EUVIMEDEUVIMED entdecken
Crossref

Photonic Integrated Circuits for End‐to‐End Spectral Analysis: Computational Sampling With Neural Network Reconstruction

Yuhang Wang, Ang Li, Xiaozhuo Zhao, Yang Zhao, Qiang Li, Ye Tian, Shilong Pan

Laser & Photonics Reviews · 08.11.2026

DOI: 10.1002/lpor.72169 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Laser & Photonics Reviews, online erschienen am 08.11.2026. Verfasst von Yuhang Wang und 6 weiteren Autor:innen. Die Arbeit nennt 37 Literaturangaben. Gefördert von: National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

On‐chip spectral analysis enables transformative applications, from smartphone‐based glucose monitoring to UAV‐mounted remote sensing. While progress has been made in on‐chip spectral sampling, on‐chip spectral reconstruction remains a critical unmet need as current approaches rely upon computer to run complex algorithms. Here, we report an on‐chip solution for end‐to‐end spectral analysis by combining high‐resolution spectral sampling and real‐time high‐fidelity spectral reconstruction on CMOS compatible platform. The system features 64 inversely designed broadband filters with highly random, low‐cross‐correlation transmission functions, enabling linearly uncorrelated sampling over a bandwidth exceeding 100 nm. Sampling data from 64 channels are further processed by a 32 × 32 photonic computing chip implementing a lightweight physics‐informed neural network, which performs high‐fidelity spectral reconstruction with drastically reduced latency and energy consumption compared to electronic processors. Various complex spectra can be accurately reconstructed and the experimental results demonstrate a spectral resolution of 0.025 nm for single‐peak and 0.05 nm for doublets with 100+ nm bandwidth, meeting requirements for diverse applications. This CMOS‐compatible, end‐to‐end solution enables miniaturized, low‐power spectrometers, paving the way for integration into consumer electronics such as smartphones and wearables.

Förderung

  • National Natural Science Foundation of China

Zitieren

APA

Wang, Y., Li, A., Zhao, X., Zhao, Y., Li, Q., Tian, Y., & Pan, S. (2026). Photonic Integrated Circuits for End‐to‐End Spectral Analysis: Computational Sampling With Neural Network Reconstruction. Laser & Photonics Reviews. https://doi.org/10.1002/lpor.72169

Vancouver

Wang Y, Li A, Zhao X, Zhao Y, Li Q, Tian Y, et al. Photonic Integrated Circuits for End‐to‐End Spectral Analysis: Computational Sampling With Neural Network Reconstruction. Laser & Photonics Reviews. 2026. doi:10.1002/lpor.72169

BibTeX
@article{Wang2026,
  author = {Wang, Yuhang and Li, Ang and Zhao, Xiaozhuo and Zhao, Yang and Li, Qiang and Tian, Ye and Pan, Shilong},
  title = {Photonic Integrated Circuits for End‐to‐End Spectral Analysis: Computational Sampling With Neural Network Reconstruction},
  journal = {Laser & Photonics Reviews},
  year = {2026},
  doi = {10.1002/lpor.72169}
}
Kostenlos starten →