Multimodal fringe detection for a self-mixing interferometry-based vibration sensor.

Robust detection of interferometric fringes is critical for accurate sensing by self-mixing interferometric (SMI) displacement sensors. Mode-hopping of a laser diode (LD) can potentially diversify SMI fringes, transforming them from mono-modal to multimodal. Thus, fringe detection of a multimodal SMI signal becomes a bigger challenge as the relative strength of each mode may be different, leading to further diversity in the fringes belonging to each regime. Also the SMI signals from each mode are incoherently added, so the composite multimodal SMI signal is of complex nature. In this paper, a robust method is proposed for the detection of multimodal fringes, which is also able to detect traditionally encountered mono-modal fringes. Since fringes are actually peaks of SMI signals, the proposed method detects all of these peaks and separates the genuine peaks that correspond to true fringes from the falsely detected peaks, corresponding to false fringes. An experimental dataset of 60 SMI signals was acquired by using two different LDs to validate our proposed method. The proposed method has correctly detected the SMI fringes with an accuracy of 99.6%. However, at the same time, 0.7% false fringes were also detected while 0.3% true fringes were undetected by the proposed method.