Activity recognition based on a small scale sparse sensor network

2018 Fiscal Year Final Research Report

• Project/Area Number: 16K00334
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Soft computing
• Research Institution: The University of Aizu
• Principal Investigator: ZHAO QIANGFU 会津大学, コンピュータ理工学部, 教授 (90260421)
• Project Period (FY): 2016-04-01 – 2019-03-31
• Keywords: 人工知能 / 機械学習 / スマートホーム / 見守りシステム / プライバシー保護 / 在宅見守り

Outline of Final Research Achievements

In this research, we developed a monitoring system for taking care of elderly people who lives alone or lives with his/her spouse only. The main results obtained are as follows. 1) We have confirmed that a sparse sensor matrix can recognize the position and activity strength of the resident(s) very accurately, and we have proposed recognition methods based on machine learning; 2) We have proposed a video-based technique for automatic data labeling, and this technique makes it possible to recognition the position / activit strength in realtime; 3) To improve the usability of the sensor matrix, we have designed several sensor modules, and finally proposed a compact module that can be installed in a normal living room easily. Based on the results, we are going to commertialize the sensor module, and provide privacy preserving monitoring services.

Free Research Field

人工知能

Academic Significance and Societal Importance of the Research Achievements

日本のような超高齢者社会において、効率よく高齢者を見守るためには、ＡＩが必要となる。本研究の成果は、プライバシーを守りながら、高齢者の日常活動をある程度見守ることができる。特に提案したセンサーモジュール（出願済み）は、通常の部屋に設置しやすい特徴がある。それを制御する機械学習モデルと合わせれば、近い将来の実用化が期待できる。