A Case of Long-Term ADL Inference and Modeling

In this research we propose to use location-based technology to assist in tracking elders’ Activities of Daily Living (ADLs) to early detect the problem of degeneration of physical functions and cognitive abilities. In collaboration with National Taiwan University Hospital, we deploy a 45-node WSN-RTLS system in its elder care center. From December 2008 to July 2009, we collect location traces and investigate the ADLs of volunteering elders. A novel expert-based system is proposed to automatically infer ADLs in real time merely from location information of the elder. Two probability-based metrics are proposed to indicate how likely an ADL is to be followed by another ADL and how often a daily ADL pattern is likely to occur respectively. The proposed mechanisms are validated by nursing staffs via questionnaires. The validation results affirm the feasibility of automated ADL inference, which in turn may offer valuable services not only to medical staffs but also to family members of the elders. Keywords-ADL Inference; Indoor Localization; Wireless Sensor Network; Elder Care