Discovering Activities to Recognize and Track in a Smart Environment

Area of Science:

• Smart Home Technology
• Machine Learning
• Health Monitoring

Background:

• Smart homes offer opportunities for health monitoring and assistance to individuals living independently.
• Current activity recognition approaches require pre-selected activities and labeled training data.
• There is a need for automated methods to track daily routines for functional health assessment.

Purpose of the Study:

• To introduce an automated approach for activity tracking in smart homes.
• To identify frequent activities naturally occurring in an individual's routine.
• To enable monitoring of functional health and detection of lifestyle changes.

Main Methods:

• Developed an automated activity mining and tracking approach.
• Algorithms designed to identify frequent activities from sensor data.
• Validated algorithms using data collected in physical smart environments.

Main Results:

• Successfully identified frequent activities in smart home residents' routines.
• Demonstrated the capability to track the occurrence of regular activities.
• Validated the effectiveness of the developed algorithms.

Conclusions:

• Automated activity tracking in smart homes is feasible for functional health monitoring.
• The proposed approach can detect changes in individual patterns and lifestyles.
• This technology holds promise for supporting independent living through continuous health assessment.