Detection of Faulty Measurements in WBANs using Gaussian Mixture Model and Ant Colony

Detection of Faulty Measurements in WBANs using Gaussian Mixture Model and Ant Colony

Osman Salem (LIPADE Laboratory, University of Paris Descartes, Paris, France), Yaning Liu (JCP-Connect, Rennes, France) and Ahmed Mehaoua (LIPADE Laboratory, University of Paris Descartes, Paris, France)
Copyright: © 2014 |Pages: 19
DOI: 10.4018/ijehmc.2014100102
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Wireless sensor networks are subject to different types of faults and interferences after their deployment. Abnormal values reported by sensors should be separated from faulty or injected measurements to ensure reliable monitoring operation. The aim of this paper is to propose a lightweight approach for the detection and suppression of faulty measurements in medical wireless sensor networks. The proposed approach is based on the combination of statistical model and machine learning algorithm. The authors begin by collecting physiological data and then they cluster the data collected during the first few minutes using the Gaussian mixture decomposition. They use the resulted labeled data as the input for the Ant Colony algorithm to derive classification rules in the central base station. Afterward, the derived rules are transmitted and installed in each associated sensor to detect abnormal values in distributed manner, and notify anomalies to the base station. Finally, the authors exploit the spatial and temporal correlations between monitored attributes to differentiate between faulty sensor readings and clinical emergency. They evaluate their approach with real and synthetic patient datasets. The experimental results demonstrate that their proposed approach achieves a high rate of detection accuracy for clinical emergency with reduced false alarm rate when compared to robust Mahalanobis distance.
Article Preview

1. Introduction

The Medical Wireless Sensor Networks (WSNs) or Wireless Body Area Networks (WBANs) are a set of wearable or implantable biomedical sensors with wireless transmission capabilities, used to collect vital signs (pulse, heart rate, blood pressure, oxygen saturation, body temperature, glucose level, galvanic skin ratio, electrocardiogram, electromyogram, etc.) from the monitored patients (Crosby, Ghosh, Murimi, & Chin, 2012), and to transmit collected data toward a central device (smartphone or tablet) (Movassaghi, Abolhasan, Lipman, Smith, & Jamalipour, 2014).

These sensors measure and transmit physiological data in real time, and allow remote and continuous health monitoring (in-home or in-work) over an extended period of time (Otto, Milenkovi, Sanders, & Jovanov, 2005). These sensors minimize the need of caregivers and allow monitored patient to continue living independent life. Their usage leads to the modernization of the way in which healthcare services are deployed and delivered. WSNs may also resolve the shortage of nursing for assisting elderly people in-home. Several clinical situations could be prevented or better monitored using WBANs, where the collected data in real-time can be used to follow the evolving state of remotely monitored patient, to early detect clinical emergency situation and to quickly react by taking the appropriate actions to save the life of the monitored patient (Ko, et al., 2010).

WBANs provide a mobile healthcare system used for remote and in-home monitoring, where the physiological data collected by sensors are transmitted to Local Processing Unit (LPU), which has more processing power, battery and transmission capabilities than sensors. The LPU must process collected measurements in real time, and raise an alarm for healthcare center (or family, neighbor, etc.) when it detects clinical emergency, in order to let them react by taking the necessary actions.

Therefore, pervasive healthcare services require the development of real-time applications for the detection of emergency situations, such as the detection of myocardial ischemia which precedes the heart attack (or infarction). The early detection prevents serious complications and damage of the heart by therapies using anticoagulants or Percutaneous Coronary Intervention (PCI) that reestablish the normal flow of blood in the obstructed coronary artery.

WBANs reduce healthcare cost and improve the usage of occupied beds in hospital by enabling the monitoring of chronicle and long term diseases outside institutions. They also enhance the life of monitored patients by allowing them to move freely and to achieve their daily life activities while being monitored. However, the use of WBANs is susceptible to several problems which range from reliability to security threats after deployment.

Sensors are subject to hardware and software faults, which are due to various reasons such as damaged device, calibration, battery exhaustion, or dislocation. Furthermore, with the small size of sensors and their underlying constrained resources, such as limited processing power, small memory and restricted transmission range, their transmitted data are extremely vulnerable to radio interference, environmental noise, function fault, breakdown, faulty measurements from badly attached sensor and malicious behavior. Consequently, the collected measurements are affected by noise and errors and have low quality and reliability.

The transmission of faulty measurements consumes energy of constrained sensors, and they might trigger a false alarm. Faulty measurements reduce the reliability and the accuracy of diagnosis results, and they affect the credibility of such monitoring system and prevent its deployment, where reliability is extremely important to ensure accuracy in the medical domain (Sahoo, 2012).

The use of WSNs in medical field has stringent requirements in terms of reliability and security. Security issues in WSNs should consider not only technical issues, but also social issues which might break the privacy of patient life. For example, the collected measurements might be accessed by an attacker which may (i) modify data, (ii) inject false data or (iii) replay previously registered data to threat patient's life. As the LPU must process the measurements in real-time to detect clinical deterioration, faulty measurements or maliciously injected data must be detected and isolated to increase the reliability of such monitoring system and to reduce false alarms.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 9: 4 Issues (2018): 1 Released, 3 Forthcoming
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing