In many countries facing the aged society, it is expected that the serious society problem like lonely death will be increasing. Therefore, a system watching over elder's life is very important. Then, we developed the system watching over elder's life using information about indoor ambient atmosphere changes. Indoor ambient atmosphere is changed by acted daily activities regardless of the age or sex of resident. Our system acquired such changes using 7 sensors (gas, temperature, humidity, brightness) and could recognize daily activities using information that extracted by observer. And, it could classify resident's daily life pattern using features that extracted by automatic feature extraction method from acquired ambient atmosphere changes. In this paper, at first, we describe an acquirement of indoor ambient atmosphere changes using multiple sensors and a classification of resident's daily activities. Next, we describe the method for detection of the information that related resident's daily activities from acquired indoor ambient atmosphere changes automatically, and the method for daily activity detection from integrated signal of multiple sensor signals. Finally, we describe result of classification of resident's daily life pattern using features extracted by automatic extraction method.
ASJC Scopus subject areas