Activity density map visualization and dissimilarity comparison for eldercare monitoring

Area of Science:

• Gerontology
• Biomedical Engineering
• Data Science

Background:

• Aging in place is a growing trend, necessitating innovative methods for monitoring senior health.
• Traditional health monitoring can be intrusive and may not capture subtle changes in daily activity.
• Passive infrared (PIR) motion sensors offer a non-intrusive way to collect data on seniors' daily routines.

Purpose of the Study:

• To develop and evaluate a methodology for analyzing PIR motion sensor data to identify health-related activity patterns in seniors.
• To enable early detection of health changes by recognizing deviations in established activity patterns.
• To provide clinicians with tools for proactive health management in aging populations.

Main Methods:

• Development of a data analysis methodology for PIR motion sensor logs.
• Creation of activity density maps for visualizing senior residents' movement patterns.
• Utilizing fuzzy rules to accurately determine time away from home.
• Employing texture features to compute dissimilarity between density maps for change detection.

Main Results:

• Demonstrated the ability to visualize activity levels using color-coded density maps.
• Successfully identified changes in activity patterns indicative of potential health issues.
• Validated the methodology through three case studies of elderly residents, tracking physical, cognitive, and mental health changes.
• Activity density maps and dissimilarity measures are actively used in an aging in place community for early illness detection.

Conclusions:

• The proposed methodology effectively analyzes PIR sensor data to monitor senior activity patterns.
• Density maps and dissimilarity measures serve as valuable tools for early detection of health declines in elderly individuals.
• This approach supports proactive healthcare and enhances the quality of life for seniors aging in place.