Signal extraction method based on spectral processing for a dual-channel SMI vibration sensor

Laser self-mixing interferometry (SMI) is a powerful measurement technique with high sensitivity and subwave-length precision. However, the simultaneous detection of multi-targets with adjacent frequencies for current SMI systems is an enormous challenge because their spectra are overlapped to be distinguished difficulty. Here, a novel method based on the short-time Fourier transform (STFT) and pre-feedback structure is presented for dual -channel laser SMI, which realizes the recovery of two vibrations with adjacent frequencies. The spectrum of the SMI signal is expanded along the time-axis by STFT, in which the pre-feedback mirror is applied to reallocate the energy of each channel. Further, the signal is extracted separately through spectrum suppression and trans-formation. Simulation results show that the proposed approach is effective in vibration frequencies over a large range of ratio of 1:1-1:5 in the weak feedback regime. The RMS errors of the experimental results with less than 11.29 nm (0.87%) indicate the reliability of this scheme, which provides an alternative solution for the simultaneous detection of multi-parameters.