Recognition of false alarms in fall detection systems

Abstract

Abstract-Falls are a major cause of hospitalization and injuryrelated deaths among the elderly population. The detrimental effects of falls, as well as the negative impact on health services costs, have led to a great interest on fall detection systems by the health-care industry. The most promising approaches are those based on a wearable device that monitors the movements of the patient, recognizes a fall and triggers an alarm. Unfortunately such techniques suffer from the problem of false alarms: some activities of daily living are erroneously reported as falls, thus reducing the confidence of the user. This paper presents a novel approach for improving the detection accuracy which is based on the idea of identifying specific movement patterns into the acceleration data. Using a single accelerometer, our system can recognize these patterns and use them to distinguish activities of daily living from real falls; thus the number of false alarms is reduced.

An On-Node Processing Approach for Anomaly Detection in Gait

“…The Average absolute Acceleration Variation (AAV) has been previously proposed to improve the accuracy of fall detection systems [34], [35]. It is found according to the following equation:…”

A supervised approach to automatically extract a set of rules to support fall detection in an mHealth system

“…Rather, these rules are laboriously generated by experts who examine fall recordings, hypothesize the higher importance of some parameters with respect to others, and consequently write some rules containing only those parameters and the values of some related thresholds, which have to be hypothesized as well. As an example, in [15] a human-conceived rule says that "a fall event is detected when acceleration magnitude is greater than 3 g and its peak is followed by a period, lasting at least 1200 milliseconds, characterized by the absence of peaks greater than the threshold". These rules are then tested on the fall database, and, if their classification ability is not satisfactory, a new set of rules has to be proposed by the experts.…”

“…[11][12][13][14][15], make reference to rules that are not automatically extracted by artificial intelligence tools. Rather, these rules are laboriously generated by experts who examine fall recordings, hypothesize the higher importance of some parameters with respect to others, and consequently write some rules containing only those parameters and the values of some related thresholds, which have to be hypothesized as well.…”

“…The third approach is often reported as offering several advantages, ranging from the possibility of monitoring people also in an outside environment, to low cost, and ease of use. Several recent papers in the literature [6,[8][9][10][11][12][13][14][15][16] deal with the issue of fall detection by means of wearable devices. Quite often, methods based on this approach have also been shown to be capable of providing a higher accuracy.…”

WSN4QoL: A WSN-Oriented Healthcare System Architecture

“…The authors in [42] use a pair of Shimmer motes [43]. Shimmer is a wireless sensor platform programmed in TinyOS [21], characterized by a small form factor, that can record and transmit physiological and kinetic data in real time using the most well-known communication technologies, such as Bluetooth or IEEE 802.15.4.…”