Monitoring daily living activities of elderly people in a nursing home using an infrared motion-detection system.

We examined whether we could identify activity patterns of elderly people in a nursing home from sensor outputs of an infrared monitoring system. The subjects consisted of three elderly people. A single passive infrared sensor installed on the ceiling of each subject's usual dwelling room provided digital output whenever the subject moved. The subjects' actual daily activities were established from questionnaires with which patients documented their living patterns for each of 7 days. Activities were classed as sleeping, getting up/breakfast, indoor activities/going out, and dinner/going to bed. The mean +/- 2 standard deviations (SDs) of the sensor outputs on each day for each period of indoor activity was used to distinguish between normal and aberrant activities. Days on which sensor outputs exceeded the means +/- 2 SDs were regarded as atypical and were identified for each subject over a 28-day period. We were unable to determine the physical condition of the subjects on these atypical days. We were able to identify the pattern of daily indoor living activities and the duration of each class of activity using sensor outputs and a questionnaire. Days were assumed to be atypical when sensor outputs deviated from the normal pattern.