| Project/Area Number |
18K02236
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 08030:Family and consumer sciences, and culture and living-related
|
|---|
| Research Institution | Akita University |
|---|
Principal Investigator |
Tanaka Motoshi 秋田大学, 理工学研究科, 准教授 (50261649)
|
|---|
| Project Period (FY) |
2018-04-01 – 2021-03-31
|
| Project Status |
Completed (Fiscal Year 2020)
|
| Budget Amount *help |
¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000)
Fiscal Year 2020: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Fiscal Year 2019: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Fiscal Year 2018: ¥1,820,000 (Direct Cost: ¥1,400,000、Indirect Cost: ¥420,000)
|
|---|
| Keywords | 生活活動音 / 時間-周波数解析 / 異常検出 / 機械学習 / 自己組織化マップ / 足音 / 特徴抽出 / 深層学習 / 信号解析 |
|---|
| Outline of Final Research Achievements |
In order to develop a detection system of abnormal situations (such as accidents) for a person living alone, a detection method using machine learning was investigated. Self-organizing map (SOM) of only sounds of daily activities recorded with a broadband microphone was obtained. The weights of the SOM after learning, as feature vectors, were classified by the hierarchical clustering, and then the stochastic model was obtained. The generation probability of the sound of an activity was calculated, and the probabilities varied corresponding to sound changes. Observing the generation probability is expected to be one of the methods for detecting abnormal situations. Also, sounds produced during sleep were analyzed and footstep recognition with deep learning was investigated, whose results indicated the feasibility of using those for detecting abnormalities.
|
| Academic Significance and Societal Importance of the Research Achievements |
事故時に発生する音の収集は困難であり，個々の音の識別には膨大なデータベースが必要なため，滅多に起きない（日常的ではない）音を検出する方法について検討した。教師無し学習の1つである自己組織化マップ（SOM）の利用を検討し，日常生活音のSOMを用いた確率モデルを作成した。音の発生確率を算出でき，転倒事故を模擬した音を異常状態（候補）として検出できた。今後，本方法を確立できれば，家庭内事故等の検出システムのみならず，故障の診断など他の異常音検出への応用も期待できる。
|
