Korbinian Frank

CoOL: A Context Ontology Language to Enable Contextual Interoperability

This paper describes a context modelling approach using ontologies as a formal fundament. We introduce our Aspect-Scale-Context (ASC) model and show how it is related to some other models. A Context Ontology Language (CoOL) is derived from the model, which may be used to enable context-awareness and contextual interoperability during service discovery and execution in a proposed distributed system architecture. A core component of this architecture is a reasoner which infers conclusions about the context based on an ontology built with CoOL.

Channel Model for Train to Train Communication Using the 400 MHz Band

This paper presents a channel model for direct train- to-train communication appropriate for the 400 MHz band. Extrapolation of theoretical and experimental results obtained for the planning of other railway communication systems like GSM-R is not obvious due to the difference in frequencies, antenna height and absence of base stations. In this paper, the analysis of the channel model covers different radio phenomena including path loss, Doppler, fading, and delay spread. Concretely we consider three scenarios (train stations, shunting yards and regional networks), for which the propagation channel characteristics are discussed. Furthermore, influence of special railway environments like cuttings which can be found near cities and towns, tunnels and bridges are encountered.

The Bayeslet Concept for Modular Context Inference

In the development and design of ubiquitous computing many challenges are arising. While there is much research done on service management systems and context provisioning, less effort is spent on the methods to actually generate context information - the process of context inference. If we are considering this field of research, we have not only to consider the pure algorithmic problem to infer otherwise unknown information from available data, we also have totarget the challenges of large scale systems with millions of users possibly spread across the world and the users requirements who is neither willing nor able to wait more thana couple of seconds for his request to be served. In this work we consider shortcomings in todays context inference systems and analyze requirements for emerging architectures relying on probabilistic algorithms, more precisely static Bayesian networks. We postulate the fragmentation of large networks into smaller so called Bayeslets, that are modular, (un)pluggable, individualisable and easy to process, as they are small and processing can be parallelised. Further on, we propose a formalism to note those Bayeslets in the Bayeslet Language (BalL). Hence, we have a way to easily exchange and deploy Bayeslets and even give application developers a way to provide their own inference rules to the pervasive system.

Hybrid fusion approach combining autonomous and cooperative detection and ranging methods for situation-aware driver assistance systems

Current driver assistance systems such as Adaptive Cruise Control (ACC) and in particular future assistance systems such as Collision Warning make high demands on reliability of detection and ranging methods for vehicles within the local vicinity. Autonomous systems such as Radar which are already integrated into a multitude of vehicles meet these requirements to only a limited extent. As an alternative, cooperative systems for detection and ranging will be enabled by future Vehicle-2-Vehicle communication. But cooperative detection and ranging also has drawbacks regarding reliability due to positioning and transmission errors if it is applied in a standalone way. Thus, the solution presented in this paper is a hybrid approach combining autonomous and cooperative methods for detection and ranging within a common architecture. A particle filter is used for state estimation. The results are a higher detection effectiveness and a lower position error compared to using standalone autonomous or cooperative detection and ranging methods.

Open vehicular data interfaces for in-car context inference

In this paper, we present a concept for an open vehicular data interface and describe its components and architecture. We discuss the enabled applications in the context of advanced driver assistance systems with a focus on human-machine interfaces, vehicle-to-x (V2X) communication and context inference systems. We conclude by a presentation of the initial implementation and deployed system.

Bayesian recognition of safety relevant motion activities with inertial sensors and barometer

Activity recognition has been a hot topic in research throughout the last years. Walking, standing, sitting or lying have been detected with more or less confidence, in more or less suitable system designs. None of these systems however has entered daily life, neither in mass market, nor in professional environments. What is required is an unobtrusive system, requiring few resources and - most important - recognizing all important activities with high confidence. To this end, our research has focused on the professional market for safety related applications: first responders or also military use. Next to the classical motion related activities, our system supports motions in three dimensions that are necessary for all kinds of movements indoors as well as outdoors. These include falling, wriggling, crawling, climbing stairs up and down and using an elevator. We have proven this approach to run in real-time with only a single wireless sensor attached to the body while achieving robust and reliable recognition with a delay lower than two seconds.

Recognition of Professional Activities with Displaceable Sensors

The applications of activity recognition have increased in the last years. Among its uses are promoting healthy life styles or monitoring professional users (like first responders) during life-threatening operations; which caused the need for new research lines. However systems using common devices (like smart phones) equipped with the necessary sensors (e.g. inertial sensors) are not fully developed. Our work develops and activity recognition system, with a focus on professional users, that uses a displaceable sensor. The activity set is comprised of the most common activities (static, walking, running, etc.) but also crawling and 3D motions like stairs walking. A detector of upright dynamic activities has been developed and implemented to cope with the changing sensor position. The system is evaluated using 10-fold cross validation with the inference network, and recall with the detector. The final system proves to perform well with respect to the initial requirements.