04-10-2012, 11:12 AM
BODY SENSOR NETWORK – A WIRELESS SENSOR PLATFORM FOR PERVASIVE HEALTHCARE MONITORING
BODY SENSOR NETWORK.pdf (Size: 170.6 KB / Downloads: 141)
ABSTRACT
With recent advances in wireless sensor networks and embedded computing technologies,
miniaturized pervasive health monitoring devices have become practically feasible. In addition to
providing continuous monitoring and analysis of physiological parameters, the recently proposed
Body Sensor Networks (BSN) incorporates context aware sensing for increased sensitivity and
specificity. To facilitate research and development in BSN and multi-sensor data fusion, a BSN
hardware development platform is presented. With its low power, flexible and compact design, the
BSN nodes provide a versatile environment for wireless sensing research and development.
Introduction
Cardiovascular disease is the main cause of death in the UK and it accounts for 39% of all death
each year. Among patients who had heart attacks, about 30% of them died even before reaching to
the hospital [5]. Although heart attack can happen suddenly without apparent indications, cardiac
rhythm disturbances can often be found before the event. They can potentially be used as the
precursor to major cardiac episodes [3]. Currently, ECG (Electrocardiogram) Holter monitoring is
the most widely used technique for providing ambulatory cardiac monitoring for capturing rhythm
disturbances. A traditional Holter monitor can record up to 24 hours of ECG signals, and the
recorded data is subsequently retrieved and analyzed by a clinician. Due to the short duration
involved and the unknown context within which the ECG signal is captured, reliable interpretation
of the recorded data is always a challenge. To address these drawbacks, some more advanced ECG
monitoring systems are emerging. They can also detect and signal a warning in real-time if any
adverse event is captured [7]. Recent research has also focused on the development of wireless
sensor networks (WSN) and pervasive monitoring systems for cardiac patients. For example, a
number of wearable systems have been proposed with integrated wireless transmission, GPS
(Global Positioning System) sensor, and local processing [2,4,6,10]. Commercial systems are also
becoming available. For example, CardioNet provides a remote heart monitoring system where
ECG signals are transmitted to a PDA (Personal Digital Assistant) and then routed to the central
server by using the cellular network [6]. Pentland recently presented the wearable MIThril system
where ECG data, GPS position, skin temperature and galvanic skin response can be captured by a
PDA [4].
BSN Architectural Design
Fig. 2 illustrates the basic structure the BSN node (left) and its architectural design (right). The
BSN node uses Texas Instrument (TI) MSP430 16-bit ultra low power RISC processor with
60KB+256B Flash memory, 2KB RAM, 12-bit ADC and 6 analog channels (connecting up to 6
sensors). The wireless module has a throughput of 250kbps with a range over 50m. In addition,
512KB serial flash memory is incorporated in the BSN node for data storage or buffering. The BSN
node runs TinyOS by U.C. Berkeley, which is a small, open source and energy efficient sensor
board operating system. It provides a set of modular software building blocks, of which designers
could choose the components they require. The size of these files is typically as small as 200 bytes
and thus the overall size is kept to a minimum. The operating system manages both the hardware
and the wireless network—taking sensor measurements, making routing decisions, and controlling
power dissipation. By using the ultra low power TI microcontroller, the BSN node requires only
0.01mA in active mode and 1.3mA when performing computation intensive calculation like a FFT.
Prototype and Demonstration
With the proposed BSN architecture, a number of wireless biosensors including 3-lead ECG, 2-lead
ECG strip, and SpO2 sensors have been developed (Fig. 3a-c). To facilitate the incorporation of
context information, context sensors including accelerometers, temperature and humidity sensors
are also integrated to the BSN node. Furthermore, a compact flash BSN card is developed for
PDAs, where sensor signals can be gathered, displayed and analyzed by the PDA, as shown in Fig.
3d. Apart from acting as the local processor, the PDA can also act as the router between the BSN
nodes and the central server, where all sensor data collected will be transmitted through a
WiFi/GRPS network for long-term storage and trend analysis.
The proposed demonstration will illustrate the latest design of the BSN nodes and their
interoperability with other WSN platforms such as MicaZ and Telos. It will also provide a unique
hands-on experience of using the BSN nodes for constructing wireless sensing daughter-boards
during the exhibition. To establish the key features of the proposed platform, a live demonstration
will be set up during the exhibition where prototype physiological and context awareness sensors
will be worn to illustrate a clinical scenario of continuous patient monitoring under normal
physiological activities. In addition, we will demonstrate the sensor fusion environment by using
the real-time central database for trend analysis and data mining. Furthermore, we will also
demonstrate our newly developed STSOM and Bayesian context detection framework for outlining
the strength and research challenges of context aware sensing.