Compact And Temperature-strain-insensitive Bending Sensor Based On An SMS-LPFG With High Sensitivity
Abstract
Fiber-optic bending sensors often suffer from cross-sensitivity to temperature and strain, necessitating complex compensation schemes or cumbersome fabrication processes. Here, we propose and experimentally demonstrate an ultra-compact bending sensor based on a cascaded single-mode-multimode-single-mode-fiber long-period grating (SMS-LPFG) that addresses these limitations. The sensor, with a total length of only 3.03 mm, is fabricated through a simple and repeatable process of periodically splicing standard SMF and MMF segments. Unlike conventional wavelength-modulated sensors, this device leverages a high-sensitivity intensity-based interrogation mechanism, achieving a bending sensitivity of up to 24.46 dB/m−1 in the range of 0.95–1.5 m−1. Critically, the sensor exhibits exceptional insensitivity to axial strain (8.055 x 10−5 dB/με) and low thermal sensitivity (9.37 x 10−3 dB/°C), corresponding to curvature crosstalk of 3.29 x 10-6 m−1/με and 3.83 x 10-4 m−1/℃, respectively, effectively eliminating thermo-mechanical crosstalk without the need for external compensation. Furthermore, to overcome the inherent nonlinearity of the intensity response and extend the operational dynamic range, we introduce and validate a Gaussian process regression (GPR) model for sensor calibration. This machine-learning approach provides a unified and accurate prediction of curvature from the raw spectral data across a broad measurement range. The combination of a compact footprint, straightforward fabrication, inherent parameter insensitivity, and a novel data-driven calibration strategy renders the proposed SMS-LPFG a highly competitive and practical solution for accurate bending measurements in multi-parameter environments.
Recommended Citation
W. Yang et al., "Compact And Temperature-strain-insensitive Bending Sensor Based On An SMS-LPFG With High Sensitivity," Infrared Physics and Technology, vol. 159, article no. 106845, Elsevier, Nov 2026.
The definitive version is available at https://doi.org/10.1016/j.infrared.2026.106845
Department(s)
Electrical and Computer Engineering
Keywords and Phrases
Bending; Long period fiber grating; Machine-learning; Strain; Temperature
International Standard Serial Number (ISSN)
1350-4495
Document Type
Article - Journal
Document Version
Citation
File Type
text
Language(s)
English
Rights
© 2026 Elsevier, All rights reserved.
Publication Date
01 Nov 2026
Comments
National Natural Science Foundation of China, Grant 11704086