Recent Advances in Signal Processing, Precision Enhancement, and Artificial Intelligence Integration in Self-Mixing Interferometry

A novel optical sensing technology, called Self-Mixing Interferometry (SMI), has been developed that uses a laser cavity as the emitter and detector, with no complicated alignment requirements. This review covers key SMI developments between 2020 and 2025, highlighting signal processing, AI integration, and precision enhancement. Ten seminal papers were analyzed to show how SMI has progressed from niche laboratory instruments to more broadly used industrial metrology instruments with a degree of precision of less than 10 nanometers. The accuracy of deep-learning-based fringe detection is over 95% in severe noise conditions; the orthogonal signal phase multiplication error is less than 11 nm (absolute value), and dual-wavelength systems feature 66 nm measurement steps. The use of wavelet-based denoising increased the reconstruction accuracy by >11%, and the use of artificial intelligence-based parameter estimation removed the need to manually calibrate the reconstruction. Multidimensional sensing was implemented to obtain simultaneous distance-velocity measurements with a relative error of 10-4. Numerous gaps exist in the real-time integration of edge AI, environmental strength, and standardized evaluation. Applications include Industry 4.0, structural monitoring, and biomedicine. Examples of future directions include physics-informed neural networks, quantum-enhanced sensing, and silicon photonics.