next up previous contents
Next: Modeling and Prediction of Up: Computational Information Processing Previous: Urban Area Classification via

Learning Control of Indoor Air

Researchers: Markus Varsta, Jouko Lampinen, and Timo Kostiainen

The project is financed by TEKES and carried out in co-operation with the Finnish Institute of Occupational Health.

The quality of indoor air is a major contributor to work atmosphere and effectivity in general and conversely perceived problems in air quality may well be symptoms of problems not at all related to inside air, consequently the aim of this project is two fold. First we aim to identify problems in the air quality that cause unpleasantness adding to work related stress and thus reduce effectiveness. Once the problems are pinpointed the air conditioning system may be adjusted accordingly. On the other hand we also want to identify problems in the working atmosphere, which problems are frequently linked with poor quality of air but in fact are caused by infected personal relations, too high work load or other causes of stress.

Our part of the project concentrates on developing neural network methods for the analysis and coupling of measured air quality data such as temperature and relative humidity with information regarding the perceived indoor air quality gathered with questionnaires. We will develop tools that aid an expert to determine and later suggest changes in air conditioning most likely to solve the problem. The tools will be based on the Self Organizing Map which is a powerful tool for grouping and visualization of high dimensional data (see Fig.5)

 
Figure 5:   A self-organizing map showing the distribution and relationships between various attributes of measured and perceived indoor air quality.
Figure 5


next up previous contents
Next: Modeling and Prediction of Up: Computational Information Processing Previous: Urban Area Classification via
Juha Merimaa
1/2/1998