Abstract
Fall is an increasing problem as people ageing. It may happen to anyone, but their incidence does increase with age. Hence, the elderly will be facing catastrophic consequences due to falls. Nevertheless, there are still vulnerable in its accuracy in categorizing and differentiating the Activities Daily Living (ADL) and falls as most of the existing systems cause false alarm. This paper presents the research and simulation of wearable device-based fall detection approach by addressing the building of wearable device-based fall detection system for elderly care by using mobile devices. Two main phases involve in this research: online phase and offline phase. Online phase covers in data acquisition step whereby the raw data of simulated fall by participants is collected via built-in-tri-axial accelerometer in a smartphone, then automatically sent towards the computer via wireless communication. Meanwhile, offline phase covers data pre-processing, feature extraction and selection and data classification where these steps are handled in offline mode. Support Vector Machine (SVM) classifier was employed, and evaluated in the analysis. Overall accuracy rate, sensitivity, specificity as well as False Positive Rate (FPR) and False Negative Rate (FNR) were calculated. The findings suggest that SVM with Polynomial (order 5) method which achieved 68.91% overall accuracy as well as producing only 24.46% FPR is the most precise model for fall detection system in this paper. This approach has the potential to be implemented and deploy in real mobile application in future.
Highlights
The falls are an increasing problem as people age
In order to find the most accurate method for wearable device-based fall detection system in detecting all falls which preventing false negative events as well as preventing from generating false alarm, this research is focusing on several types of classification algorithms which are proved to have high acceptance towards the sensitivity, specificity as well as overall accuracy including Artificial Neural Network (ANN) algorithm, K-Nearest Neighbor (KNN) algorithm and Support Vector Machine (SVM) algorithm
Several types of kernel function methods are selected to be implemented in the SVM algorithm, including the default kernel, Linear, as well as non-linear such as Gaussian/ Radial Basis Function (RBF) as well as Polynomial kernel which are further divided into several orders (2, 5, 7, and 10)
Summary
The falls are an increasing problem as people age. They have been identified as one of the most prevalent public health problems facing the elderly as they usually associated with high morbidity and mortality, a public health concern. Recent researches explore about using wearable device-based approach as it uses sensors embedded to the garments to detect the posture and motion of the body of the wearer and use a classifier to detect fall. This approach is flexible for any environment. It is easy to use as the embedded sensor is small and light [9] This approach relatively low cost due to the complexity of the device that's being used [10].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: International Journal of Engineering & Technology
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
