Lecture3.1Factors Influencing Sensor Network

7/29/2019 Lecture3.1Factors Influencing Sensor Network

1/75

1

Factors Influencing Sensor NetworkFactors Influencing Sensor Network

DesignDesign


2/75

2

Factors Influencing Sensor NetworkFactors Influencing Sensor Network

DesignDesign

A. Hardware ConstraintsA. Hardware Constraints

B. Fault Tolerance (Reliability)B. Fault Tolerance (Reliability)

C. ScalabilityC. Scalability

D. Production CostsD. Production CostsE. Sensor Network TopologyE. Sensor Network Topology

F. Operating Environment (Applications)F. Operating Environment (Applications)

G. Transmission MediaG. Transmission Media

H. Power Consumption (Lifetime)H. Power Consumption (Lifetime)


3/75

3

Sensor Node HardwareSensor Node Hardware

Power UnitPower Unit AntennaAntenna

Sensor ADCSensor ADC

ProcessorProcessor

MemoryMemory

TransceiverTransceiver

Location Finding SystemLocation Finding System MobilizerMobilizer

SENSING UNIT PROCESSING UNIT


4/75

4

Fault ToleranceFault Tolerance

(Reliability)(Reliability)

Sensor nodes may fail due to lack of power,Sensor nodes may fail due to lack of power,physical damage or environmental interferencephysical damage or environmental interference

The failure of sensor nodes should not affect theThe failure of sensor nodes should not affect the

overall operation of the sensor networkoverall operation of the sensor network

This is calledThis is calledRELIABILITY or FAULT TOLERANCE,RELIABILITY or FAULT TOLERANCE,

i.e., ability to sustain sensor network functionalityi.e., ability to sustain sensor network functionality

without any interruptionwithout any interruption


5/75

5

Fault Tolerance (Reliability)Fault Tolerance (Reliability)

Reliability R (Fault Tolerance) of a sensor node k isReliability R (Fault Tolerance) of a sensor node k ismodeled:modeled:

i.e., by Poisson distribution, to capture the probability ofi.e., by Poisson distribution, to capture the probability ofnot having a failure within the time interval (0,t) with lnot having a failure within the time interval (0,t) with l kk isisthe failure rate of the sensor node k and t is the time period.the failure rate of the sensor node k and t is the time period.

)()(

t

kketR

=

G. Hoblos, M. Staroswiecki, and A. Aitouche,G. Hoblos, M. Staroswiecki, and A. Aitouche, Optimal Design of Fault Tolerant SensorOptimal Design of Fault Tolerant SensorNetworks,Networks, IEEE Int. Conf. on Control ApplicationsIEEE Int. Conf. on Control Applications, pp. 467-472, Sept. 2000., pp. 467-472, Sept. 2000.


6/75

6


Reliability (Fault Tolerance) of a broadcast rangeReliability (Fault Tolerance) of a broadcast rangewith N sensor nodes is calculated fromwith N sensor nodes is calculated from

])(1[1)(1

=

=N

k

k tRtR


7/757


EXAMPLE:EXAMPLE:

How many sensor nodes are needed within aHow many sensor nodes are needed within abroadcast radius (range) to have 99% fault toleratedbroadcast radius (range) to have 99% fault tolerated

network?network?

Assuming all sensors within the radio range haveAssuming all sensors within the radio range havesame reliability, previous equation becomes:same reliability, previous equation becomes:

Drop t and substitute f = (1-R) 0.99 = (1 fN) N=2

NtRtR )](1[1)( =


8/758


REMARKREMARK::

1. Protocols and algorithms may be designed to1. Protocols and algorithms may be designed toaddress the level of fault tolerance required byaddress the level of fault tolerance required by

sensor networks.sensor networks.

2. If the environment has little interference, then2. If the environment has little interference, thenthe requirements can be more relaxed.the requirements can be more relaxed.


9/759


Examples:Examples:1.1. HouseHouse to keep track of humidity and temperatureto keep track of humidity and temperature

levelslevels the sensors cannot be damaged easily orthe sensors cannot be damaged easily orinterfered by environmentinterfered by environment lowlow fault tolerancefault tolerance(reliability) requirement!!!!(reliability) requirement!!!!

2.2. BattlefieldBattlefield for surveillance the sensed data are criticalfor surveillance the sensed data are criticaland sensors can be destroyed by enemiesand sensors can be destroyed by enemies highhigh faultfault

tolerance (reliability) requirement!!!tolerance (reliability) requirement!!!

Bottom line:Bottom line: Fault Tolerance (Reliability)Fault Tolerance (Reliability)

depends heavily on applications!!!depends heavily on applications!!!


10/7510

ScalabilityScalability

The number of sensor nodes may reach thousandsThe number of sensor nodes may reach thousandsin some applicationsin some applications

The density of sensor nodes can range from few toThe density of sensor nodes can range from few toseveral hundreds in a region (cluster) which can beseveral hundreds in a region (cluster) which can be

less than 10m in diameterless than 10m in diameter


11/7511

Node DensityNode Density:: The number of expected nodes per unit areaThe number of expected nodes per unit area::

N is the number of scattered sensor nodes in region AN is the number of scattered sensor nodes in region A

Node DegreeNode Degree: The number of expected nodes in the transmission range of a: The number of expected nodes in the transmission range of anodenode

R is the radio transmission rangeR is the radio transmission range

Basically:Basically: mm(R(R)) is the number of sensor nodes within the transmissionis the number of sensor nodes within the transmissionradius R of each sensor node in region A.radius R of each sensor node in region A.


AN/=

2)( RR =


12/75

12


EXAMPLE:EXAMPLE:

Assume sensor nodes are evenly distributed in the sensorAssume sensor nodes are evenly distributed in the sensor

field. Determine the node density and node degree if 200 sensorfield. Determine the node density and node degree if 200 sensornodes are deployed in a 50x50 mnodes are deployed in a 50x50 m22 region where each sensorregion where each sensor

node has a broadcast radius of 5m.node has a broadcast radius of 5m.

Use the eq.Use the eq.

6508.0)( 2 = R

08.0)5050/(200 ==


13/75

13

ScalabilityScalabilityExamples:Examples:

1.1. Machine Diagnosis Application:Machine Diagnosis Application:less than 50 sensor nodes in a 5 m x 5 m region.less than 50 sensor nodes in a 5 m x 5 m region.

2.2. Vehicle Tracking Application:Vehicle Tracking Application:Around 10 sensor nodes per cluster/region.Around 10 sensor nodes per cluster/region.

3.3. Home Application:Home Application: tens depending on the size of the house.tens depending on the size of the house.

4.4. Habitat Monitoring Application:Habitat Monitoring Application:Range from 25 to 100 nodes/clusterRange from 25 to 100 nodes/cluster

5.5. Personal Applications:Personal Applications:Ranges from tens to hundreds, e.g., clothing, eye glasses, shoes,Ranges from tens to hundreds, e.g., clothing, eye glasses, shoes,watch, jewelry.watch, jewelry.


14/75

14

Production CostsProduction Costs

Cost of sensors must be low so that sensorCost of sensors must be low so that sensornetworks can be justified!networks can be justified!

PicoNode: less than $1PicoNode: less than $1

Bluetooth system: around $10,-Bluetooth system: around $10,- THE OBJECTIVE FOR SENSOR COSTSTHE OBJECTIVE FOR SENSOR COSTS

must be lower than $1!!!!!!!must be lower than $1!!!!!!!

CurrentlyCurrently ranges from $25 to $180ranges from $25 to $180(STILL VERY EXPENSIVE!!!!)(STILL VERY EXPENSIVE!!!!)


15/75

15

Sensor Network TopologySensor Network Topology

Internet,Internet,Satellite, UAVSatellite, UAV

Sink

Sink

Task

Manager


16/75

16


Topology maintenance and change:Topology maintenance and change:

Pre-deployment and Deployment PhasePre-deployment and Deployment Phase

Post Deployment PhasePost Deployment Phase Re-Deployment of Additional NodesRe-Deployment of Additional Nodes


17/75

17


Pre-deployment and Deployment PhasePre-deployment and Deployment Phase

Dropped from aircraft(Random deployment) Well Planned, Fixed (Regular deployment) Mobile Sensor Nodes

Adaptive, dynamicCan move to compensate for deployment

shortcomings

Can be passively moved around by some

external force (wind, water)Can actively seek out interesting areas


18/75

18


Initial Deployment SchemesInitial Deployment Schemes

Reduce installation costReduce installation cost

Eliminate the need for any pre-organization andEliminate the need for any pre-organization andpre-planningpre-planning

Increase the flexibility of arrangementIncrease the flexibility of arrangement

Promote self-organization and fault-tolerancePromote self-organization and fault-tolerance


19/75

19


POST-DEPLOYMENT PHASEPOST-DEPLOYMENT PHASE

Topology changes may occur:Topology changes may occur:

PositionPosition

Reachability (due to jamming, noise, movingReachability (due to jamming, noise, moving

obstacles, etc.)obstacles, etc.)Available energyAvailable energy

MalfunctioningMalfunctioning


20/75

20

Operating EnvironmentOperating Environment

* SEE ALL THE APPLICATIONS discussed before* SEE ALL THE APPLICATIONS discussed before


21/75

21

TRANSMISSION MEDIATRANSMISSION MEDIA

Radio, Infrared, Optical, Acoustic, Magnetic MediaRadio, Infrared, Optical, Acoustic, Magnetic Media

ISMISM (Industrial, Scientific and Medical)(Industrial, Scientific and Medical) Bands (433Bands (433

MHz ISM Band in Europe and 915 MHz as well asMHz ISM Band in Europe and 915 MHz as well as2.4 GHz ISM Bands in North America)2.4 GHz ISM Bands in North America)

REASONS:REASONS: Free radio, huge spectrum allocationFree radio, huge spectrum allocation

and global availability.and global availability.


22/75

22

POWER CONSUMPTIONPOWER CONSUMPTION

Sensor node has limited power sourceSensor node has limited power source

Sensor node LIFETIME depends on BATTERY lifetimeSensor node LIFETIME depends on BATTERY lifetime

Goal: Provide as much energy as possible at smallestGoal: Provide as much energy as possible at smallestcost/volume/weight/rechargecost/volume/weight/recharge

Recharging may or may not be an optionRecharging may or may not be an option

OptionsOptions

Primary batteries not rechargeablePrimary batteries not rechargeable

Secondary batteries rechargeable, only makesSecondary batteries rechargeable, only makes

sense in combination with some form of energysense in combination with some form of energyharvestingharvesting


23/75

23

Battery ExamplesBattery Examples

Energy per volume (Joule per cubic centimeter):Energy per volume (Joule per cubic centimeter):Primary batteries

Chemistry Zinc-air Lithium Alkaline

Energy (J/cm3) 3780 2880 1200

Secondary batteries

Chemistry Lithium NiMHd NiCd

Energy (J/cm3) 1080 860 650


24/75

24

Energy ScavengingEnergy Scavenging (Harvesting)(Harvesting)Ambient Energy Sources (their power density)Ambient Energy Sources (their power density)

Solar (Outdoors)Solar (Outdoors) 15 mW/cm 15 mW/cm22 (direct sun)(direct sun)

Solar (Indoors)Solar (Indoors) 0.006 mW/cm 0.006 mW/cm22 (office desk)(office desk)

0.57 mW/cm0.57 mW/cm22

(


25/75

25


Sensors can be aSensors can be a DATA ORIGINATORDATA ORIGINATOR or aor a DATADATAROUTER.ROUTER.

Power conservation and power management arePower conservation and power management areimportantimportant

POWER AWARE COMMUNICATION PROTOCOLSPOWER AWARE COMMUNICATION PROTOCOLSmust be developed.must be developed.


26/75

26



27/75

27

Power ConsumptionPower Consumption

Power consumption in a sensor network can bePower consumption in a sensor network can bedivided into three domainsdivided into three domains

SensingSensingData Processing (Computation)Data Processing (Computation)

CommunicationCommunication


28/75

28






29/75

29


SensingSensing

Depends onDepends on

ApplicationApplication

Nature of sensing: Sporadic or ConstantNature of sensing: Sporadic or Constant

Detection complexityDetection complexity

Ambient noise levelsAmbient noise levels

Rule of thumb (ADC power consumption)Rule of thumb (ADC power consumption)

FFss - sensing frequency, ENOB - effective number of bits- sensing frequency, ENOB - effective number of bits

PsFS2

EN


30/75

30





Power Consumption inPower Consumption in


31/75

31


Data Processing (Computation)Data Processing (Computation)(Wang/Chandrakarasan: Energy Efficient DSPs for Wireless Sensor(Wang/Chandrakarasan: Energy Efficient DSPs for Wireless Sensor

Networks. IEEE Signal Proc. Magazine, July 2002. also from Shih paper)Networks. IEEE Signal Proc. Magazine, July 2002. also from Shih paper)

)(***/2 TVndd

dd

V

OddP eIVVCfP +=

The power consumption in data processing (PThe power consumption in data processing (Ppp) is) is

f clock frequency

C is the aver. capacitance switched per cycle (C ~ 0.67nF);

Vdd is the supply voltage

VT is the thermal voltage (n~21.26; Io ~ 1.196 mA)


32/75

32


Data ProcessingData Processing (Computation)(Computation)

The second term indicates the power loss due toThe second term indicates the power loss due toleakage currentsleakage currents

In general, leakage energy accounts for about 10%In general, leakage energy accounts for about 10%of the total energy dissipationof the total energy dissipation

In low duty cycles, leakage energy can becomeIn low duty cycles, leakage energy can become

large (up to 50%)large (up to 50%)


33/75

33


Data ProcessingData Processing

This is much less than in communication.This is much less than in communication.

EXAMPLE:EXAMPLE: (Assuming: Rayleigh Fading wireless(Assuming: Rayleigh Fading wireless

channel; fourth power distance loss)channel; fourth power distance loss)

Energy cost of transmittingEnergy cost of transmitting 1 KB1 KB over a distance ofover a distance of100 m is approx. equal to executing100 m is approx. equal to executing 0.25 Million0.25 Millioninstructionsinstructions by a 8 million instructions per secondby a 8 million instructions per secondprocessor (MicaZ).processor (MicaZ).

Local data processing is crucial in minimizingLocal data processing is crucial in minimizingpower consumption in a multi-hop networkpower consumption in a multi-hop network


34/75

34

Memory Power ConsumptionMemory Power Consumption

Crucial part: FLASH memoryCrucial part: FLASH memory

Power for RAM almost negligiblePower for RAM almost negligible

FLASH writing/erasing is expensiveFLASH writing/erasing is expensiveExample: FLASH on Mica motesExample: FLASH on Mica motes

Reading: 1.1 nAh per byteReading: 1.1 nAh per byte

Writing: 83.3 nAh per byteWriting: 83.3 nAh per byte


35/75

35






36/75

36

Power Consumption forPower Consumption for


A sensor spends maximum energy in dataA sensor spends maximum energy in datacommunication (both for transmission and reception).communication (both for transmission and reception).

NOTE:NOTE:

For short range communication with low radiationFor short range communication with low radiationpower (~0 dbm), transmission and reception powerpower (~0 dbm), transmission and reception powercosts are approximately the same,costs are approximately the same,

e.g., modern low power short range transceiverse.g., modern low power short range transceiversconsume betweenconsume between 15 and 300 mW15 and 300 mW of power whenof power when

sending and receivingsending and receiving Transceiver circuitry has both active and start-upTransceiver circuitry has both active and start-up

power consumptionpower consumption


37/75

37



Power consumption forPower consumption fordata communicationdata communication (P(Pcc))

PPcc = P= P00 + P+ Ptxtx + P+ Prxrx

PPte/rete/re is the power consumed in the transmitter/receiveris the power consumed in the transmitter/receiver

electronics (including the start-up power)electronics (including the start-up power)

PP00 is the output transmit poweris the output transmit power

TX RXTX RX


38/75

38



START-UP POWER/ START-UP TIMESTART-UP POWER/ START-UP TIME

A transceiver spends upon waking up from sleep mode,A transceiver spends upon waking up from sleep mode,e.g., to ramp upe.g., to ramp up phase locked loops or voltagephase locked loops or voltagecontrolled oscillatorscontrolled oscillators..

During start-up time, no transmission or reception ofDuring start-up time, no transmission or reception ofdata is possible.data is possible.

Sensors communicate in short data packetsSensors communicate in short data packets

Start-up power starts dominating as packet size isStart-up power starts dominating as packet size isreducedreduced

It is inefficient to turn the transceiver ON and OFFIt is inefficient to turn the transceiver ON and OFFbecause a large amount of power is spent in turning thebecause a large amount of power is spent in turning thetransceiver back ON each time.transceiver back ON each time.


39/75

39

Wasted EnergyWasted Energy

Fixed cost of communication:Fixed cost of communication: Startup TimeStartup Time High energy per bit for small packetsHigh energy per bit for small packets (from Shih paper)(from Shih paper)

Parameters: R=1 Mbps; TParameters: R=1 Mbps; Tstst ~ 450 msec, P~ 450 msec, Ptete~81mW; P~81mW; Poutout = 0 dBm= 0 dBm


40/75

40

Energy vs Packet SizeEnergy vs Packet Size

TR 1000 (115kbps)

0

1020

30

40

50

60

10 100 1000 10000

Packet Size (bits)

E

bit(p

J)

Energy per Bit

(pJ)

As packet size is reduced the energy consumption is dominated by the startup time on the orderof hundreds of microseconds during which large amounts of power is wasted.

NOTE: During start-up time NO DATA CAN BE SENT or RECEIVED by the

transceiver.


41/75

41

Start-Up and SwitchingStart-Up and Switching

Startup energy consumptionStartup energy consumption

EEstst = P= PLOLO x tx tstst

PPLOLO, power consumption of the circuitry (synthesizer, power consumption of the circuitry (synthesizer

and VCO); tand VCO); tstst, time required to start up all, time required to start up allcomponentscomponents

Energy is consumed when transceiver switchesEnergy is consumed when transceiver switchesfrom transmit to receive modefrom transmit to receive mode

Switching energy consumptionSwitching energy consumptionEEswsw = P= PLOLO x tx tswsw


42/75

42

Start-Up Time and Sleep ModeStart-Up Time and Sleep Mode

The effect of the transceiver startup time willThe effect of the transceiver startup time willgreatly depend on the type of MAC protocol used.greatly depend on the type of MAC protocol used.

To minimize power consumption, it is desirable toTo minimize power consumption, it is desirable tohave the transceiver in ahave the transceiver in a sleep modesleep mode as much asas much aspossiblepossible

Energy savings up to 99.99% (59.1mWEnergy savings up to 99.99% (59.1mW 33mmW)W) BUTBUT

Constantly turning on and off the transceiver alsoConstantly turning on and off the transceiver alsoconsumes energy to bring it to readiness forconsumes energy to bring it to readiness fortransmission or reception.transmission or reception.


43/75

43

Receiving and Transmitting EnergyReceiving and Transmitting Energy

ConsumptionConsumption

Receiving energy consumptionReceiving energy consumption

EErxrx = (P= (PLOLO + P+ PRXRX ) t) trxrx

PPRXRX, power consumption of active components, e.g.,, power consumption of active components, e.g.,

decoder, tdecoder, trxrx, time it takes to receive a packet, time it takes to receive a packet Transmitting energy consumptionTransmitting energy consumption

EEtxtx = (P= (PLOLO + P+ PPAPA ) t) ttxtx

PPPAPA, power consumption of power amplifier, power consumption of power amplifier

PPPAPA = 1/= 1/ PPoutout ,, power efficiency of power amplifier, Ppower efficiency of power amplifier, Poutout, desired, desired

RF output power levelRF output power level


44/75

44

RF output powerRF output power

http://memsic.com/support/documentation/wireless-sensor-networks/category/7-datasheets.html?download=148%3Amicazhttp://memsic.com/support/documentation/wireless-sensor-networks/category/7-datasheets.html?download=148%3Amicaz


45/75

45

Power Amplifier Power ConsumptionPower Amplifier Power Consumption

Receiving energy consumptionReceiving energy consumption

PPPAPA = 1/= 1/ PAPA rr ddnn

PAPA, amplifier constant (antenna gain, wavelength,, amplifier constant (antenna gain, wavelength,thermal noise power spectral density, desiredthermal noise power spectral density, desiredsignal to noise ratio (SNR) at distance d),signal to noise ratio (SNR) at distance d),

r, data rate,r, data rate,

n, path loss exponent of the channel (n=2-4)n, path loss exponent of the channel (n=2-4)

d, distance between nodesd, distance between nodes


46/75

46

Lets put it togetherLets put it together

Energy consumption for communicationEnergy consumption for communication

EEcc = E= Estst + E+ Erxrx + E+ Eswsw + E+ Etxtx

= P= PLOLO ttstst + (P+ (PLOLO + P+ PRXRX)t)trxrx + P+ PLOLO ttswsw + (P+ (PLOLO+P+PPAPA)t)ttxtx

Let tLet trxrx = t= ttxtx = l= lPKTPKT/r/r

EEcc = P= PLOLO (t(tstst+t+tswsw)+(2P)+(2PLOLO + P+ PRXRX)l)lPKTPKT/r + 1//r + 1/ PAPA llPKTPKT ddnnDistance-independentDistance-independent Distance-dependentDistance-dependent


47/75

47

A SIMPLE ENERGY MODELA SIMPLE ENERGY MODEL

Operation EnergyDissipated

Transmitter Electronics ( ETx-elec)

Receiver Electronics ( ERx-elec)

( ETx-elec = ERx-elec = Eelec )

50 nJ/bit

Transmit Amplifier {eamp} 100pJ/bit/m2

TransmitElectronics

TxAmplifier

ETx (k,D)

Eelec * keamp* k* D

2

k bit packet

ReceiveElectronics

Eelec * k

k bit packet

D

tx (k,D) = Etx-elec (k) + Etx-amp (k,D)

tx (k,D) = Eelec * k + eamp * k * D2

ERx (k) = Erx-elec (k)

ERx (k) = Eelec * k

ERx (k)

ETx-elec (k) ETx-amp (k,D)


48/75

48

Power ConsumptionPower Consumption(A Simple Energy Model)(A Simple Energy Model)

Assuming a sensor node is only operating in transmit andreceive modes with the following assumptions:Energy to run circuitry:

Eelec = 50 nJ/bit

Energy for radio transmission:

eamp = 100 pJ/bit/m2

Energy for sending k bits over distance D

ETx

(k,D) = Eelec

* k + eamp

* k * D2

Energy for receiving k bits:

ERx (k,D) = Eelec * k


49/75

49

Example using the Simple Energy ModelExample using the Simple Energy Model

What is the energy consumption if 1 Mbit ofinformation is transferred from the source to the sinkwhere the source and sink are separated by 100meters and the broadcast radius of each node is 5meters?

Assume the neighbor nodes are overhearing each

others broadcast.


50/75

50

EXAMPLEEXAMPLE

100 meters / 5 meters = 20 pairs of transmitting andreceiving nodes (one node transmits and one node receives)

ETx (k,D) = Eelec * k + eamp * k * D2

ETx = 50 nJ/bit . 106 + 100 pJ/bit/m2 . 106 . 52 =

= 0.05J + 0.0025 J = 0.0525 J

ERx (k,D) = Eelec * k

ERx = 0.05 J

Epair = ETx + ERx = 0.1025J

ET = 20 . Epair = 20. 0.1025J = 2.050 J


51/75

51

VERY DETAILED ENERGY MODEL

sleepsleepononTPTPE +=

Simple Energy Consumption Model

A More Realistic ENERGY MODEL*

( ) LTPTPBTGP

BT

LNE trsynoncond

b

on

BT

L

BT

L

f

on

on /2

214

ln123

41

2

2

++

+=

* S. Cui, et.al., Energy-Constrained Modulation* S. Cui, et.al., Energy-Constrained ModulationOptimization,Optimization, IEEE Trans. on Wireless CommunicationsIEEE Trans. on Wireless Communications,,September 2005.September 2005.

fD t il f th R li ti M d l


52/75

52

Details of the Realistic ModelDetails of the Realistic Model

onTB

L

M

MM

=

+

=

=

2

113

1

L packet lengthL packet length

B channel bandwidthB channel bandwidth

NNff receiver noise figure receiver noise figure

22 power spectrum energy power spectrum energyPPbb probability of bit error probability of bit error

GGdd power gain factor power gain factor

PPcc circuit power consumption circuit power consumption

PPsynsyn frequency synthesizer power frequency synthesizer power

consumptionconsumption

TTtrtr frequency synthesizer settling time (duration of frequency synthesizer settling time (duration of

transient mode)transient mode)TTonon transceiver on time transceiver on time

M Modulation parameterM Modulation parameter

ANOTHER EXAMPLE


53/75

53

Enery Consumption: Important Variables:Enery Consumption: Important Variables:

PPrere 4.5 mA4.5 mA (energy consumption at receiver)(energy consumption at receiver)PPtete 12.0 mA12.0 mA (energy consumption at transmitter)(energy consumption at transmitter)PPclcl 12.0 mA12.0 mA(basic consumption without radio)(basic consumption without radio)PPslsl 8mA (0.008 mA)8mA (0.008 mA) (energy needed to sleep)(energy needed to sleep)

ANOTHER EXAMPLE

EXAMPLE


54/75

54

Capacity (Watt) = Current (Ampere) * Voltage (Volt)Capacity (Watt) = Current (Ampere) * Voltage (Volt)Rough estimation for energy consumption and sensor lifetime:Rough estimation for energy consumption and sensor lifetime:

Let us assume that each sensor should wake up once aLet us assume that each sensor should wake up once a

second, measure a value and transmit it over the network.second, measure a value and transmit it over the network.

a) Calculations needed: 5K instructions (for measurement anda) Calculations needed: 5K instructions (for measurement and

preparation for sending)preparation for sending)

b) Time to send information: 50 bytes for sensor data,b) Time to send information: 50 bytes for sensor data,

(another 250 byte for forwarding external data)(another 250 byte for forwarding external data)

c) Energy needed to sleep for the rest of the time (sleepc) Energy needed to sleep for the rest of the time (sleep

mode)mode)

EXAMPLE

EXAMPLE


55/75

55

Time for Calculations and Energy Consumption:Time for Calculations and Energy Consumption:

MSP430 running at 8 MHz clock rateMSP430 running at 8 MHz clock rate one cycleone cycletakes 1/(8*10takes 1/(8*1066) seconds) seconds

1 instruction needs an average of 3 cycles1 instruction needs an average of 3 cycles 3/3/(8* 10(8* 1066) sec, 5K instructions, 15/(8*10) sec, 5K instructions, 15/(8*1033) sec) sec

15/(8*1015/(8*1033) * 12mA = 180/8000 = 0.0225 mAs) * 12mA = 180/8000 = 0.0225 mAs

EXAMPLE


56/75

56

Time for Sending Data and Energy Consumption:Time for Sending Data and Energy Consumption:

Radio sends with 19.200 baud (approx. 19.200 bits/sec)Radio sends with 19.200 baud (approx. 19.200 bits/sec)

1 bit takes 1/19200 seconds1 bit takes 1/19200 seconds

We have to send 50 bytes (own measurement)We have to send 50 bytes (own measurement)

and we have to forward 250 bytes (externaland we have to forward 250 bytes (external

data): 250+50=300 which takesdata): 250+50=300 which takes

300*8/19200s*24mA (energy basic + sending) = 3mAs300*8/19200s*24mA (energy basic + sending) = 3mAs

EXAMPLE

EXAMPLE


57/75

57

Energy consumed while sleeping:Energy consumed while sleeping:

Time for calculation 15/8000 + time for transmissionTime for calculation 15/8000 + time for transmission

300*8/19200 ~ 0.127 sec300*8/19200 ~ 0.127 sec Time for sleep mode = 1 sec 0.127 = 0.873 sTime for sleep mode = 1 sec 0.127 = 0.873 s

Energy consumed while sleepingEnergy consumed while sleeping

0.008mA * 0.873 s = 0.0007 mAs0.008mA * 0.873 s = 0.0007 mAs

EXAMPLE

EXAMPLE


58/75

58

Total Amount of energy and resulting lifetimeTotal Amount of energy and resulting lifetime::

The ESB needs to be supplied with 4.5 V so we needThe ESB needs to be supplied with 4.5 V so we need

3 * 1.5V AA batteries.3 * 1.5V AA batteries.

3*(0.0225 + 3 + 0.007) ~ 3 * 3.03 mWs3*(0.0225 + 3 + 0.007) ~ 3 * 3.03 mWs

Energy of 3AA battery ~ 3 * 2300 mAh = 3*2300*60*60 mWsEnergy of 3AA battery ~ 3 * 2300 mAh = 3*2300*60*60 mWs

Total lifetimeTotal lifetime 3*2300*60*60/3*3.03 ~ 32 days.3*2300*60*60/3*3.03 ~ 32 days.

EXAMPLE

EXAMPLE


59/75

59

NOTES:NOTES:

Battery suffers from large current (losing about 10% energy/year)Battery suffers from large current (losing about 10% energy/year)

Small network (forwarding takes only 250 bytes)Small network (forwarding takes only 250 bytes)

Most important:Most important:

Only sending was taken into account, not receivingOnly sending was taken into account, not receiving

If we listen into the channel rather than sleeping 0.007 mA has to beIf we listen into the channel rather than sleeping 0.007 mA has to bereplaced by (12+4.5)mAreplaced by (12+4.5)mA

which results in a lifetime of ~ 5 days.which results in a lifetime of ~ 5 days.

EXAMPLE


60/75

60

Power Consumption forPower Consumption forCommunicationCommunication(Detailed Formula)(Detailed Formula)

)]([)]()([stonreRonOstonteTc

RRPNTPTTPNP ++++=

wherewherePPtete is power consumed by transmitteris power consumed by transmitter

PPrere is power consumed by receiveris power consumed by receiver

PPOO is output power of transmitteris output power of transmitter

TTonon is transmitter on timeis transmitter on time

RRonon is receiver on timeis receiver on timeTTstst is start-up time for transmitteris start-up time for transmitter

RRstst is start-up time for receiveris start-up time for receiver

NNTT is the number of times transmitteris the number of times transmitter

is switched on per unit of timeis switched on per unit of time

NNRRis the number of times receiveris the number of times receiver

is switched on per unit of timeis switched on per unit of time

E. Shih et al.,Physical Layer Driven Protocols and Algorithm Design forE. Shih et al.,Physical Layer Driven Protocols and Algorithm Design forEnergy-Efficient Wireless Sensor Networks, ACM MobiCom, Rome, JulyEnergy-Efficient Wireless Sensor Networks, ACM MobiCom, Rome, July2001.2001.

P C ti fP C ti f


61/75

61



TTonon = L / R= L / R

where L is the packet size in bits and R is thewhere L is the packet size in bits and R is thedata rate.data rate.

NNTT and Nand NRR depend on MAC and applications!!!depend on MAC and applications!!!

What can we do to Reduce Energy ConsumptionWhat can we do to Reduce Energy Consumption


62/75

62

What can we do to Reduce Energy ConsumptionWhat can we do to Reduce Energy Consumption

Multiple Power Consumption ModesMultiple Power Consumption Modes

Way out:Way out: Do not run sensor node at full operation all theDo not run sensor node at full operation all thetimetime

If nothing to do, switch toIf nothing to do, switch topower safe modepower safe mode

Question: When to throttle down? How to wake upQuestion: When to throttle down? How to wake up

again?again? Typical modesTypical modes

Controller: Active, idle, sleepController: Active, idle, sleep

Radio mode: Turn on/offRadio mode: Turn on/off

transmitter/receiver, bothtransmitter/receiver, both


63/75

63


Multiple modes possibleMultiple modes possible Deeper sleep modesDeeper sleep modes

Strongly depends on hardwareStrongly depends on hardwareTI MSP 430, e.g.: four different sleep modesTI MSP 430, e.g.: four different sleep modes

Atmel ATMega: six different modesAtmel ATMega: six different modes


64/75

64


MicrocontrollerMicrocontroller

TI MSP 430TI MSP 430

Fully operation 1.2 mWFully operation 1.2 mW

Deepest sleep mode 0.3Deepest sleep mode 0.3 W only woken up byW only woken up byexternal interrupts (not even timer is running anyexternal interrupts (not even timer is running anymore)more)

Atmel ATMegaAtmel ATMega

Operational mode: 15 mW active, 6 mW idleOperational mode: 15 mW active, 6 mW idle

Sleep mode: 75Sleep mode: 75 WW

S it hi b t M d


65/75

65

Switching between ModesSwitching between Modes

Simplest idea: Greedily switch to lower mode wheneverSimplest idea: Greedily switch to lower mode wheneverpossiblepossible

Problem: Time and power consumption required to reachProblem: Time and power consumption required to reachhigher modes not negligiblehigher modes not negligible

Introduces overheadIntroduces overhead

Switching only pays off if ESwitching only pays off if Esavedsaved > E> Eoverheadoverhead

S i hi b M dS it hi b t M d


66/75

66

Switching between ModesSwitching between Modes

Example: Event-triggered wake up from sleep modeExample: Event-triggered wake up from sleep mode

Scheduling problem with uncertaintyScheduling problem with uncertainty

Pactive

Psleeptimeteventt1

Esaved

tdown tup

Eoverhead


67/75

67

Alternative: Dynamic Voltage ScalingAlternative: Dynamic Voltage Scaling

Switching modes complicated by uncertainty onSwitching modes complicated by uncertainty onhow long a sleep time is availablehow long a sleep time is available

Alternative: Low supply voltage & clockAlternative: Low supply voltage & clock

Dynamic Voltage Scaling (DVS)Dynamic Voltage Scaling (DVS)

A controller running at a lower speed, i.e., lowerA controller running at a lower speed, i.e., lowerclock rates, consumes less powerclock rates, consumes less power

Reason: Supply voltage can be reduced at lowerReason: Supply voltage can be reduced at lowerclock rates while still guaranteeing correctclock rates while still guaranteeing correct

operationoperation


68/75

68


Reducing the voltage is a very efficient way toReducing the voltage is a very efficient way toreduce power consumption.reduce power consumption.

Actual power consumption P depends quadraticallyActual power consumption P depends quadraticallyon the supply voltage Von the supply voltage VDDDD, thus,, thus,

P ~ VP ~ VDDDD22

Reduce supply voltage to decrease energyReduce supply voltage to decrease energyconsumption !consumption !


69/75

69

69


Gate delay also depends on supply voltageGate delay also depends on supply voltage

K and a are processor dependent (a ~ 2)K and a are processor dependent (a ~ 2)

Gate switch periodGate switch period TT00=1/f=1/f

For efficient operationFor efficient operation

TTgg


70/75

70


)(~)(

cVKVdd

VVKf dd

a

thdd


71/75

71

Computation vs. Communication EnergyComputation vs. Communication Energy

costcost

Tradeoff?Tradeoff?

Directly comparing computation/communicationDirectly comparing computation/communicationenergy cost not possibleenergy cost not possible

But: put them into perspective!But: put them into perspective!

Energy ratio of sending one bit vs. computingEnergy ratio of sending one bit vs. computingone instruction: Anything between 220 and 2900one instruction: Anything between 220 and 2900

in the literaturein the literature

To communicate (send & receive)To communicate (send & receive) one kilobyteone kilobyte ==

computingcomputing three million instructions!three million instructions!

C t ti C i ti EComputation vs Communication Energy


72/75

72

Computation vs. Communication EnergyComputation vs. Communication Energy

CostCost

BOTTOMLINEBOTTOMLINE

Try to compute instead of communicateTry to compute instead of communicatewhenever possiblewhenever possible

Key technique in WSN Key technique in WSN in-network processingin-network processing!!

Exploit compression schemes, intelligent codingExploit compression schemes, intelligent codingschemes, aggregation, filtering, schemes, aggregation, filtering,

BOTTOMLINE


73/75

73

BOTTOMLINE:BOTTOMLINE:

Many Ways to Optimize Power ConsumptionMany Ways to Optimize Power Consumption

Power aware computingPower aware computing

Ultra-low power microcontrollersUltra-low power microcontrollers

Dynamic power management HWDynamic power management HW

Dynamic voltage scaling (e.g Intels PXA, TransmetasDynamic voltage scaling (e.g Intels PXA, TransmetasCrusoe)Crusoe)

Components that switch off after some idle timeComponents that switch off after some idle time

Energy aware softwareEnergy aware software

Power aware OS: dim displays, sleep on idle times, powerPower aware OS: dim displays, sleep on idle times, poweraware schedulingaware scheduling

Power management of radiosPower management of radios

Sometimes listen overhead larger than transmit overheadSometimes listen overhead larger than transmit overhead

O OBOTTOMLINE


74/75

74

BOTTOMLINE:BOTTOMLINE:

Many Ways to Optimize Power ConsumptionMany Ways to Optimize Power Consumption

Energy aware packet forwardingEnergy aware packet forwarding

Radio automatically forwards packets at a lowerRadio automatically forwards packets at a lowerpower level, while the rest of the node is asleeppower level, while the rest of the node is asleep

Energy aware wireless communicationEnergy aware wireless communication

Exploit performance energy tradeoffs of theExploit performance energy tradeoffs of thecommunication subsystem, better neighborcommunication subsystem, better neighbor

coordination, choice of modulation schemescoordination, choice of modulation schemes


75/75

COMPARISONCOMPARISON

Technology Data RateTx

CurrentEnergy per

bitIdle

CurrentStartup

time

Mote 76.8 Kbps 10 mA 430 nJ/bit 7 mA Low

Bluetooth 1 Mbps 45 mA 149 nJ/bit 22 mA Medium

802.11 11 Mbps 300 mA 90 nJ/bit 160 mA High

IEEE 802.11

Bluetooth

Mote

Energyper bit

Startuptime

Idlecurrent

Lecture3.1Factors Influencing Sensor Network

Documents