Odor Sensor System Using Chemosensitive Resistor Array and Machine Learning

Rui Yatabe, Atsushi Shunori, Bartosz Wyszynski, Yosuke Hanai, Atsuo Nakao, Masaya Nakatani, Akio Oki, Hiroaki Oka, Takashi Washio, Kiyoshi Toko

Research output: Contribution to journalArticlepeer-review

8 Citations (Scopus)


In this study, we developed an odor sensor system using chemosensitive resistors, which outputted multichannel data. Mixtures of gas chromatography stationary materials (GC materials) and carbon black were used as the chemosensitive resistors. The interaction between the chemosensitive resistors and gas species shifted the electrical resistance of the resistors. Sixteen different chemosensitive resistors were fabricated on an odor sensor chip. In addition, a compact measurement instrument was fabricated. Sixteen channel data were obtained from the measurements of gas species using the instrument. The data were analyzed using machine learning algorithms available on Weka software. As a result, the sensor system successfully identified alcoholic beverages. Finally, we demonstrated the classification of restroom odor in a field test. The classification was successful with an accuracy of 97.9%.

Original languageEnglish
Article number9167294
Pages (from-to)2077-2083
Number of pages7
JournalIEEE Sensors Journal
Issue number2
Publication statusPublished - Jan 15 2021

All Science Journal Classification (ASJC) codes

  • Instrumentation
  • Electrical and Electronic Engineering


Dive into the research topics of 'Odor Sensor System Using Chemosensitive Resistor Array and Machine Learning'. Together they form a unique fingerprint.

Cite this