FINDING LONG TERM TENDENCIES IN DAILY ACTIVITIES OF PEOPLE WITH DEMENTIA BASED ON LOCATION DATA

Abstract

The development of radio frequency identi?cation (RFID) technology makes it possible to track the daily activities of residents when installed in a residential home. We studied what valuable information can be extracted from location data of the daily activities of residents collected over several years. We installed antennas into a group home housing six people with dementia and asked them to wear slippers into which RFID tags were inserted. Data was collected over thirty months. We found a seasonal trend with all residents but one person who was wheel-chaired. For example, their activity level was lowered during summer. We identi?ed a resident whose activity was in rapid decline three months prior to her death. For another resident who fell down, we identi?ed a rapid decline in activity decline a week prior to the incident. Both also showed a change in their behavioral patterns. Long-term data collection is useful for assessing residents' health conditions. With sufficient data, we can detect abnormalities.