Extracting Routines of the Living Alone Occupant's Daily Activities Using Multiple Sequence Alignment

The increasing number of single member households is a critical issue worldwide, especially for the elderly. Since the living alone may be unaware of their health status or ways to improve their routines, a continuous... [ view full abstract ]

The increasing number of single member households is a critical issue worldwide, especially for the elderly. Since the living alone may be unaware of their health status or ways to improve their routines, a continuous monitoring system for healthcare would offer feedback. Assessing the adequacy of activities of daily living (ADL) can act as a measure of an individual's health status; previous research has focused on detecting the person’s daily activities and extracting the most frequently performed patterns using camera recording or wearable sensing techniques. However, extracting common patterns of an occupant's activities in the home, using existing methods, still requires more investigation to consider the spatio-temporal dimensions of human activities simultaneously. Though the multiple sequence alignment (MSA) has some advantages, such as containing the spatio-temporal data in sequence format inherently and finding the hidden patterns easily, little research has applied the MSA to extracting ADL routines at home. This research proposes a method to extract an occupant's ADL routines from a cumulative spatio-temporal data log of occupancy collected through a non-intrusive approach (i.e., tomographic motion detection system). The results from an occupant’s 14 day spatio-temporal activity log demonstrate that the proposed approach is capable of identifying routine patterns of an occupant’s daily activities and revealing the order, duration, and frequency of routine activities. Routine ADL patterns identified from the proposed approach are expected to provide a basis in detecting/evaluating abrupt or gradual changes of an occupant’s ADL patterns due to a physical or mental disorder, as well as to provide valuable information in home automation applications by allowing the prediction of ADL patterns.