Loading...
Flexible, wireless smart wearable microelectronics for fall risk monitoring among older adults
Jabr, Mariam
Jabr, Mariam
Citations
Altmetric:
Files
Loading...
Abstract.pdf
Adobe PDF, 49.57 KB
Authors
Other Names
Location
Time Period
Advisors
Original Date
Digitization Date
Issue Date
2022-04-15
Type
Abstract
Genre
Keywords
Subjects (LCSH)
Citation
Jabr, Mariam. 2022.
Flexible, wireless smart wearable microelectronics for fall risk monitoring among older adults -- In Proceedings: 21st Annual Undergraduate Research and Creative Activity Forum. Wichita, KS: Wichita State University, p. 15
Abstract
Falling presents a serious health issue and financial burden for elderly people
aged 65 years or older. According to the WHO Global Report, 28%-35% of adults 65
years or older experience fall related injuries more than once per year. The CDC has also
reported that more than $50 billion is spent on fall related injuries every year, which calls
for the need of an automatic fall detection system for older adults. Current fall detection
systems are classified into wearable and non-wearable systems. Non-wearable systems
include camera-based systems; however, these systems have very complex setups, are
highly expensive and are area constraint. Wearable systems use accelerometers and
gyroscopes to obtain motion data; however, they are usually bulky and/or visually
displeasing for older adults. Therefore, the overall objective of this research is to design a
fall detection system using flexible, wireless smart micro-electronics which offer accurate
fall detection and user comfort for long-term use due to their skin-like properties. To
achieve this objective, the following specific aims were accomplished. 1) The smart skinwearable
device was designed and fabricated using microfabrication techniques. 2)
Motion data was collected from 10 participants (5 younger adults aged 20-30 years and 5
older adults aged 65 years or older). Participants were instructed to perform 5 different
human activities (falling, running, sitting, walking, and stairs to collect linear and
rotational motion data. 3) Various deep-learning models (CNN, LSTM, CNN-LSTM, and
ConvLSTM) were explored to train the data for accurate fall detection. The highest
accuracy so far achieved among all models is 89%. Current work on this study includes
optimizing hyperparameters of deep-learning models to achieve highest performance and
accuracy. Deep-learning models will be compared to create the highest performance
algorithm in order to achieve the automatic fall detection system.
Table of Contents
Description
Presented to the 21st Undergraduate Research and Creative Activity Forum (URCAF) held at the Rhatigan Student Center, Wichita State University, April 15, 2022.
Publisher
Wichita State University
Journal
Book Title
Series
URCAF
v.21
v.21