Context Information Framework for pervasive computing environments

Goal of this project work is to find new methods, dealing with abstract semantic description of objects and their interaction and to manifest this methods in a software artefact.

Forthcoming research and new hardware concerning kommunikation technology through radio (IEEE802.11b, Bluetooth, Rfid) enables the use of interconected sensors and actors. This intelligent sensors and actors are embedded in their environment and are collecting masses of unstructured data about this environment.

For this purpose literature defines "Context Information" as follows:

"Any information that can be used to characterize the situation of an entity"[A.Dey]

In order to enable a device to feal its environment and to react specificly, it is necessary to preprocess, to interprete and to wrap this data in a suitable form. (=contextsensitive services, reactions). This embedded digital devices equipped with context information about their environment (=office, school, household, ..etc) is often called "Smart Environment" Smart Environments (=digital devices + context information about their environment) smooth the way for new ways of interaction between the real and the digital world.

To enable effective use of context information and embedded devices, it is necessary to represent the description of all different objects in an environment. The representation of a object changes dynamicly with data received by different sensors in the environment, and by interaction with other objects. We distinguise between two basic ways of retrieving context information: event driven and time triggered context information sensing.

Solutions, which are currently in use, often are isolated applications which are not able to share their valuable data with other applications like cellular phones, HomePCs, paddles, air conditions, stereos, videos and even security and alarm equipment, etc.... Many different digital solutions are currently in use which produce data about their environments, but is is not possible to use this data in form of context information. If the different devices of daily use could represent their data, context information could be modeled in order to build a detailed picture of different environments, for people and for other smart devices.

To give a digital device the possibility to react on different events occuring in its environment and to specify the details of interaction with other objects, we try to specify abstrakte rules for each object. This abstract rules should be very flexible, because the environment of a mobile device is considered to change very often.

The context framework team considers following challenges:

Context Sensing: Definition of generic methods, to gain context information event triggered and time triggered (via polling)
Representation of context information: To find a suitable abstract representation to describe semantic context information.
Integration of abstract rules: To find a context rule syntax definition, in order to control interaction of heterogeneous devices.
Communication: Usage of popular transport protocols to enable the retrieve of context information

Contact: Wolfgang Beer
Publications: OCG Workshop Paper Vienna 2001
Further information