I am working as a researcher in Tampere University Computer Human Interaction group (TAUCHI) [1] in Finland in Ubiquitous Computing project funded by the Academy of Finland. I found UbiComp during my HCI studies through my interest on context aware computing and user modelling. UbiComp interested me, because it combines the sci-fi like view of computers and user centered design in an interesting way.

I am currently working on an application called "Ovimies", a doorman in English. We are building a smart communicational environment in TAUCHI premises. In our project we are building a system to help the visitors and staff of our group in their every day communicational tasks and problems, such as finding a person or a room in our group premises and handling phone calls while being away from the phone.

The system is going to be constructed by equipping TAUCHI premises. The main sensory channel used in our project is sound, especially speech. The application is done in collaboration with our Adaptive speech user interface project, so the use of speech recognition and synthesis is crucial part of our project.

We are using a distributed JASPIS architecture [2] as a base of our software architecture. JASPIS (Java-based Adaptive SPeech user Interface development System) is being created in TAUCHI by the Adaptive Speech User Interfaces project. JASPIS provides adaptive base for our system and it can be used also with other modalities than speech.

At the moment we are trying to find the most suitable hardware architecture to connect and control all the devices needed in the system. We have considered CAN (Control Area Network) to be a possible solution.

My main interests are context awareness and user modelling. It is challenging to gather information about the user who is not carrying any equipment or identification tag. The other methods for user recognition, for example voice or image recognition, used alone in everyday situation are unreliable. The picture of the user and the situation must be gathered from small pieces of information.

There are several topic areas that I am interested in concerning our project and the infrastructure. Sensory fusion must be used to find optimal combination of different kind of sensory devices to get reliable results. Many sorts of data must be gathered, stored and processed all the time to get information about the situation. Data fusion must be used to combine and analyse information in context to draw conclusions about the situation.

[2] Markku Turunen and Jaakko Hakulinen. Jaspis - A Framework for Multilingual Adaptive Speech Applications. In Proceedings of 6th International Conference of Spoken Language Processing, 2000 (to appear).