Research on hunter assistance system using wireless sensing and machine learning

2023 Fiscal Year Final Research Report

• Project/Area Number: 20K19780
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 60060:Information network-related
• Research Institution: Radiation Effects Research Foundation
• Principal Investigator: Ono Satoru 公益財団法人放射線影響研究所, 情報技術部, 部長 (50818309)
• Project Period (FY): 2020-04-01 – 2024-03-31
• Keywords: ワイヤレスセンシング / RSSI / 動体検出 / WiFi / 機械学習 / 安全管理

Outline of Final Research Achievements

We verified the sensing of moving objects using radio waves at outdoors. As a result, it was found that the propagation range of radio waves was too small, especially in mountain forests. This is thought to be due to the fact that trees in mountain forests absorb radio waves due to their high moisture content. Therefore, the sensing of moving objects was verified indoors. The results showed that significant motion detection was possible indoors. In addition, it was confirmed that a specific pattern shown by the time-series transition of radio wave strength can be detected as human moving. We have found that it is possible to distinguish between the three basic types of changes in radio wave strength: "people moving," "people stopping," and "nothing". These three situation can be classified using machine learning that model was constructed by this study.

Free Research Field

計算機ネットワーク

Academic Significance and Societal Importance of the Research Achievements

本研究では2.4GHz帯の電波リソースを用いた．これは計算機ネットワークの伝送媒体として利用されているため，センシングのため新たなインフラを導入する必要がないというメリットがある．RSSIやCSIを用いた電波の変遷を様々な用途に利活用するための既存研究は数多く行われているが，その中で本研究は室内における電波の挙動を人やモノの動きによって顕著且つ特異的なパターンで変遷することを明らかにした点において社会的意義があると考えている．また，基本的な動作パターンの検出にあたり，センサ設置条件，検出間隔，データ処理方法，機械学習のモデル構築の4つの条件を導出できたことは学術的な成果が得られたと考えている．