Wireless sensor networks for pervasive health applications
Majid Nabi · TU/e Research Portal · 2013
Wireless Sensor Networks for Pervasive Health Applications Wireless sensor networks (WSNs) enable many new applications. Sensor nodes are developing into tiny, light and inexpensive devices, paving the road for widespread use. Health is one of the domains that can greatly benefit from WSNs. Advances in ultra-low power tiny wireless (bio-)sensor devices allow deployment on (and in) a human body during daily life. This forms a Wireless Body Area Network (WBAN) to measure, process, and transmit a variety of physical and physiological data. The combination of WBANs and ambient WSNs provides an infrastructure for pervasive health applications such as elderly care, remote healthcare monitoring of patients with chronic diseases such as COPD, and post-surgery recovery monitoring. Such technology enables patients to live more independently and still receive sufficient care. It also reduces the cost of medical care. Exploiting well-known general-purpose protocol stacks in WSNs benefits from maturity and technology availability, but suffers from being sub-optimal for specific applications. The protocols need to be adapted to specifications of the target application to fulfill the application requirements. Mobility of individual nodes and of clusters of nodes (WBANs), heterogeneity of sensing characteristics and of quality-of-service (QoS) requirements, and dynamics in the network and its context are special characteristics of pervasive health application. Energy consumption of nodes should be minimized to extend the life-time of the networks and to reduce maintenance cost. This thesis makes the following contributions. An application-aware Medium Access Control (MAC) protocol, a data prioritization mechanism for data dissemination, and an adaptive intra-WBAN communication protocol are proposed, all aiming to prepare communication protocols for pervasive health applications. The performance of the protocols is studied through real-world experiments with a large-scale ambient WSN and several persons with WBANs. The experiments are complemented with computer simulations to evaluate the protocols in a wider range of use-cases. Models, tools and methods needed to assess performance of the proposed protocols, including a mobility model for WBANs, are developed. We consider WSNs that exhibit high mobility in the form of moving clusters of nodes (WBANs). The MAC layer needs to efficiently manage medium access in the presence