Multitaper FFT-Based Full-Spectrum Phase Demodulation for High-Precision GaAs F–P Temperature Sensors

Authors:
Publisher website
Google Scholar
Abstract
Conventional demodulation methods for fiber-optic Fabry–Pérot (F-P) temperature sensors are limited either by the spectral resolution of wavelength-drift approaches or by spectral leakage in FFT-based phase demodulation. To address these issues, this study proposes a multi-taper FFT full-spectrum phase demodulation method (Multi-Taper FFT-FSPD) for high-precision GaAs F-P temperature sensors. In the proposed dual-band cooperative scheme, the absorption-edge shift of GaAs in the 880 nm band is used for coarse temperature localization and free spectral range (FSR) counting, while high-precision phase demodulation is performed in the 1550 nm band using multi-taper FFT for fine temperature compensation. By combining one Hamming window and two mutually orthogonal DPSS windows through coherent averaging in the complex spectrum domain, the proposed method effectively suppresses spectral leakage and improves dominant-frequency phase extraction. Experimental results show that the proposed sensor achieves a phase sensitivity of 0.235 rad/°C and a demodulation error of ±0.0984 °C. Compared with the traditional wavelength-drift method and the multi-peak average wavelength-drift method, the proposed method improves demodulation accuracy by 63.6% and 38.8%, respectively. These results demonstrate the strong potential of the proposed method for high-precision real-time temperature monitoring.
Funding Information
  • National Key Research and Development Program of China (2025YFF0523500)
  • Guangxi Natural Science Foundation (2024GXNSFAA010339)
  • in part by Guangxi Science and Technology Program Project (AD25069073)
  • in part by Shenzhen Research Foundation (ZDCY20250901095803004)
  • Strategic Priority Research Program of Chinese Academy of Sciences (XDC0190400)
  • Science and Technology Program of Suzhou (2025B01J2002)

This publication has 27 references indexed in Scilit: