Related Experiment Video
Updated: Sep 1, 2025
08:05
Design and Analysis for Fall Detection System Simplification
Published on: April 6, 2020
10.8K
Deep Learning-Based Near-Fall Detection Algorithm for Fall Risk Monitoring System Using a Single Inertial Measurement
Summary
This study introduces a deep learning algorithm using a waist-worn sensor to accurately detect falls and near-falls. This technology can help prevent elderly injuries and monitor rehabilitation progress.
Area of Science:
- Gerontology
- Biomedical Engineering
- Artificial Intelligence
Background:
- Fall detection and prevention are crucial for elderly health.
- Near-fall monitoring offers insights into fall risks and rehabilitation status for individuals with balance issues.
Purpose of the Study:
- To develop a novel deep learning algorithm for classifying falls, near-falls, and activities of daily living (ADLs).
- To utilize a single inertial measurement unit (IMU) for precise fall risk assessment.
Main Methods:
- A modified directed acyclic graph-convolution neural network (DAG-CNN) was developed and optimized.
- Data was collected from 34 young participants performing 36 types of activities using an IMU (accelerometer and gyroscope).
Main Results:
- The modified DAG-CNN achieved higher accuracy than traditional CNNs for fall, near-fall, and ADL classification.
- Near-fall detection accuracy exceeded 98% by integrating gyroscope and accelerometer data.
- Combined sensor data (acceleration and angular velocity) improved model performance.
Conclusions:
- The developed deep learning model enables accurate classification of falls, near-falls, and ADLs.
- This technology can provide valuable data for preemptive fall risk management and quantitative rehabilitation evaluation in the elderly.
