Research Interests

General Interests

The central idea of ubiquitous computing is the emergence of a networked environment of computational enhanced tools and devices. These augmented thinks can range from designated computational tools like PDAs to everyday objects that contain embedded computing systems and are often referred to as "artifacts". What I am interested in are the aspects of the proclaimed "Ubicomp Environment". The term environment is often used in the Ubicomp community to describe the idea of communication and seamless interconnection of devices. But it also stands for the implicit assumption that by introducing ubiquitous computing we will create a variety of synergetic effects that make an Ubicomp Environment more than just the single parts of it.

My research focuses on different aspects of Ubicomp environments. The most absorbing question is how can we "built" these environments. The Ubicomp research community agrees that the best way to augment artifacts and to built applications that can support the user in her tasks in an unobtrusive way is to make use of context information. So one of my central concerns is to look into providing an architecture for context processing in Ubicomp environments that is able to provide the recommended degree of flexibility to interconnect all kinds of applications for ubiquitous computing. The idea is to have some very few agreements on the structure of the data (context information) that has to be exchanged in an Ubicomp environment to enable all applications in an area to make use of all the data that is known in this environment. Only by this the synergetic effects can be exploited if more independent applications share the environment. The goal has to be to provide an architecture for systematic context use.

Context Processing

As context is the new data type of Ubicomp systems and first context models that describe what context can be are available, it is necessary to have a closer look into the options of context processing systems and context architectures. I am not interested in inventing yet another semantic context model for the application layer of a Ubicomp system, but seek to gain some insight in how we can structure the processing and use of context information on a the lower, system oriented layers. Therefore I am studying context attributes that apply to a large variety of semantically different contexts to get an idea how a minimal generalized architecture for structural context representation could look like, that provides access to these generic attributes to all applications in an ubiquitous computing environment.

Context Algorithms

An interesting problem in designing Ubicomp applications is to determine how to get a context out of the sensor data that is available in an environment. You obviously need a reliable classifier for the contexts you want to recognize. There are very different approaches in the discussion like neural networks, fuzzy logic, Bayesian networks, self organizing maps. I think there is a "killer algorithm" to get rid of the classification problem for all time. So I'm interested in figuring out which approach performs best on data of a given characteristics.

Context Communication

The communication of context information can benefit from structural data that is implicitly provided with the context data itself (see publications for details). On the other hand communication systems can provide a large variety of data on the configuration and situation in the environment that and add to the context data available to applications in this environment. One goal to improve context aware systems is to bridge this gap between system data and application data and make context information available to all layers of an ubiquitous computing system.

Social and Technical Implications of Ubiquitous Computing Technologies

The research in the Ubicomp community is more application driven than in other research communities. This is good because you always get an idea on what a new technology is good for, but as you scan through some more publications you encounter some "standard Ubicomp applications" over and over again. But what makes a "killer app" in a new field, or does something like a killer app exist in new fields? I belive that there are some features a technology must have to have major impact to society and social behaviour. One of these is the potential to build communities of users, like SMS did for the youth and of cause the existence of a possible business model for that service.

My Ph.D. Thesis

In my Ph.D. thesis I look into the above aspects of ubiquitous computing and develop an architecture for systematic context processing in Ubicomp environments. The resulting framework will form the basis for the development of applications and artifacts that fully exploit the synergetic effects of communication and context awareness in ubiquitous computing environments.

Other Interests

All the above points give a more detailed description of what I am working on. This work involves lots of other components of information systems that I will list here only briefly:

For those who are interested in doing a thesis under my supervision: There are nearly always opportunities to do a thesis in one of those fields, even when it is not explicitly mentioned in the theses section of this page. If you are interested feel free to contact me.

For more information on what I am doing, visit the "projects" section of my page. This will be updated soon...

This site was last modified: August 17 2004 15:29:57.
Copyright Tobias Zimmer 2003, 2004, 2005, 2006.