Paper Abstract and Keywords |
Presentation |
2023-11-10 14:45
Time-series AI analysis of whole-body movements in independent elderly and elderly requiring care Hiroki Kawasaki, Takeshi Kohama, Masanobu Kusunoki (Kindai Univ.), Soichi Okada (Ukai Hospital), Eiji Fujita (National Institute of Fitness and Sports in Kanoya), Nobuo Takeshima (Asahi Univ.), William F. Brechue (Kirksville College of Osteopathic Medicine) |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
In this study, we performed a 20-second march test (foot-stomping exercise) on women aged 55–90 years. We measured their movements using a general-purpose motion sensor to clarify the differences in characteristics between independent elderly individuals and those requiring care. We built an AI that learned time series from joints throughout the body and trained it to learn the movements of a healthy person in their 20s. The trained AI was given data from independent elderly individuals and elderly individuals requiring care as inputs, and the prediction error was calculated. The results showed that the prediction errors were significantly higher in the following order: healthy young adults, independent elderly, and elderly requiring care. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Time Series AI / Motion Sensor / Elderly requiring care / Independent Elderly / / / / |
Reference Info. |
ITE Tech. Rep., vol. 47, no. 35, HI2023-19, pp. 17-20, Nov. 2023. |
Paper # |
HI2023-19 |
Date of Issue |
2023-11-03 (HI) |
ISSN |
Online edition: ISSN 2424-1970 |
Download PDF |
|