Optimizing Energy Efficiency in Wireless Sensor Networks

eggplantcinnabarΚινητά – Ασύρματες Τεχνολογίες

21 Νοε 2013 (πριν από 3 χρόνια και 8 μήνες)

74 εμφανίσεις

Optimizing Energy Efficiency in Wireless Sensor Networks


Almohanad S. Fayez

(afayez@vt.edu)

and Theresa
M.
Nelson

(tnelson@vt.edu)
1

Michael I. Brownfield,
Electrical and Computer Engineering, Virginia Tech

Nathaniel Davis IV,
Electrical and Computer Engin
eering, Virginia Tech


With the progression of computer networks extending boundaries and joining distant locations,
wireless sensor networks (WSNs) emerge as the new frontier in developing opportunities to collect and
process data from remote locations.
WSNs rely on hardware simplicity to make sensor field deployments
both affordable and long
-
lasting
,

without any maintenance support.
Nodes used in WSN
s
,
also known as
mo
t
es
,

are self
-
sustained; they have a built
-
in radio, microcontroller, RAM, and flash m
emory.
WSN
designers strive to extend network lifetimes while meeting application
-
specific throughput and latency
requirements. Effective power management places sensor nodes into one of the available energy
-
saving
modes based upon the sleep period durati
on
, opportunities to turn off

the radio,
and th
e current state of the
radio. Although t
he new generation of sensor platform radios with a 250

kbps data rate will not provide
adequate time to completely power off the radio between most 128
-
byte IEEE 802.15
.4 transmissions, but
the nodes may be able to transition to an intermediate power level. This
research

characterizes the newest
generation of WSN radio power levels, state transition times, and state transition energy costs.
Incorporating these results
into an embedded energy consumption model implemented in OPNET
, a
network simulation program,

provides increased accuracy for simulating node radio state transitions.
Additionally, the
research

proposes a new radio power management

(RPM) algorithm which o
ptimizes
radio sleep transitions.

In order to
accurately

characterize the motes,

we produced our own mote parameter
measurements,

rather than

using the provided spec
ification

sheet

values
. The voltage
consumption levels

w
ere

measured during
the
various op
erational states

in motes
: transmit, receive, and
low
power
. Additional
transition costs are associated with each of the aforementioned states.

Motes are low power consuming
devices
,

which means that voltages consumed during the various states
are very sm
all. In order to achieve
more accurate consumption readings from these states, an amplifying circuit was constructed.

A sample
mote characterization is shown in Figure 1; the current consumption levels were derived from voltages
measured
on

a
Telos A

mote
.




Using OPNET, we have coded various WSN specific MAC protocols, protocols that dictate how
packets are transmitted
in networks,

in order to evaluate various WSN scenarios. The goal of the model
was to use
obtained
hardware characterization
to

develop

an energy model in OPNET
which

should
allow
s

evaluating the
efficiency

of various MAC protocols and hardware technology.
By knowing how much
energy is available for each mote,
usually
2 AA batteries, the energy model account
s

for how much energy
motes co
nsume which was used
during in simulations to

determin
e

how many days would motes last before
they deplete their
batteries
.
The simulations
are

composed
of
20 motes; in each simulation motes’ utilize
various MAC protocols under various
traffic conditions.

The simulations help in evaluating
two things,
how
energy
efficient the motes are and how efficient the various WSN MAC protocols are. The developed
models, both the
hardware and MAC protocol models
, can be used

in the future
in

evaluat
ing

new
generatio
n motes and MAC protocols.


U
nderstanding
how the
mote
circuitry
works and
learning
various techniques used in developing
MAC protocols

helped us in developing

a Radio Power Management (RPM) algorithm
that can be used to

increase WSN energy savings

and lif
etimes
. The RPM algorithm is hardware
specific;

therefore
,

it is





1

We

wou
ld like to thank Mr. Wayne Donald and Mr. Randy Marchany for supporting
our

research in the
Virginia Tech

Information Technology and Security Lab.
We

would also like to thank Mrs. Sally
Brownfield for her invaluable help in reviewing
our

work.



Figure 1
.

Sample Mote Characterization
. The figure shows the
Telos A

mote transitions associated
with
receive (RX), power off,
low power
, i
dle, and receive, in addition to the current consumption
levels at each state.


Using OPNET, we have coded various WSN specific MAC protocols,
which dictate network
packet transmittions
, in order to evaluate various WSN scenarios. The goal of the model was to use
obta
ined hardware characterization
s

to develop an energy model in OPNET
. This model provides
evaluations of
the efficiency of various MAC protocols and hardware technology.
By factoring in available
energy for each mote,
usually 2 AA batteries,
the simulation

is able to predict the lifetime, in days, until the
mote batteries are depleted.

The simulations are composed of 20 motes; in each simulation motes’ utilize
various MAC protocols under
a range of

traffic conditions. The simulations help evaluat
e two thi
ngs:

how
energy efficient the motes are and how efficient the various WSN MAC protocols are. The developed
models, both hardware and MAC protocol, can be used in
future evaluations of

new generation motes and
MAC protocols.


Understanding how the mote cir
cuitry works and learning various techniques used in developing
MAC protocols helped us in developing a Radio Power Management (RPM) algorithm that can be used to
increase WSN energy savings and lifetimes. The RPM algorithm is hardware specific; therefore
, it is
implemented differently on
various

motes. The crux of energy savings in WSN’s is for motes to utilize as
many
low power

opportunities as possible during their lifetime. The more time motes spend powered
-
down the longer their batteries will last.

The most important consideration for motes when they power
-
down is whether or not they have enough time to power
-
down and power back up without losing network
synchronization. Therefore, motes will only power
-
down if

they have enough time to do so;

if th
ey
determine that there is not enough time to power
-
down then they

will remain powered
-
up. T
he RPM
algorithm proposes various intermediate
low power

levels for motes to utilize when they do not have
enough time to power all the way down. Using the WSN mo
del developed in OPNET, it was determined
that the RPM algorithm provides invaluable intermediate
low power

opportunities that were not previously

available in WSN
s. Figure 2 is a flow chart representation of the RPM algorithm; it illustrates the basic
lo
gic motes follow in order to utilize
low power

opportunities optimally. By allowing motes more sleep
opportunities,

the
motes are able to
better
conserve their batteries
, which leads to an operational lifetime
extension.






















Figure 2
.

RPM Algorithm
.

The
figure shows the RPM
algorithm
and how it is
used
in

determin
ing

optimal
low
power

level
s
. Note:

NAV stands for Network Allocation Vector
,

which

determines how long the network medium will
be busy with a transmission. In

the RPM algorithm, NAV is used in determining how long a
low power

opportunity
will last.