Recursive least squares fuzzy modeling of chemoresistive gas sensors for drift compensation

Responses of chemoresistive gas sensors suffer from the influences of the variations of the ambient humidity and temperature. An appropriate countermeasure is required if any qualitative and quantitative analysis is going to be implemented based on these responses. Here, a novel compensation method based on the fuzzy modeling of the sensor behavior is presented. Gas sensor is treated as a nonlinear system that is affected simultaneously by three inputs, the partial pressure of the target gas, the humidity, and the temperature of the surrounding atmosphere. The single output of this system is the sensor´s resistance that is referred to as the sensor response. A large database was created out of the experimental results, i.e. the inputs and outputs of the system in different conditions. It was shown that an appropriate recursive least squares (RLS) fuzzy model can be employed for modeling of the system and a quantitative analysis of the responses in different conditions. The results of the analysis were employed for the partial compensation of the drift caused errors. The method reduced the humidity and temperature variation caused errors by a factor of 5 in the laboratory tests conducted.