Development of a Biocompatible Structural Color type Sensor to Identify "Good" or "Bad" Forceps Grasping Conditions

2022 Fiscal Year Final Research Report

• Project/Area Number: 21K18090
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 90130:Medical systems-related
• Research Institution: Kagawa National College of Technology
• Principal Investigator: Maeda Yusaku 香川高等専門学校, 機械工学科, 講師 (00803404)
• Project Period (FY): 2021-04-01 – 2023-03-31
• Keywords: 構造色 / 機械学習 / 生体医工学 / 低侵襲治療 / 触覚センサ / 機械学習 / 畳み込みニューラルネットワーク

Outline of Final Research Achievements

(1) Development of biocompatible structural color type sensor elements: We established a manufacturing process for biocompatible structural color-type sensors by combining mechanical structure formation by optical 3D printing, mirror surface formation by film polishing, and 1um-level gap control technology by crimping.
(2) Construction of a machine learning model for "good/bad" evaluation of grasping condition: Using convolutional neural networks, we have realized a model linking the time-series waveform of load, which is 1D data, and human sensation.
(3) Verification of operation in an biological environment: We demonstrated that there is no problem in measuring information by structural color even in a glycerin environment. We demonstrated that there is no problem in measuring information by structural color even in a glycerin environment with a viscosity of 1499 mPas, which is sufficiently larger than the viscosity of gastric juice of 40 mPas or less.

Free Research Field

生体医工学・計測工学・機械学習

Academic Significance and Societal Importance of the Research Achievements

生体適合型構造色式センサと，内視鏡カメラの組み合わせによる計測は単一の内視鏡視野から複数の情報を，同時に取得可能であり，構造的には，現行の内視鏡構成に対してセンサチップの追加のみで実現できる。一般的なアプローチである，無線素子および生体適合性を有するためのパッケージを使用する場合に比べ，実装体積を小さくでき，多様な計測への拡張性の高い技術である。加えて，取得データによる感性的な情報の取得は，難易度の高い内視鏡治療の習得を少しでも早くし，治療法自体の波及を早める貢献が期待できる。