Development of machine-learning type odor sensing technology by integrating image output type sensor and molecular sieve mechanism

2022 Fiscal Year Final Research Report

• Project/Area Number: 21K18718
• Research Category: Grant-in-Aid for Challenging Research (Exploratory)
• Allocation Type: Multi-year Fund
• Review Section: Medium-sized Section 21:Electrical and electronic engineering and related fields
• Research Institution: Toyohashi University of Technology
• Principal Investigator: Noda Toshihiko 豊橋技術科学大学, エレクトロニクス先端融合研究所, 准教授 (20464159)
• Project Period (FY): 2021-07-09 – 2023-03-31
• Keywords: CMOSセンサ / においセンサ / 機械学習型センシング / ブロードセンシング / 分子ふるい

Outline of Final Research Achievements

Image output type odor sensors that can output odor information as imaging data were fabricated by attaching odor-sensitive membranes onto a CMOS potentiometric sensor array. The sensor was used to measure gases, and the data was analyzed by machine learning (Light GBM). Three types of gases were discriminated, and the results showed that more than 1000 measurement data could be discriminated with more than 98% accuracy. It has developed a molecular sieve mechanism that enhances the effectiveness of machine learning, which consists of bridged aluminum stripes on a Si substrate with a structure that allows gas permeation through the stripes. The gas response characteristics were confirmed to vary with the stripe width and the applied voltage. The combination of these results demonstrated the feasibility of the proposed odor measurement.

Free Research Field

CMOS集積化センサ

Academic Significance and Societal Importance of the Research Achievements

あえてブロードな特性のセンサを用いることで，さまざまな対象を一括計測を実現するブロードセンシングという計測概念が実証された。ソサエティ5.0では「におい」のデジタル化も求められるが，本技術の中核であるCMOSセンサは様々な機器への組み込みが容易であり，サイバー-フィジカルインタフェースとしての展開が期待できる。