Abstract
Indoor air-pollution monitorning system was developed using a semiconductive tin oxide gas sensor. The concentration of carbon dioxide has been used as an indicator for a grade of the air-pollution in Japan. However, this method has drawbacks because there are many polluting gases besides carbon dioxide in an indoor environment.
In this study, the grade of pollution was relatively represented as the sensor output. Seasonal characteristics of the data obtained in a room for three weeks were calculated from daily output patterns with a computer. Its data were processed using OPS 5 and Rete algorithm to match condition elements of the rules and working memory elements. With this technique, temporal change such as season of the of f-set level of the sensor can be fairly negligible.
