Artificial intelligence approach for detecting and classifying abnormal behaviour in older adults using wearable...

Area of Science:

• Gerontology
• Computer Science
• Biomedical Engineering

Background:

• Increasing global elderly population necessitates improved elder care solutions.
• Nursing homes face challenges in providing adequate care for a growing number of older adults.
• Existing monitoring systems may lack real-time detection and classification capabilities for abnormal behaviors.

Purpose of the Study:

• To propose and evaluate a smart wearable device for detecting and classifying abnormal behaviors in nursing home residents.
• To leverage artificial intelligence (AI) for enhanced monitoring and safety of older adults.
• To lay the groundwork for AI-based technological advancements in elder care.

Main Methods:

• Utilized a smart wearable device with a triple-axis acceleration sensor for motion data collection.
• Employed Zigbee for wireless data transfer within a wireless sensor network (WSN).
• Implemented a Backpropagation (BP) neural network for intelligent recognition and classification of behaviors using simulated data.

Main Results:

• The system successfully detected and classified routine motions and abnormal behaviors, including falls.
• Simulations demonstrated the effectiveness of the AI-based system in detecting behavioral changes.
• The device provides indoor positioning, behavior detection, and classification capabilities.

Conclusions:

• The proposed smart wearable device offers a viable solution for real-time monitoring of older adults in nursing homes.
• AI-powered behavior detection and classification can significantly improve resident safety and care.
• This technology represents a step forward in implementing advanced AI solutions for elder care facilities.