High-Precision Temperature Sensing Based on Sagnac Interferometer and BPNN in Fiber Ring Laser

Temperature sensors are critical across diverse applications, yet their performance is inherently limited by measurement errors. Here, we proposed a hybrid-intelligent sensing paradigm combining a Sagnac interferometer (SI)-based fiber ring laser (FRL) with Backpropagation Neural Network (BPNN) optimization. The sensor, constructed with 30 cm polarization-maintaining fiber (PMF) with a high birefringence coefficient, achieves -0.548 nm/°C sensitivity at a range of 15 °C to 25 °C and maintains the signal-to-noise ratio (SNR) at 47.5 dB with a linewidth of 0.05 nm. Experimental results demonstrate a mean squared error (MSE) reduction from 0.01034 °C<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> in broadband light source (BBS) to 0.00691 °C<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> via FRL integration. To address residual errors impacting ultra-precision applications, BPNN-based spectral analysis further reduces MSE to 6.396×10<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-6</sup> °C<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> —a three-order-of-magnitude improvement over conventional methods. To the best of our knowledge, this work represents the first implementation of neural network-assisted error analysis in FRL system, demonstrating a hybrid-intelligent approach to enhance measurement accuracy. The designed sensing system offers transformative potential for monitoring temperature changes in quantum computing, superconducting electronics, and semiconductor fabrication.