Pre-impact Alarm System for Fall Detection Using MEMS Sensors and HMM-based SVM Classifier

Area of Science:

• Biomedical Engineering
• Gerontology
• Wearable Technology

Background:

• Accidental falls pose significant health risks, particularly for elderly individuals living alone, necessitating timely intervention systems.
• Existing fall detection methods often lack the accuracy and reliability required for effective elderly care.
• The development of proactive health monitoring solutions is crucial for improving the quality of life and safety of the aging population.

Purpose of the Study:

• To propose and evaluate a pre-fall alarm system designed for early detection of falling events in elderly individuals.
• To assess the reliability and accuracy of a novel fall detection algorithm integrated into a wearable device.
• To explore the potential of the system in monitoring physical function and providing timely assistance to prevent fall-related injuries.

Main Methods:

• Development of a wearable device to capture acceleration and angular velocity time series data during human motion.
• Utilizing a Hidden Markov Model (HMM)-based Support Vector Machine (SVM) classifier to distinguish between falls and Activities of Daily Living (ADLs).
• Conducting experiments with eighteen subjects wearing the device to collect motion data for algorithm training and validation.

Main Results:

• The fall detection system achieved a high accuracy rate of 94.91%.
• The system demonstrated excellent performance with 97.22% Sensitivity and 93.75% Specificity in fall event recognition.
• The proposed device is characterized by its small size, low power consumption, and ability to accurately recognize falls.

Conclusions:

• The developed pre-impact fall alarm system effectively detects falls with high accuracy, offering a promising solution for elderly safety.
• The system's ability to recognize fall events and its advantageous features (small size, low power) make it suitable for continuous monitoring.
• This technology holds significant potential for monitoring the physical function of the elderly population and enabling prompt assistance, thereby reducing fall-related complications.