Triaxial Accelerometer-Based Falls and Activities of Daily Life Detection Using Machine Learning

Abstract

<jats:p>The detection of activities of daily living (ADL) and the detection of falls is of utmost importance for addressing the issue of serious injuries and death as a consequence of elderly people falling. Wearable sensors can provide a viable solution for monitoring people in danger of falls with minimal external involvement from health or care home workers. In this work, we recorded accelerometer data from 35 healthy individuals performing various ADLs, as well as falls. Spatial and frequency domain features were extracted and used for the training of machine learning models with the aim of distinguishing between fall and no fall events, as well as between falls and other ADLs. Supervised classification experiments demonstrated the efficiency of the proposed approach, achieving an F1-score of 98.41% for distinguishing between fall and no fall events, and an F1-score of 88.11% for distinguishing between various ADLs, including falls. Furthermore, the created dataset, named “ShimFall&amp;ADL” will be publicly released to facilitate further research on the field.</jats:p>

Event-Centered Data Segmentation in Accelerometer-Based Fall Detection Algorithms

“… is a harmonic mean of sensitivity and precision and is often used as a single standard measure for evaluation of fall detection systems [ 33 , 40 , 41 ]. (from Activity Miss Rate) and (from Activity False Positive Rate) express percentage of records from a particular ADL or fall type that are misclassified.…”

Detection of Generalized Tonic Clonic Seizures and Falls in Unconstraint Environment Using Smartphone Accelerometer

“…In [4], tri-axial accelerometer data is collected from 35 healthy individuals performing ADLs including jumping, lying down, sitting, walking, Falls, etc. Spatial frequency domain features were extracted from the data and a final feature vector was generated as the concatenation of the computed features.…”

“…to determine type of activity or action accordingly [3]. HAR is one of the most active research areas with its applications in human computer interaction, healthcare, and security surveillance [4,5,6]. The accuracy of automated HAR is enhanced by using data from environmental and body worn sensors.…”

“…However, it offers a level of difficulty due to sensor motion, its placement, background management, and variations in the way activities are performed by an individual [9]. Monitoring of various ADLs is employed in treating various neurological disorders, detection of Falls and eventually helping to improve quality of life of individuals with various health issues [4] [5]. It is to be noted that several ADLs have common postures and levels of ambulation and it is very difficult to classify these activities correctly.…”

“…It influences individuals of all age groups generally causing SUDEP fatalities (0. [3][4][5][6] in every one of thousand individuals. In addition, SUDEP is also responsible for 8-12% of deaths in epileptic patients [15].…”