Abstract:Introduction: Falls and gait disorders often result in hospitalization and immobilization. Near-falls may be one of the earliest signs of increased fall risk. In the literature, several sensor positions are used for fall detection, but few studies include the head as a sensor position. Hearables and hearing aids are increasingly equipped with inertial measurement units (IMUs) and are therefore of particular interest for measuring the risk of falling in everyday life. Methods: Therefore, we investigate the suitability of the ear as a sensor position for near-fall detection in comparison to the standard sensor positions. The motion data of one study participant (female, 63 years) was exemplary analyzed. The participant walked at her individually preferred gait speed on a perturbation treadmill while nine different perturbations (anterior-posterior, medio-lateral and pitch) were applied with a time interval of 20-30 seconds. We used seven IMUs during the measurement at the positions ear, sternum, lumbar, wrist (left/right), foot (left/right). Results: The absolute acceleration signals at the seven different positions show the periodicity of the normal gait before the perturbation. During and after the perturbation changes in the motion pattern can be seen, whereby the response to the perturbation occurs with a slight time lag. The Pearson correlations show that the sensor positions sternum, lumbar and ear correlate well with each other and thus show similar signal characteristics in the reaction to this perturbation. Conclusion: This provides evidence that the ear sensor position is at least comparable to the preferred sensor positions in the literature on the torso. However, these results were obtained under laboratory conditions. Further research is needed to investigate the sensor position at the ear in everyday life.

Comparison of machine learning approaches for near-fall-detection with motion sensors

Deep Learning-Based Near-Fall Detection Algorithm for Fall Risk Monitoring System Using a Single Inertial Measurement Unit

Accelerometer-Based Fall Detection Using Machine Learning: Training and Testing on Real-World Falls

Fall Detection for Elderly People Using Machine Learning

Inertial Sensing-Based Pre-Impact Detection of Falls Involving Near-Fall Scenarios

Development of a Strategy to Predict and Detect Falls Using Wearable Sensors

A Machine Learning Approach for Fall Detection Based on the Instantaneous Doppler Frequency

A Machine Learning Multi-Class Approach for Fall Detection Systems Based on Wearable Sensors with a Study on Sampling Rates Selection

Machine learning algorithms for fall detection using kinematic and heart rate parameters-a comprehensive analysis

Towards the Application of Hearables for Near-Fall Detection

Machine Learning Prediction of Fall Risk in Older Adults Using Timed Up and Go Test Kinematics

Analysis of waist and wrist positioning wearable machine learning models to detect falls

Machine Learning based Human Fall Detection System

Fall Detection: Threshold Analysis of Wrist-Worn Motion Sensor Signals

Deep Learning to Predict Falls in Older Adults Based on Daily-Life Trunk Accelerometry

Fall Risk Assessment in Stroke Survivors: A Machine Learning Model Using Detailed Motion Data from Common Clinical Tests and Motor-Cognitive Dual-Tasking

Accuracy Comparison of ML-Based Fall Detection Algorithms Using Two Different Acceleration Derived Feature Vectors

Fall Detection in the Elderly using Different Machine Learning Algorithms with Optimal Window Size

Detection Of Human Fall Using Floor Vibration And Multi-Features Semi-Supervised Svm

Comparison of four machine learning algorithms for a pre-impact fall detection system