2020 Fiscal Year Final Research Report

Research on IoT devices for environmental measurement and time series data forecasting

Research Project

PDF

Project/Area Number	17K01339
Research Category	Grant-in-Aid for Scientific Research (C)
Allocation Type	Multi-year Fund
Section	一般
Research Field	Natural disaster / Disaster prevention science
Research Institution	Keio University
Principal Investigator	Matsumoto Yoshinori 慶應義塾大学, 理工学部(矢上), 教授 (60252318)
Project Period (FY)	2017-04-01 – 2021-03-31
Keywords	センサ / IoT / 機械学習 / 可視化 / クラウド / 時系列データ / 環境計測
Outline of Final Research Achievements	We researched and developed a system that measures and collects environmental information using sensors and IoT technology, and provides time-series data prediction and warning on a cloud server. In the first year, we mainly developed hardware and conducted demonstration experiments and visualization. In the third year, we conducted machine learning using Long Short Term Memory (LSTM) for prediction from time-series environmental measurement data. In the second year, we conducted machine learning using Long Short Term Memory (LSTM) for forecasting from time-series environmental measurement data. As for social implementation, we developed solar-powered IoT PM2.5, NO2, temperature and humidity meters and installed them in Uganda, East Africa, to improve the accuracy of weather forecasting.
Free Research Field	センサ、IoT
Academic Significance and Societal Importance of the Research Achievements	研究期間を通じて申請書記載の３つの技術課題を達成することで「途上国向けIoT環境モニタリングポスト」を実現して途上国が導入・長期運用可能な装置を実現する事ができた。高耐久自立電源が実現できた事から、電源がない地方や山林・荒野における環境モニタリングが増加させるすることができ大気環境の定量化が可能となる。このデータと機械学習予測を組み合わせることで大気環境の把握が可能となり適切な環境改善対策の立案や防災に生かすことが可能となる。さらに、IoT技術によりクラウドを介在させ可視化しつつ長時間の環境データを蓄積して環境研究に役立てる事や市民が環境情報にアクセスすることで健康被害を防ぐことが可能となる。