing new systems.

Mobile robot motion planning in unstructured dynamic environments is a challenging task. Thus, often suboptimal methods are employed which perform global path planning and local obstacle avoidance separately. This work introduces a holistic planning algorithm which is based on the concept of state × time lattices with variable dimensionality and multiple resolutions. The adaptive planning method makes it possible to rapidly plan local maneuvers while still pursuing global optimality.

The availability of video data is an opportunity and a challenge for law enforcement agencies. Face recognition methods can play a key role in the automated search for persons in the data. This work targets efficient representations of low-quality face sequences to enable fast and accurate face search. Novel concepts for multi-scale analysis, dataset augmentation, CNN loss function, and sequence description lead to improvements over state-of-the-art methods on surveillance video footage.

Für die Schätzung der Höhenkarten aus Sidescan-Sonardaten liefert die Arbeit mehrere Beiträge: Ein neues Schätzverfahren, das Sonarmessungen aus vorberechneten Sonarantworten von Basiselementen, sog. Kerneln, zusammensetzt und so zu einer Höhenschätzung gelangt. Des Weiteren ein dreidimensionales Verfahren, das auf Markov Random Fields basiert und eine Sidescan-Sonarsimulationsumgebung für beliebige dreidimensionale Szenen, die auch verschiedene Sonaraufnahmemodi und Terraingeneratoren bietet. This work makes several contributions to the process of estimating elevation maps from side-scan sonar data: A new estimation method that recreates sonar measurements by pre-computed known sonar responses (so called kernels) and then derives a height profile from the kernels used. Additionally, a 3D method based on Markov Random Fields and a side-scan sonar simulation environment for arbitraty 3D scenes featuring different sonar modes and several terrain generators have been developed.

Diese Arbeit beschäftigt sich mit der automatischen Erkennung komplexer Situationen in Bildfolgen im Videoüberwachungskontext. Bei der Behandlung von Daten aus natürlichen Umgebungen ergeben sich Schwierigkeiten. Diese Arbeit erweitert dazu den verwendeten Formalismus um die Behandlung von Unschärfe, fehlender Information und Komplexität, zeigt die Robustheit der Situationserkennung bei natürlichen Szenarien und stellt die generische Anwendbarkeit auch über Diskursbereichsgrenzen hinaus heraus. This work addresses the automatic detection of complex situations in image sequences in the video surveillance context. There are difficulties when dealing with data from natural environments. This work expands the formalism of FMTHL and SGTs to deal with erroneous, missing, and noisy data and complexity, demonstrates the robustness of situational recognition in natural scenarios, and expands generic applicability beyond discourse boundaries.

This research introduces a conceptual framework for enforcing privacy-related restrictions in smart video surveillance systems based on danger levels and incident types to be handled. It increases the selectivity of surveillance by restricting data processing to individuals associated to incidents under investigation. Constraints are enforced by usage control, which is instantiated for video surveillance for the first time and enables tailoring such systems to comply with data protection law.