Robust Time-Domain-Based Spectral Processing Method for Laser Self-Mixing Vibration Measurement

Robust and contactless sensing schemes have always been enormous challenges in mechanical vibration measurement and need to be solved urgently. A novel vibration measurement technique based on laser self-mixing interference (SMI) is widely used in noncontact and integrated measurement schemes because of its simple structure. However, the key parameter of SMI, which is defined as the optical feedback parameter C, is time-varying resulting from the influence of the actual industrial environment and rough target surface, making it difficult to calculate and affecting the robustness of the signal extraction. In this article, a time-domain-based spectral processing method (TSPM) is proposed to solve the difficulty of robust mechanical vibration measurement with various C values. By suppressing spectral components, the TSPM can reduce different C values to an extremely low regime mathematically, which significantly minimizes the influence of C. The reconstruction of vibration can be realized without determining C or estimating other parameters. A series of experiments has been carried out under different amplitudes and speckles, and the root-mean-squared (rms) error is 7.6 nm (0.25%) for simple harmonic vibration. The experimental results agree well with the simulation. Additionally, the proposed method is noise-proof and robust, which provides a novel solution for mechanical vibration measurement and fault diagnosis.