An Energy-Efficient Multi-Tier Architecture for Fall Detection on Smartphones

Abstract

Automatic detection of fall events is vital to providing fast medical assistance to the causality, particularly when the injury causes loss of consciousness. Optimization of the energy consumption of mobile applications, especially those which run 24/7 in the background, is essential for longer use of smartphones. In order to improve energy-efficiency without compromising on the fall detection performance, we propose a novel 3-tier architecture that combines simple thresholding methods with machine learning algorithms. The proposed method is implemented on a mobile application, called uSurvive, for Android smartphones. It runs as a background service and monitors the activities of a person in daily life and automatically sends a notification to the appropriate authorities and/or user defined contacts when it detects a fall. The performance of the proposed method was evaluated in terms of fall detection performance and energy consumption. Real life performance tests conducted on two different models of smartphone demonstrate that our 3-tier architecture with feature reduction could save up to 62% of energy compared to machine learning only solutions. In addition to this energy saving, the hybrid method has a 93% of accuracy, which is superior to thresholding methods and better than machine learning only solutions.

Accelerometer-Based Fall Detection Using Machine Learning: Training and Testing on Real-World Falls

“…To evaluate the added value of the proposed multiphase features, we also compared the classifiers’ performance using the proposed features with the classifiers’ performance using conventional features that are commonly used in fall detection studies [ 13 , 25 , 33 , 37 ]. These conventional features were max, min, mean, and standard deviation of a x , a y , a z , and Norm acc .…”

“…Several researchers have already used machine learning for fall detection [ 29 ]. Different classifiers have been used to detect falls, such as support vector machines (SVM) [ 19 , 25 , 30 , 31 , 32 , 33 , 34 , 35 , 36 ], Naïve Bayes [ 33 , 36 , 37 ], logistic regression [ 30 , 34 , 36 ], k-nearest neighbors (KNN) [ 31 , 33 , 34 , 36 , 38 ], decision trees and random forests [ 33 , 36 , 39 ], artificial neural networks, and deep learning [ 40 , 41 , 42 ]. However, in most of these studies, learning and testing were performed with simulated falls.…”

Using Temporal Covariance of Motion and Geometric Features via Boosting for Human Fall Detection

“…Shi et al proposed a human fall detection algorithm for classifying the human motion using the J48 decision tree classifier and achieved a sensitivity of 98.9%, a specificity of 98.5% and an overall accuracy of 98.6% [ 36 ]. In 2017, Guvensan et al developed a system that implements the decision tree learning algorithm of J48, using five features, for detection of fall events [ 37 ]. Motivated by these algorithms, we chose the boosted J48 classifier due to its significantly higher F-measure, low computational cost and robustness to outliers and reduction of the feature space.…”

A Survey on Recent Advances in Wearable Fall Detection Systems

“…Guvensan et al [38] focus on energy efficiency in fall detection. A combination of threshold-based method and ML-based algorithms-K-Star, Naïve-Bayes, and J48-was applied to data generated from a 3D accelerometer attached to a smartphone.…”