Influence of indoor thermal indexes to a tin oxide gas sensor output and prediction of its output

Abstract

There are various researches on a tin oxide gas sensor. It is effective to adopt the various information in an indoor environment. The information is derived from the sensor output characteristics. And, it is also understood that the sensor output is influenced by the indoor climate. The purpose of this research is to clarify the influence of the temperature, humidity and the atmospheric pressure to the gas sensor output. However, the fluctuation of the base level of the sensor output every day is considerably large. So, it is difficult to introduce directly some information from the sensor level. It is attempted to characterize the variance by smoothing the sensor output using the moving average method. As for results, a strong correlation coefficient presents between the sensor output and the atmospheric pressure. This result can be applied to predict a sensor output. The sensor output is derived roughly using an approximate equation, which is introduced by the trigonometric function. The predicted value, however, does not have good accuracy in summer and winter. A theory of Markov source is used to predict an accurate gas sensor output still more. The Markov prediction value is good at an actual sensor output in summer. Both methods are very effective to predict an indoor thermal indexes using tin oxide gas sensor.