02-06-2012, 12:23 PM
Using virtualisation on mobile phones for remote health monitoring
Using virtualisation on mobile phones.pdf (Size: 681.26 KB / Downloads: 197)
Abstract
We present vNurse, a system based on a smartphone
platform that permits comprehensive, secure and modular patient
remote monitoring outside a clinical environment, e.g. in the
home. Using both virtualisation of the phone OS and virtual
mobile networks of sensors with full Internet Protocol (IP)
connectivity, we enable real-time remote sensor readings of
patient Wireless Body Area Networks (WBANs) to be stored.
INTRODUCTION
vNurse builds upon and extends technologies that are inexpensive
and readily available to allow for better patient
monitoring. Our proof-of-concept system leverages three technologies,
(i) virtualisation; (ii) network mobility; and (iii)
wireless sensor networking. We first utilise existing wireless
sensors to collect relevant patient sensor readings, such as
temperature, heart-rate and environmental readings such as
GPS location. These readings are streamed ‘live’ from the
patient to the healthcare practitioner, or stored on a smartphone
device (which acts as both a data collector and mobile router)
for later upload or retrieval on request. The smartphone
connections can be maintained even while the patient is mobile
(within the home or travelling outside), permitting 24-hour
monitoring. This functionality is achieved by aggregating the
WBAN sensors as a single mobile network, utilising mobility
protocols such as Mobile IP1 and NEMO2 to achieve this.
PROPOSED ARCHITECTURE
Overview
Our architecture (See Figure 1) has been designed to solve
the problem of remote monitoring in eHealth scenarios. This
system comprises of four main components (i) WBAN, (ii)
virtualisation, (iii) network mobility and (iv) egress connectivity.
We are proposing to use the smartphone as a data collector,
aggregating and partially processing (e.g. [1]), data gathered
from the sensors, and as a mobile router for the sensors. It
will host a wireless network via its WiFi interface to provide
network connectivity to local wireless sensors (and of course,
it can use other technology such as bluetooth). It will also
provide uplink capability via the existing 3G interface. As
patients roam, we leverage the prevalence of 3G coverage to
provide network connectivity at all times.
EVALUATION
Power Consumption
To evaluate the feasibility of vNurse we have examined the
power consumption of the smartphone in different configurations
and scenarios during data collection. Because we do
not want to bias our evaluation with the power effects of any
network handovers, in the following tests, the smart phone
is in a stationary position. Our main method of collecting
battery information has been through the Python API on the
Android Scripting Environment (ASE) 8. This can be installed
on Android platforms via the Android Market. Because we did
not wish to run additional Python scripts outside of the Debian
VM image, we exported the ‘AP-PORT’ values into the VM.
This made it possible to access the Python Android API from
within the shell of the Debian file system. Using this API we
are able to get information about the battery levels in terms of
an overall remaining percentage and also in terms of power.
One issue using this API is that the LCD screen has to remain
lit in order for the sensor readings to be collected.