Development, Analysis and Implementation of Regression Algorithm For MOS Gas Sensors
Air quality monitoring, electronic noses, breath analyzers and many other applications use gas sensors to detect the concentration of various gases and notably assist in maintaining decent air quality standards. Reliability in detection and prediction of gas concentration using sensors is of utmost importance in the course of developing sensor based applications. Sensor accuracy and system integrity can be assured by adhering to a mathematical approach that involves the formulation of inferences which is preceded by calibration. In this paper, we present the findings drawn up from performing calibration for metal oxide semiconductor gas sensors MICS 2710 (NO2 sensor) and MICS 5521 (CO sensor). The real time plotting of important data and further mathematical modeling of sensor responses were done using Python. The equations obtained subsequently were applied to ENVIROBAT 2.0, a portable urban air pollution monitoring device and the results were compared. The comparison ratified the accuracy of modeling and thus was an affirmative of sensor efficiency.
Index Terms— Calibration, gas sensors, mathematical modeling, Python, regression analysis.