Optical Frequency Comb-Based Aerostatic Micro Pressure Sensor Aided by Machine Learning

Although the whispering gallery mode (WGM) microcavity sensor is known for its high sensitivity, conventional monitoring using single-mode methods poses challenges in detecting very low signals. Here, an aerostatic micro pressure sensor based on the optical frequency comb (OFC) generated in a WGM microbubble resonator (MBR) is verified experimentally. The proposed OFC-based aerostatic sensing is implemented by analyzing the multiple comb tooth (i.e., multimodes) information instead of single mode. Notably, the true aerostatic micro pressure can be accurately predicted using an artificial neural network, while, in the same condition, the wavelength shift of a single mode is masked by noise. The detection limit (DL) is one-to-two orders of magnitude lower than that of single-mode sensor, capable of reaching 0.3672 mbar. The correlation coefficient
${R}^{{2}}$
is as high as 0.99988. We believe that this novel method represents an important step toward the practical application of OFC-based sensor outside the laboratory, particularly when combined with machine learning techniques.