Electrical Only Calibration of Barometric Pressure Sensors Using Machine Learning

This paper reports a novel, rapid and economical approach to calibrate barometric micromechanical capacitive pressure sensors via electrical measurement only using a machine learning approach. Conventional calibration approaches that use physical means are time consuming and require specialized equipment. We proposed and experimentally validated an electrical only approach through machine learning to calibrate the pressure sensor over a wide range of pressures. Using this novel approach, the experimental results show an average absolute accuracy of 1.74hPa within 600hPa to 1100hPa pressure range. The typical absolute accuracy falls well within the ±1hPa. This approach reduces the test time by more than 85%.