A Novel Elderly Tracking System Using Machine Learning to Classify Signals from Mobile and Wearable Sensors

Area of Science:

• Gerontology and Health Informatics
• Machine Learning Applications in Healthcare
• Assistive Technology for Elderly Care

Background:

• Growing elderly population necessitates advanced health and activity monitoring systems.
• Existing systems are insufficient to meet the increasing demand for comprehensive elderly care.
• Need for integrated solutions covering activity tracking, geolocation, and personal information.

Purpose of the Study:

• To develop an innovative elderly tracking system integrating multiple technologies and machine learning.
• To enhance real-time monitoring capabilities for both indoor and outdoor environments.
• To provide a comprehensive solution for elderly care, including data collection, tracking, and emergency alerts.

Main Methods:

• Integration of multiple technologies for activity tracking and geolocation.
• Application of machine learning, specifically the k-nearest neighbor (k-NN) model, for activity classification.
• Collaboration with local agencies for system planning and development, including case study testing.

Main Results:

• The k-nearest neighbor (k-NN) model with k=5 demonstrated high effectiveness, achieving 96.40% accuracy in classifying nine distinct elderly activities.
• The developed system enables real-time monitoring, provides timely alerts, and displays elderly information spatially.
• The system facilitates emergency help requests via a messaging device.

Conclusions:

• The developed elderly tracking system effectively supports elderly care through data collection, real-time monitoring, and notifications.
• The system provides valuable supporting information to relevant agencies involved in elderly care.
• This integrated approach enhances the safety, independence, and quality of life for the elderly population.