On-chip Fourier-transform spectrometers and machine learning

a new route to smart photonic sensors

Alaine Herrero-Bermello, Jiangfeng Li, Mohammad Khazaei, Yuri Grinberg, Aitor V. Vellasco, Martin Vachon, Pavel Cheben, Lina Stankovic, Vladimir Stankovic, Dan-Xia Xu, Jens H. Schmid, Carlos Alonso-Ramos

Research output: Contribution to journalArticle

Abstract

Miniaturized silicon photonics spectrometers capable of detecting specific absorption features have great potential for mass market applications in medicine, environmental monitoring, and hazard detection. However, state-of-the-art silicon spectrometers are limited by fabrication imperfections and environmental conditions, especially temperature variations, since uncontrolled temperature drifts of only 0.1 °C distort the retrieved spectrum precluding the detection and classification of the absorption features. Here, we present a new strategy that exploits the robustness of machine learning algorithms to signal imperfections, enabling recognition of specific absorption features in a wide range of environmental conditions. We combine on-chip spatial heterodyne Fourier-transform spectrometers and supervised learning to classify different input spectra in the presence of fabrication errors, without temperature stabilization or monitoring. We experimentally show differentiation of four different input spectra under an uncontrolled 10 °C range of temperatures, about 100x increase in operational range, with a success rate up to 82.5% using state-of-the-art support vector machines and artificial neural networks.
Original languageEnglish
Number of pages5
JournalOptics Letters
Publication statusAccepted/In press - 4 Nov 2019

Fingerprint

machine learning
learning
chips
routes
photonics
spectrometers
sensors
fabrication
environmental monitoring
temperature
defects
silicon
medicine
hazards
stabilization

Keywords

  • spectrometers
  • absorption
  • machine learning

Cite this

Herrero-Bermello, A., Li, J., Khazaei, M., Grinberg, Y., Vellasco, A. V., Vachon, M., ... Alonso-Ramos, C. (Accepted/In press). On-chip Fourier-transform spectrometers and machine learning: a new route to smart photonic sensors. Optics Letters.
Herrero-Bermello, Alaine ; Li, Jiangfeng ; Khazaei, Mohammad ; Grinberg, Yuri ; Vellasco, Aitor V. ; Vachon, Martin ; Cheben, Pavel ; Stankovic, Lina ; Stankovic, Vladimir ; Xu, Dan-Xia ; Schmid, Jens H. ; Alonso-Ramos, Carlos. / On-chip Fourier-transform spectrometers and machine learning : a new route to smart photonic sensors. In: Optics Letters. 2019.
@article{c29d5198c40f4511aa180f1ce465d96c,
title = "On-chip Fourier-transform spectrometers and machine learning: a new route to smart photonic sensors",
abstract = "Miniaturized silicon photonics spectrometers capable of detecting specific absorption features have great potential for mass market applications in medicine, environmental monitoring, and hazard detection. However, state-of-the-art silicon spectrometers are limited by fabrication imperfections and environmental conditions, especially temperature variations, since uncontrolled temperature drifts of only 0.1 °C distort the retrieved spectrum precluding the detection and classification of the absorption features. Here, we present a new strategy that exploits the robustness of machine learning algorithms to signal imperfections, enabling recognition of specific absorption features in a wide range of environmental conditions. We combine on-chip spatial heterodyne Fourier-transform spectrometers and supervised learning to classify different input spectra in the presence of fabrication errors, without temperature stabilization or monitoring. We experimentally show differentiation of four different input spectra under an uncontrolled 10 °C range of temperatures, about 100x increase in operational range, with a success rate up to 82.5{\%} using state-of-the-art support vector machines and artificial neural networks.",
keywords = "spectrometers, absorption, machine learning",
author = "Alaine Herrero-Bermello and Jiangfeng Li and Mohammad Khazaei and Yuri Grinberg and Vellasco, {Aitor V.} and Martin Vachon and Pavel Cheben and Lina Stankovic and Vladimir Stankovic and Dan-Xia Xu and Schmid, {Jens H.} and Carlos Alonso-Ramos",
note = "{\circledC} 2019 Optical Society of America. One print or electronic copy may be made for personal use only. Systematic reproduction and distribution, duplication of any material in this paper for a fee or for commercial purposes, or modifications of the content of this paper are prohibited.",
year = "2019",
month = "11",
day = "4",
language = "English",
journal = "Optics Letters",
issn = "0146-9592",
publisher = "Optical Society of America",

}

Herrero-Bermello, A, Li, J, Khazaei, M, Grinberg, Y, Vellasco, AV, Vachon, M, Cheben, P, Stankovic, L, Stankovic, V, Xu, D-X, Schmid, JH & Alonso-Ramos, C 2019, 'On-chip Fourier-transform spectrometers and machine learning: a new route to smart photonic sensors', Optics Letters.

On-chip Fourier-transform spectrometers and machine learning : a new route to smart photonic sensors. / Herrero-Bermello, Alaine; Li, Jiangfeng; Khazaei, Mohammad; Grinberg, Yuri; Vellasco, Aitor V.; Vachon, Martin; Cheben, Pavel; Stankovic, Lina; Stankovic, Vladimir; Xu, Dan-Xia; Schmid, Jens H.; Alonso-Ramos, Carlos.

In: Optics Letters, 04.11.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - On-chip Fourier-transform spectrometers and machine learning

T2 - a new route to smart photonic sensors

AU - Herrero-Bermello, Alaine

AU - Li, Jiangfeng

AU - Khazaei, Mohammad

AU - Grinberg, Yuri

AU - Vellasco, Aitor V.

AU - Vachon, Martin

AU - Cheben, Pavel

AU - Stankovic, Lina

AU - Stankovic, Vladimir

AU - Xu, Dan-Xia

AU - Schmid, Jens H.

AU - Alonso-Ramos, Carlos

N1 - © 2019 Optical Society of America. One print or electronic copy may be made for personal use only. Systematic reproduction and distribution, duplication of any material in this paper for a fee or for commercial purposes, or modifications of the content of this paper are prohibited.

PY - 2019/11/4

Y1 - 2019/11/4

N2 - Miniaturized silicon photonics spectrometers capable of detecting specific absorption features have great potential for mass market applications in medicine, environmental monitoring, and hazard detection. However, state-of-the-art silicon spectrometers are limited by fabrication imperfections and environmental conditions, especially temperature variations, since uncontrolled temperature drifts of only 0.1 °C distort the retrieved spectrum precluding the detection and classification of the absorption features. Here, we present a new strategy that exploits the robustness of machine learning algorithms to signal imperfections, enabling recognition of specific absorption features in a wide range of environmental conditions. We combine on-chip spatial heterodyne Fourier-transform spectrometers and supervised learning to classify different input spectra in the presence of fabrication errors, without temperature stabilization or monitoring. We experimentally show differentiation of four different input spectra under an uncontrolled 10 °C range of temperatures, about 100x increase in operational range, with a success rate up to 82.5% using state-of-the-art support vector machines and artificial neural networks.

AB - Miniaturized silicon photonics spectrometers capable of detecting specific absorption features have great potential for mass market applications in medicine, environmental monitoring, and hazard detection. However, state-of-the-art silicon spectrometers are limited by fabrication imperfections and environmental conditions, especially temperature variations, since uncontrolled temperature drifts of only 0.1 °C distort the retrieved spectrum precluding the detection and classification of the absorption features. Here, we present a new strategy that exploits the robustness of machine learning algorithms to signal imperfections, enabling recognition of specific absorption features in a wide range of environmental conditions. We combine on-chip spatial heterodyne Fourier-transform spectrometers and supervised learning to classify different input spectra in the presence of fabrication errors, without temperature stabilization or monitoring. We experimentally show differentiation of four different input spectra under an uncontrolled 10 °C range of temperatures, about 100x increase in operational range, with a success rate up to 82.5% using state-of-the-art support vector machines and artificial neural networks.

KW - spectrometers

KW - absorption

KW - machine learning

UR - https://www.osapublishing.org/ol/home.cfm

M3 - Article

JO - Optics Letters

JF - Optics Letters

SN - 0146-9592

ER -