remote notification method for welfare care service utilizing sound detection

2018 Fiscal Year Final Research Report

• Project/Area Number: 15K13109
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Social welfare and social work studies
• Research Institution: Hiroshima National College of Maritime Technology
• Principal Investigator: Kajihara Kazunori 広島商船高等専門学校, 電子制御工学科, 教授 (10300617)
• Research Collaborator: WATANABE Yuta
• Project Period (FY): 2015-04-01 – 2019-03-31
• Keywords: 異常検出 / 音検出 / 介護福祉 / プライバシー配慮

Outline of Final Research Achievements

The purpose of this study was to find out a method of data analysis and faint detection that can prevent loitering and provide emergency notification by adding a source of information based on lifestyle to the urgent detection system.We focused on privacy protection of watching target person and aimed at method not to acquire information by image such as camera, but it turned out that it was more effective for detection and notification to add supplemental information to sound information.
The sound detection method that contributes to the estimation of the situation grasp of the elderly by tracking was tested. As a result, specifying the generation source by the array microphone as an abnormal sound that a falling object is generated, even if the difference between the average level of the ambient sound and the abnormal sound is small. We could identify the direction of the incident.

Free Research Field

物理量の高精度な計測方法の研究

Academic Significance and Societal Importance of the Research Achievements

異常検出時に通知するシステムであるため、対象となる在宅者は日常の生活を送っている状況での異常検出が必要である。健常な状況下では音声を活用することができるが、意識低下による転倒や物の落下など異常な状況にある場合の検出を目指した。そのため背景音から異常音の発生を抽出する必要があり、エアコンや扇風機の使用によるほどの空気の大きな流れがない場合には、音の発生方向を特定できる周波数分析の手順が見い出せた。これきっかけに、音の発生方向にカメラを向け映像を遠隔地の支援者・介護者のもとに送信することが可能となった。