Highly Accurate Bathroom Activity Recognition Using Infrared Proximity Sensors

Area of Science:

• Gerontology and Health Monitoring
• Biomedical Engineering and Sensor Technology

Background:

• A significant portion of the global elderly population suffers from physical or mental diseases, impacting their quality of life.
• Elderly individuals prefer to age in place, necessitating automated solutions for health monitoring to reduce reliance on human resources.
• Activities related to personal hygiene are strongly correlated with health status and can indicate the onset of disorders.

Purpose of the Study:

• To develop a simple, affordable, and reliable sensor platform for automatic detection of basic bathroom activities in elderly populations.
• To leverage sensor data to monitor health status and facilitate early detection of potential health issues.
• To support the goal of enabling elderly individuals to live independently in their homes for longer periods.

Main Methods:

• Development of a sensor platform utilizing infrared proximity sensors.
• Accurate recognition of basic bathroom activities, including toilet use and showering.
• Validation of the system's reliability through experimental testing and duration accuracy assessment.

Main Results:

• The developed sensor system achieved a high F-Score of 96.94% in recognizing bathroom activities.
• Collected activity durations were found to be approximately 6 seconds different from actual durations, confirming system reliability.
• The system demonstrates potential for accurate and unobtrusive health status monitoring in elderly individuals.

Conclusions:

• The infrared sensor platform offers a reliable and cost-effective solution for monitoring elderly individuals' health through activity recognition.
• The system supports independent living by providing automated health insights, reducing the need for constant human supervision.
• This technology can serve as an important tool for early detection of health decline in aging populations.