Date of Award
Doctor of Philosophy (PhD)
Electrical & Computer Engineering
Modeling & Simulation Engineering
World Health Organization (WHO) data show that around 684,000 people die from falls yearly, making it the second-highest mortality rate after traffic accidents . Early detection of falls, followed by pneumatic protection, is one of the most effective means of ensuring the safety of the elderly. In light of the recent widespread adoption of wearable sensors, it has become increasingly critical that fall detection models are developed that can effectively process large and sequential sensor signal data. Several researchers have recently developed fall detection algorithms based on wearable sensor data. However, real-time fall detection remains challenging because of the wide range of gait variations in older. Choosing the appropriate sensor and placing it in the most suitable location are essential components of a robust real-time fall detection system.
This dissertation implements various detection models to analyze and mitigate injuries due to falls in the senior community. It presents different methods for detecting falls in real-time using deep learning networks. Several sliding window segmentation techniques are developed and compared in the first study. As a next step, various methods are implemented and applied to prevent sampling imbalances caused by the real-world collection of fall data. A study is also conducted to determine whether accelerometers and gyroscopes can distinguish between falls and near-falls.
According to the literature survey, machine learning algorithms produce varying degrees of accuracy when applied to various datasets. The algorithm’s performance depends on several factors, including the type and location of the sensors, the fall pattern, the dataset’s characteristics, and the methods used for preprocessing and sliding window segmentation. Other challenges associated with fall detection include the need for centralized datasets for comparing the results of different algorithms. This dissertation compares the performance of varying fall detection methods using deep learning algorithms across multiple data sets.
Furthermore, deep learning has been explored in the second application of the ECG-based virtual pathology stethoscope detection system. A novel real-time virtual pathology stethoscope (VPS) detection method has been developed. Several deep-learning methods are evaluated for classifying the location of the stethoscope by taking advantage of subtle differences in the ECG signals. This study would significantly extend the simulation capabilities of standard patients by allowing medical students and trainees to perform realistic cardiac auscultation and hear cardiac auscultation in a clinical environment.
In Copyright. URI: http://rightsstatements.org/vocab/InC/1.0/ This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).
Yhdego, Haben G..
"Wearable Sensor Gait Analysis for Fall Detection Using Deep Learning Methods"
(2023). Doctor of Philosophy (PhD), Dissertation, Electrical & Computer Engineering, Old Dominion University, DOI: 10.25777/91hp-ty43