Application of sensor data fusion techniques to multi chemical sensor

Devices that perform odor detection in electronic systems are called odor sensors or electronic noses. The spread of the odor in the air occurs when the odor particles move in the air like heat and light. The odor sensors sense the changes in the air and transmit the obtained information to the decision-making unit. The decision-making unit, on the other hand, combines the information from the sensors to try to guess what the smell in the air is. In this thesis, an electronic nose that determines the identity of the incoming gas using a porous silicon-based Fabry-Perot chemical sensor has been developed. This sensor has a resonant wavelength in the reflection spectrum for each chemical gas. When the sensor is given a different organic solution vapor, the initial resonance wavelength shifts to the right over time. By utilizing this amount of shift in the wavelength, the identification of the solution vapor is realized. In the literature, it is indicated that the identification of gases is realized by taking advantage of the amount of shift in the wavelength taken from a single point. In this study, the identification of the gases is obtained as a result of the evaluation of shift amount measured from the multiple points of gas identification process different from the processes mentioned in the literature. Experimental studies in this context indicate that the proposed method has a higher sensitivity to gas identification than the methods available in the literature.