Application of principal component analysis method for micro-resonator weak resonant signal detection

The characteristic of weak resonant signal output by an electrothermally excited microresonator is analyzed, and the principal component analysis (PCA) method is proposed and applied in the resonant frequency detection of the output weak signal by separating the noises. Simulation on weak resonant signals under different levels of noises and different quality factors of the resonator was conducted after the influence of the data selection window width and the principal component number on detection results was analyzed. Finally, the experiment platform was built, and the output signal of the sensor sample was used to verify the detection effect of the proposed PCA method. Simulation and experiment results show that the proposed PCA method could accurately obtain the resonant frequency compared with the resonant frequency obtained by the smoothing filter and Lorentzian fitting method.