A joint research team from YSU and Tsinghua University announces a breakthrough in optical fiber interferometric sensing. By applying a long short-term memory (LSTM) neural network, the researchers achieve full-spectrum recognition of optical interference spectra, overcoming the long-standing trade-off between high sensitivity and wide measurement range. The study, led by Professor Hailiang Chen of YSU’s School of Science in collaboration with Professor Sigang Yang of Tsinghua University, is published in Light: Science & Applications (Impact Factor ~23.4) under the title “LSTM-assisted Optical Fiber Interferometric Sensing: Breaking the Limitation of Free Spectral Range”.
Research roadmap for the study “LSTM-assisted Optical Fiber Interferometric Sensing”
Optical interferometric sensors are prized for their high sensitivity and quality factor and are widely used in rapid detection across physics, chemistry, biology and medicine. Their broader use, however, is constrained by the free spectral range (FSR), which makes it difficult to achieve both high sensitivity and a wide measurement range. Researchers now report a way to address this challenge. The team integrates a LSTM neural network into the sensing system, as shown in the figure, and demonstrates wide-range detection beyond a single FSR with a Mach–Zehnder interferometer built from tapered single-mode optical fibers. Rather than tracking peaks and valleys in the spectra, their full-spectrum recognition approach effectively overcomes the FSR barrier.
Optical interferometric sensors rely on real-time spectral data, but dense sampling during rapid parameter changes can compromise their performance. The YSU–Tsinghua team shows that its LSTM model overcomes this challenge. Even with sparse sampling, it preserves precise one-to-one mapping between spectra and measurement parameters, a capability that conventional demodulation methods still struggle to deliver. The findings underscore the growing role of artificial intelligence in optical sensing. By enabling low-cost systems that combine high sensitivity with a wide measurement range, the approach opens new opportunities for next-generation sensors in medical diagnostics, environmental monitoring and industrial process control.
The project brings together researchers from YSU and Tsinghua University. YSU doctoral students Junling Hu and Sa Zhang serve as co–first authors, with Professor Hailiang Chen as corresponding author. The study is supported by the National Natural Science Foundation of China and the Beijing–Tianjin–Hebei Basic Research Cooperation Project, among others.
Citation: Junling Hu#, Sa Zhang#, Meiyu Cai, Mingjian Ma, Shuguang Li, Hailiang Chen*, and Sigang Yang. LSTM-assisted Optical Fiber Interferometric Sensing: Breaking the Limitation of Free Spectral Range. Light: Science & Applications, 2025. DOI: 10.1038/s41377-025-02008-4.
