A comparison of accuracy of fall detection algorithms (threshold-based vs. machine learning) using waist-mounted tri-axial accelerometer signals from a comprehensive set of falls and non-fall trials

Abstract

Falls are the leading cause of injury-related morbidity and mortality among older adults. Over 90 % of hip and wrist fractures and 60 % of traumatic brain injuries in older adults are due to falls. Another serious consequence of falls among older adults is the 'long lie' experienced by individuals who are unable to get up and remain on the ground for an extended period of time after a fall. Considerable research has been conducted over the past decade on the design of wearable sensor systems that can automatically detect falls and send an alert to care providers to reduce the frequency and severity of long lies. While most systems described to date incorporate threshold-based algorithms, machine learning algorithms may offer increased accuracy in detecting falls. In the current study, we compared the accuracy of these two approaches in detecting falls by conducting a comprehensive set of falling experiments with 10 young participants. Participants wore waist-mounted tri-axial accelerometers and simulated the most common causes of falls observed in older adults, along with near-falls and activities of daily living. The overall performance of five machine learning algorithms was greater than the performance of five threshold-based algorithms described in the literature, with support vector machines providing the highest combination of sensitivity and specificity.

A User-Adaptive Algorithm for Activity Recognition Based on K-Means Clustering, Local Outlier Factor, and Multivariate Gaussian Distribution

“…The experimental results showed that their proposed method has low computational complexity and is robust among embedded systems. Khojasteh et al [25] compared the performance of threshold-based algorithms and various machine learning algorithms in detecting falls, using data from waist-located tri-axial accelerometers. The experimental results showed that the machine learning algorithms outperformed the threshold-based algorithms.…”

Fall Detection using Lifting Wavelet Transform and Support Vector Machine

“…However, to detect weather a fall occurred relies on the performance of the detection mechanism. The most common and simplest fall detection is the threshold method [8]. Nevertheless, the performance heavily depends on the fixed threshold level.…”

“…Hence, it is rarely used alone, and often combined with other machine learning methods such as decision tree (DT) [9], [10], artificial neural networks (ANN) [11], hidden Markov model (HMM) [12] and Support Vector Machine (SVM) [4], [14], [15] can be combined to outperform the threshold method [8], [14]. Among the machine learning methods, SVM was found the most robust for fall detection when compared to other methods such as threshold-based methods and the decision tree method [8]. However, most works which deploy SVM for fall detection use time-series features [8], [16].…”

“…Among the machine learning methods, SVM was found the most robust for fall detection when compared to other methods such as threshold-based methods and the decision tree method [8]. However, most works which deploy SVM for fall detection use time-series features [8], [16]. It was found that SVM fall detection performance can be improved by a combination of time and frequency domain features [4].…”

Combining novelty detectors to improve accelerometer-based fall detection

“…The authors used directly the acceleration values and the signal magnitude area as features. In [3], five threshold-based algorithms and five machine learning algorithms were compared. The overall performance of the machine learning algorithms was greater.…”