Proposal of Unsupervised Gas Classification by Multimode Microresonator

Chun-Hua Dong,Huan Zheng,Yulu Zhang,Zichun Le,Peiqiong Yu,Jin Lu,Weisheng Hu,Hongliang Ren,Yali Qin,Chang-Ling Zou

doi:10.1109/jphot.2021.3069582

Abstract

A high-accuracy unsupervised classification model is developed in a multimode self-interference microring resonator (SIMRR). For the SIMRR, there are many whispering gallery modes (WGMs) present. Each of these resonance modes supported by the SIMRR has a different response to different target parameters, so that the SIMRR-based sensor has the super capability to distinguish between multiple components. In the classification model, principal component analysis (PCA) is firstly used to reduce the dimensionality of this multimode sensing information from the SIMRR-based sensor. When the original higher-dimensional data points are projected onto the lower-dimensional data with only the first few principal components, they can be easily categorized into several different types by using density-based spatial clustering of application with noise (DBSCAN) algorithm. As an example, the unsupervised classification method is numerically validated based on a designed three-gas SIMRR-based sensor. The numerical results prove that the classification model can achieve an ultra high classification accuracy for the designed three-gas sensor with more than 60 dB in signal-to-noise ratio.

Highlights

Optical whispering gallery mode (WGM) microresonators can significantly enhance the light-matter interaction due to its ultrahigh quality factor (Q) and relatively small mode volume [1]–[4]
To quantitatively evaluate the cluster quality in principal component analysis (PCA) space, the Davies–Bouldin index (DBI) is introduced to estimate the cluster separation, which is a metric for evaluating clustering algorithms
Each point in the PCA space corresponds to a specific mixed ratio of three gases

Summary

Introduction

Optical whispering gallery mode (WGM) microresonators can significantly enhance the light-matter interaction due to its ultrahigh quality factor (Q) and relatively small mode volume [1]–[4]. A WGM-based multi-parameter sensing often depends on a WGM resonator sensor array integrated on a single chip [6]–[8]. When the array sensor is used to simultaneously monitor multiple gas concentrations, the multi-parameter measurement is often impacted severely by the cross-sensitivity to these parameters [10]. By combining an optical WGM-based sensor with artificial intelligence algorithms, an alternative method has been proposed to realize multi-parameter detection [11]–[13]. A back propagation artificial neural network (BP-ANN) was applied to realize multi-parameter sensing by merging the multiple resonant mode changes of a single microresonator [12]. A neglected issue in these approaches is that the supervised learning algorithms require a big training dataset, and it is often difficult to obtain the labeled data in practical applications

Methods

Results

Conclusion