Sensor
Networks
I attended an
interesting session on sensor network. It was organized by Institute of
Engineers.
The topic was
“Experiences, strategies and challenges in building wireless sensor
network”.
Speaker: Dr.
S.S. Inynger [email protected] Chairman / Roy Paul
Daniels Professor of Computer Science, Louisiana State University
First of all
he emphasized with some references that Bangalore is the best city for R&D
and IT industry compare to US.
His team has
started research on this topic since 1991. Nowadays sensor network, ZigBee has
become buzzwords. Ember is pioneer company in this field. We have already seen
networks of sensors like smoke detectors/heat detectors.
Architecture:
Basic
architecture is distributed sensors provide information to sink. Sink can be
connected via Internet to Task Manager. There are large numbers of sensors also
called MOTE, form a pure random wireless network without any topology. We can
have biomedical sensors too. It is a dynamic, scalable, ad-hoc network of self
configurable, self healing sensors with multi hop adaptive routing.
Specifications:
Maximum size
of sensor is 2”, minimum as small as dust particle.
Range of
communication 10ft to 100ft.
Battery life:
2 hours
CPU clock
frequency: 125 MHz
RAM: 4kb to
10kb
Cost: $10 to
$40. Low cost. So they are throw away sensors
It has
interrupt routine, its own Operating System. In one application memory is
divided in 3.5kb routing table/algorithm and 4.5kb for application.
One sensor is
a programmable device. It consists of sensing element, battery, processor,
memory and communication circuit.
Applications:
Medical
application to monitor parameters inside the body
Object
tracking
Environment
monitoring
MEMS = Micro
Electro-Mechanical System
Monitor
migration of birds
Forest
wild-land fire management system
Home care
system
Military
applications
To track
terrorist
US government
is using it to find out Osama Bin Laden
Challenges:
Major
challenge is extend battery life. The sensors go to sleep mode whenever needed.
It is also vulnerable to virus/warm attack. Here virus tries to generate dummy
data to drain the battery. As per Murphy’s
law, the sensor stops working exactly at the moment, when the monitored
object is nearby :) There are many redundant data to be filtered out. The
measured value is not accurate. It does not provide deterministic value of
monitored parameter.
At last it was
an interactive question answer session also. Audience also asked about why US
government fails to locate Osama Bin Laden and why Indian government fails to
locate Virappan?
Note: I have just described my experience about this session. I
tried my best, yet the information presented here may be incorrect. There is not
any motive to promote any technology, brand name or institute. .