PERFORMANCE MEASUREMENTS OF WIRELESS SENSOR NETWORKS

wizzstuffingΠολεοδομικά Έργα

16 Νοε 2013 (πριν από 3 χρόνια και 4 μήνες)

46 εμφανίσεις

PERFORMANCE
MEASUREMENTS OF
WIRELESS SENSOR NETWORKS



Gizem

ERDOĞAN

WIRELESS SENSOR NETWORKS


Wireless


Battery powered


Ad
-
hoc


Sense & monitor : temperature, humidity, light
intensity, voltage, current, volume, acceleration,
sound, pressure ,etc.


Various application fields: military, health care,
traffic control, scientific monitoring


Important aspects: Energy efficiency, self
configuration

S
TUDIES

OF

ANASTASI

ET

AL
.


Berkeley family motes


(a) mica2


(b) mica2dot


4
-
Mhz, 8
-
bit Atmel microprocessor


512 KB of non
-
volatile flash memory


32
-
KHz clock


RFM
ChipCon

Radio bit rate of 19.2 Kbps


CSMA/CA


TinyOS

E
XPERIMENTAL

ENVIRONMENT


Different traffic conditions


Outdoor environment


Temperature


Humidity


Fog


Rain


10 replicas in different times


Average values as well as upper &lower bounds


Virtual ground


Limits reflection and bad electromagnetic wave’s
perturbation


P
ARAMETERS

VALUES

DEFINITIONS


Transmission Range
(
TX_range
): the range
within which a transmitted frame can be
successfully received



the transmission power


the radio propagation properties



Carrier Sensing Range
(
CS_range
) :the range
within which the other sensor nodes can detect a
transmission


sensitivity of the receiver


the radio propagation properties




EXPERIMENTAL RESULTS

AVAILABLE

BANDWIDTH


Maximum size message 56 bytes


18
-
byte preamble


2
-
byte synchronization +


36 bytes data



Theoretical throughput







EXPERIMENTAL RESULTS

AVAILABLE

BANDWIDTH

CONT
.


m :

the amount of data to be transmitted.



For maximum size frame 36 bytes


Tframe

:
time required to transmit a MAC data
frame at 19.2 Kbps.


56*8 /19.2 * 10
-
3
= 23.33 ms;


(
B
min

+
IB
max
)/2 : the average backoff time


( 15 +68.3)/2=41.65 ms


Expected throughput : 4.43 Kbps


Estimated throughput: 4.4 Kbps





EXPERIMENTAL RESULTS

POWER

CONSUMPTION


EXPERIMENTAL RESULTS

POWER

CONSUMPTION

CONT
.

EXPERIMENTAL RESULTS

POWER

CONSUMPTION

CONT
.


Real World Application


Mica2 motes


Sample the light in every 1 second


Transmit 8 byte message to another node


When no messages to be sent, the radio turns off and
the motes power down.


Leaked current while sampling is 20mA


Leaked current while transmitting 18mA.


Leaked current when powered down 10uA.


Average current leaked in a cycle 0.19
mA
,


Average power consumption 0.19*3=0.57mW


Lifetime of the network: more than a year!


EXPERIMENTAL RESULTS

TRANSMISSION

RANGE


Two sensor nodes with the antennas in a back to
back disposition


Sequence numbers in each packet transmitted


Vary the distance between the nodes, keep the
track of the number of lost packets


Assume threshold as the distance at which the
percentage of received packets are below 85%



Transmission range is approximately 55 m for
mica2 and 135 m for mica2dot.

EXPERIMENTAL RESULTS

TRANSMISSION

RANGE

CONT
.

EXPERIMENTAL RESULTS

TRANSMISSION

RANGE

CONT
.


Factors that may affect the transmission range


Transmission power


Orientation of the antenna


Data rate


Sensor nodes location with respect to the ground


Environmental conditions


Transmission power: more than linear increase
this increase for both kinds of motes.



At the maximum transmission power 5dBm


transmission range: 70 m for mica2


Transmission range: 230 m for mica2dot


EXPERIMENTAL RESULTS

TRANSMISSION

RANGE

CONT
.


Influence of the antenna


Change the relative antennas’ disposition to see
the effect of the communication quality in terms
of received packets


mica2 antennas are very directional


mica2dot nodes are resistant

EXPERIMENTAL RESULTS

TRANSMISSION

RANGE

CONT
.


Influence of the data rate


Inversely proportional in IEEE 802.11 wireless
networks


Does not have a significant influence on in mica2
and mica2dot


Different scale of data rates


Motes: Kbps whereas



IEEE802.11 stations: Mbps.


EXPERIMENTAL RESULTS

TRANSMISSION

RANGE

CONT
.

EXPERIMENTAL RESULTS

TRANSMISSION

RANGE

CONT
.


Effect of sensor node’s height


When the nodes are close to the ground under 1
meter, significant power loss.


This loss is due to the interference between the ground.

EXPERIMENTAL RESULTS

TRANSMISSION

RANGE

CONT
.


Influence of environmental conditions


Slight variations of temperature or humidity do
not have any significant influence


In the presence of fog or rain, we saw that
transmission range decreased significantly


Due to signal attenuation caused by the interference of fog
and rain particles with the electromagnetic waves


EXPERIMENTAL RESULTS

CARRIER

SENSING

RANGE


Fixed distance between the nodes in a couple


Vary the distance between the two couples


Until no correlation is measured between the couples



Until Throughput achieved =4.4 Kbps




EXPERIMENTAL RESULTS

CARRIER

SENSING

RANGE


275 m :end of the carrier sensing range


Minor interference


450 m: interference becomes negligible


MAC PROTOCOLS


S
-
MAC


Reduce energy consumption caused by idle listening


Schedule coordinated transmission and listen periods


Overhead due to coordination and schedule maintenance.


B
-
MAC


Wake up for a very small time and sleep for a longer time.


Poll the channel, if nothing interesting, go back to sleep


Long preamble guaranteed to intersect with polling


SCP
-
MAC


Ultra
-
low duty cycle


Synchronizing the polling times

DATA DISSEMINATION


PUSH BASED STRATEGY


Nodes detecting the interesting event broadcast the
relevant information


Efficient when there is constant need of information


Broadcast bandwidth is wasted when the demand for the
information is low


PULL BASED STRATEGY


Querier broadcasts a query for the information when it is
needed.


Nodes that have the relevant information send the
information back.


Communication takes place only when it is needed.

DATA DISSEMINATION
CONT
.


COMB
-
NEEDLE STRATEGY


Integrates both push and pull based techniques


Each sensor node pushes its data through some
number of


The querier pulls the data in a certain
neighborhood


In most cases it is more efficient than both pure
push and pure pull strategies.


A
NY

Q
UESTIONS
?


THANK YOU!