Great Northern Haven: From Raw Smart Home Data to Health Metrics

Background: 

Great Northern Haven (GNH) is a demonstration housing project consisting of 16 purpose built homes, each equipped with a combination of sensor and interactive technology to support Ambient Assisted Living for older adults, promote independent living and support active and healthy ageing. Data have been collected from 15 of the apartments, occupied by older adults, since June 2010. There are a total of 2,240 sensors and actuators throughout the development; the sensors include Passive InfraRed (PIR) sensors to detect motion, window and door sensors to detect openings and closings, electricity and water usage sensors, temperature and brightness sensors. Importantly, Great Northern Haven is not a pilot or trial project; realworld data is being captured from people living in their homes over extended periods. 

Methods: 

Algorithms have been developed to extract daily metrics from the raw smart home data including total daily activity, time spent outside the house, number of room transitions, time in bed, nocturnal restlessness and number and duration of bed exits. Multiple health questionnaire (cognition, depression, loneliness, quality of life, sleep quality) data, repeated at multiple time points, have been collected from the residents. 

Results: 

Comparisons of the extracted metrics were made across all residents and within each resident’s data as well as against the longitudinal health questionnaire’s data and observations by health care experts. Changes in health status (e.g. depression) may be detected using the ambient data (e.g. total activity and time spent outside the house) through comparisons made against the person’s own baseline data. 

Conclusions: 

Ambient sensing is a practical solution which can provide an insight into a person's daily patterns. Longitudinal derivations in these metrics may be used to inform health and wellness interventions.