Automation & Robotics Research Institute The University of

Automation & Robotics Research Institute The University of Texas at Arlington

F.L. LewisMoncrief-O’Donnell Endowed Chair

Head, Controls & Sensors Group

Talk available online athttp://ARRI.uta.edu/acs

Wireless Sensor Networks

Invited by I-Ming Chen

Wireless Sensor Networks2006 Conferences on Cybernetics & Intelligent Systems (CIS)

and Robotics, Automation, & Mechatronics (RAM)

Bangkok

K.C. TanAbdullah Al Mamun

IEEE Thailand Section

Ivan Boo

First CIS-RAM Singapore 2004IEEE SMC ChapterIEEE R&A ChapterNUS / NTU / SIMTECH

Sam GeK.C. TanDanwei Wangand team

2008 ?

Papers Due 15 Nov

Bhumibol Adulyadej- Rama IX, 1946

SILVER JUBILEE – 60 years

Kingdoms of Thailand

454 Emerald Buddha found in Chiang Rai1279 Sukhotai Kingdom1351 Ayutthaya Kingdom - U Thong1767 Thonburi – Taksin

Emerald Buddha brought to Bangkok1782 Bangkok – Chakri – Rama I1868 Chulalongkorn – Rama V1946 Bhumibol Adulyadej – Rama IX

PDA

BSC(Base Station

Controller, Preprocessing)BST

WirelessSensor

Machine Monitoring

Medical Monitoring

Wireless SensorWireless

Data Collection Networks

Wireless(Wi-Fi 802.11 2.4GHz

BlueToothCellular Network, -

CDMA, GSM)

Printer

Wireland(Ethernet WLAN,

Optical)

Animal Monitoring

Vehicle Monitoring

Onlinemonitoring Server

transmitter

Any where, any time to access

Notebook Cellular Phone PC

Ship Monitoring

I. Wireless Sensor Networks

RovingHumanmonitor

Data Distribution Network

Management Center(Database large storage,

analysis)Data Acquisition

Network

Applications

Wide area monitoring for personnel / vehiclesSecure area intrusion monitoring and denialEnvironmental monitoring

animal habitatsmigrationforest firesnatural disasters

Subsea monitoringEnvironmental toxin detectionBuilding monitoring

Integrity after earthquakes and tornadoesUrban area environmental monitoring

sensors on buildingssensors in taxis or buses

Vehicle traffic monitoring & controlsensors on roadways and traffic lightssensors on vehicles

Remote site power substation monitoringRemote site patient medical monitoringSmart homeInventory management

Latency (delay)Energy efficiencyAccuracyFault-toleranceScalabilitySecurity

Metrics / QoS

• COTS Wireless SensorsBerkeley CrossbowMicrostrain

• Wireless NetworksCellular networkWLANOther short range RF networksMultiple linked networks

Wireless CBM Research Areas• Sensor Technology

MEMS ?• Node Technology

DSPPowerRF link

• Remote Access TerminalsWireless PDA, Wireless LaptopCellphone, Internet

• Data managementSensor data storageDSPData Access

• Fault & Diagnostic Decision-Making• Alarming

Self-OrganizeCommunicationLocalizationForm clusters Monitoring

continuousevent-basedquery

Random vs. Structured topology

Post-deploymentRedeployment of new nodesFault recovery

Deploy Operate Reconfigure

1. Program Missions2. Accomplish Missions

Mobilityuser / observersensorsphenomena / target

Changing Topologymobile nodesevent occurrencemobile target / phenomenachanging user queries/interests

node failuredeploy additional nodes

Coverage

Research TopicsDeploy

Self-organizationComms. wakeupLocalization

Sensing

Event detectionInterpret data

Responsive actionUser broadcast interestRespond to queries

CooperationDynamic Clustering

CommunicationsDynamically reconfigurableEvent-based routing

Data TransmissionEvent-basedData aggregationSensor Data fusionInformation fusionDecision fusion

Fault tolerancenode failurelink failure

Security

Programmable WSNProgram missions quickly

Task schedulingDynamic resource assignment

Scalability- NP complexityDistributed local algorithms vs. global

Meet QoS requirementsCommunicationsSensing High priority data

Decision-making & control

Use Mobility toLocalize nodesMaintain connectivityOptimize comms.Optimize sensor coverageReduce measurement uncertainty Lack of testbeds

Energy conservation

Principal Problems in Army Communications

Paul Baran, Rand Corp.

Cellular Technologies

2G Systems

2.5G Systems

3G Systems

Other Short-range Technologies

Home RF

Bluetooth

IrDA

IEEE 802.11

Wireless LAN Technology

• 2.4 GHz Wireless LAN

• 5 GHz Wireless LAN

• Ad-hoc Mode

• Infrastructure Mode

Which Technology?

IEEE 1451 Standard for Smart Sensor Networks

Long Range Technologies

Cordless Telephony (cellphone)Internet

Frequency Bands for WSN

The Industrial, Scientific and medical (ISM) radio bands were originallyreserved internationally for non-commercial use of RF EM fields forindustrial, scientific and medical purposes. In recent years they have beenused for license free communications applications such as sensor networksand bluetooth.

The bands generally used for WSN in ISM are:

433MHz933 MHz2.4 GHz5.8 GHz

http://www.pctechguide.com/29network.htm

Network Topology

Self Healing Net – Dual Ring

Uniform- for placed sensors-e.g. environmental monitoring

Random- for deployed sensors-e.g. secure area monitoring

Paul Baran, Rand Corp.The Spider Web Net

Paul Baran, Rand Corp.

Centralized, Decentralized, Distributed

Neighbor Connectivity and RedundancyNetwork Topology

Paul Baran, Rand Corp.Connectivity and Number of Links

Number of links increases exponentially

The Problem of Complexity

Communication Protocols in a network must be restricted and organized to avoid Complexity problems

e.g. in ManufacturingThe general Job Shop allows part flows between all machinesThe Flow Line allows part flows only along specific Paths

We have shown that the job shop is NP-complete

but the reentrant flow line is of polynomial complexity

Lewis, Horne, Abdallah, 2000

Think of the military chain of command

Ethernet. The Ethernet was developed in the mid 1970’s by Xerox, DEC, and Intel, and was standardized in 1979. The Institute of Electrical and Electronics Engineers (IEEE) released the official Ethernet standard IEEE 802.3 in 1983. The Fast Ethernet operates at ten times the speed of the regular Ethernet, and was officially adopted in 1995. It introduces new features such as full-duplex operation and auto-negotiation. Both these standards use IEEE 802.3 variable-length frames having between 64 and 1514-byte packets.

Token Ring. In 1984 IBM introduced the 4Mbit/s token ring network. The system was of high quality and robust, but its cost caused it to fall behind the Ethernet in popularity. IEEE standardized the token ring with the IEEE 802.5 specification. The Fiber Distributed Data Interface (FDDI) specifies a 100Mbit/s token-passing, dual-ring LAN that uses fiber optic cable. It was developed by the American National Standards Institute (ANSI) in the mid 1980s, and its speed far exceeded current capabilities of both Ethernet and IEEE 802.5.

Gigabit Ethernet. The Gigabit Ethernet Alliance was founded in 1996, and the Gigabit Ethernet standards were ratified in 1999, specifying a physical layer that uses a mixture of technologies from the original Ethernet and fiber optic cable technologies from FDDI.

Historical Development and Standards

http://www.pctechguide.com/29network.htm

Client-Server networks became popular in the late 1980’s with the replacement of large mainframe computers by networks of personal computers. Application programs for distributed computing environments are essentially divided into two parts: the client or front end, and the server or back end. The user’s PC is the client and more powerful server machines interface to the network..

Peer-to-Peer networking architectures have all machines with equivalent capabilities and responsibilities. There is no server, and computers connect to each other, usually using a bus topology, to share files, printers, Internet access, and other resources.

Peer-to-Peer Computing is a significant next evolutionary step over P2P networking. Here, computing tasks are split between multiple computers, with the result being assembled for further consumption. P2P computing has sparked a revolution for the Internet Age and has obtained considerable success in a very short time. The Napster MP3 music file sharing application went live in September 1999, and attracted more than 20 million users by mid 2000.

802.11 Wireless Local Area Network. IEEE ratified the IEEE 802.11 specification in 1997 as a standard for WLAN. Current versions of 802.11 (i.e. 802.11b) support transmission up to 11Mbit/s. WiFi, as it is known, is useful for fast and easy networking of PCs, printers, and other devices in a local environment, e.g. the home. Current PCs and laptops as purchased have the hardware to support WiFi. Purchasing and installing a WiFi router and receivers is within the budget and capability of home PC enthusiasts.

Bluetooth was initiated in 1998 and standardized by the IEEE as Wireless Personal Area Network (WPAN) specification IEEE 802.15. Bluetooth is a short range RF technology aimed at facilitating communication of electronic devices between each other and with the Internet, allowing for data synchronization that is transparent to the user. Supported devices include PCs, laptops, printers, joysticks, keyboards, mice, cell phones, PDAs, and consumer products. Mobile devices are also supported. Discovery protocols allow new devices to be hooked up easily to the network. Bluetooth uses the unlicensed 2.4 GHz band and can transmit data up to 1Mbit/s, can penetrate solid non-metal barriers, and has a nominal range of 10m that can be extended to 100m. A master station can service up to 7 simultaneous slave links. Forming a network of these networks, e.g. a piconet, can allow one master to service up to 200 slaves.

Currently, Bluetooth development kits can be purchased from a variety of suppliers, but the systems generally require a great deal of time, effort, and knowledge for programming and debugging. Forming piconets has not yet been streamlined and is unduly difficult.

Home RF was initiated in 1998 and has similar goals to Bluetooth for WPAN. Its goal is shared data/voice transmission. It interfaces with the Internet as well as the Public Switched Telephone Network. It uses the 2.4 GHz band and has a range of 50 m, suitable for home and yard. A maximum of 127 nodes can be accommodated in a single network. IrDA is a WPAN technology that has a short-range, narrow-transmission-angle beam suitable for aiming and selective reception of signals.

A catalyst for the standardisation of communications protocols and the functions of a protocol layer.

OSI-Open Systems Interconnection

Cabling

L1 Physical Layer

L2 Link Layer

L3 Network Layer

L4 Transport Layer

L5 Session Layer

L6 Presentation Layer

L7 Applications Layer

Applications Programs

TelnetFTP

TCP, UDP

IP, ICMP

EthernetToken ringFDDIEtc.

OSI/RM

Cabling

L1 Physical Layer

L2 Link Layer

L3 Network Layer

L4 Transport Layer

L5 Session Layer



Applications Programs

TelnetFTP

TCP, UDP

IP, ICMP

EthernetToken ringFDDIEtc.

OSI/RM

L1 Physical Layer

L2 Link Layer

L3 Network Layer

L4 Transport Layer

L5 Session Layer



Communicationsinfrastructure

L1 Physical Layer

L2 Link Layer

L3 Network Layer

L4 Transport Layer

L5 Session Layer



Sensingapplication

Akyildiniz, Su, et al. 2002

Cross-layer design

ConfigureMaintainOptimize

e.g. Integrate navigation, communication, congestion control, and sensing

http://ieee1451.nist.gov/intro.htm

Objective of IEEE 1451 - Smart Transducer Interface Standard.To make it easier for transducer manufacturers to develop smart devices and to

interface those devices to networks, systems, and instruments by incorporating existing and emerging sensor- and networking technologies.


History of IEEE-1451In September 1993, the National Institute of Standards and Technology (NIST) and the Institute of Electrical and Electronics Engineers (IEEE)'s Technical Committee on Sensor Technology of the Instrumentation and Measurement Society co-sponsored a meeting and set up four working groups -

P1451.1 Defining a common object model for smart transducers along with interface specifications for the components of the model.

P1451.2 Defining a smart transducer interface module (STIM), a transducer electronic data sheet (TEDS), and a digital interface to access the data.

P1451.3 working group aims at defining a digital communication interface for distributed multidrop systems.

P1451.4 Defining a mixed-mode communication protocol for smart transducers. The working groups created the concept of smart sensors to control networks

interoperability and to ease the connectivity of sensors and actuators into a device or field network.

sensor signalconditioning

DSP

local userinterface

applicationalgorithms

data storage

communicationanalog-to-

digitalconversion

NETWORK

hardwareinterface

Network SpecificNetwork Independent

Virtual Sensor


Concept of Smart Sensorcontains functions in addition to those needed foraccurate presentation of the measurand


XDCR ADC

XDCR DAC

XDCR Dig. I/O

XDCR ?

TransducerElectronic DataSheet (TEDS)

addresslogic

Smart Transducer Interface Module (STIM)

NETWORK

Network CapableApplication

Processor (NCAP)

1451.1 ObjectModel

TransducerIndependent

Interface (TII)

1451.2 Interface

XDCR ADC

XDCR DAC

XDCR Dig. I/O

XDCR ?

TransducerElectronic DataSheet (TEDS)

addresslogic

Smart Transducer Interface Module (STIM)

NETWORK

Network CapableApplication

Processor (NCAP)

1451.1 ObjectModel

TransducerIndependent

Interface (TII)

1451.2 Interface

Sensors &Transducers

Xbow wireless sensor boards

• Temperature, ambient light, acoustic sensors, accelerometer, and magnetometer, (can get GPS)

• Each node is endowed with a microcontroller, programmable with a C-based operating system

• Cricket motes have ultrasound rangefinders• 433 or 933 MHz ISM band

Environmental Monitoring & Secure Area Denial

Microstrain V-Link

Transceiver

MicrostrainTransceiver

Connect to PC

MicrostrainG-Sensor

Microstrain, Inc., Wireless Sensors

RFID node

http://www.microstrain.com/index.cfm

MicrostrainG-Sensor

MicrostrainTransceiver

Connect to PC

Microstrain V-Link

Transceiver

WSN for Machinery Monitoring- Diagnostics & Prognostics

Signal ConditioningTemperature compensation

Low pass filter for noise rejection

V1 V2

R1 R2

C

V1 V2

R1 R2

CAnalog LPF

asaksH+

=)( kk szzKsα−+

=1ˆ

Digital LPF

)(ˆ 11 kkkk ssKss ++= ++ αDifference equation

Filtered velocity )( 11 kkkk ssKvv −+= ++ α

R1 R2

R3 R4

Vref

Vo

R1 R2

R3 R4

Vref

Vo

Wheatstone BridgeChanges in R converted to changes in VImproved sensitivity

DSP- Kalman Filter

kkkk AKzBuxKHIAx ++−=+ ˆ)(ˆ 1

LabVIEW Real-time Signaling & Processing

CBM Database and real time Monitoring

PDA access Failure Data from anytime and

anywhere

User Interface, Monitoring, & Decision AssistanceWireless Access over the Internet

II. ARRI Distributed Intelligence & Autonomy LabDIAL

UnattendedGroundSensors

SmallmobileSensor-Dan Popa

Testbed containing MICA2 network (circle), Cricket network (triangle), Sentry robots, Garcia Robots & ARRI-bots ( Dan Popa)

!

DEC for WSN

Mission result:

False fire alarmRobot1 goes to

sensor 2Sensor 1

output:alarmRobot 1 picks

sensor 2Robot2 follows

robot1Sensor 3

measurementSensor 4

measurementRobot1 places

sensor 2Sensor2

measurement

Robot1 goes to sensor1

Robot2 smoke detection

Programmable Missions

Table I- Mission1-Task sequence

Mission 1 completedy1output

S2 takes measurementS2m1Task 11

R1 takes measurementR1m1Task 10

R1 deploys UGS2R1dS21Task 9

R2 takes measurementR2m1Task 8

R1 gores to UGS1R1gS11Task 7

R1 listens for interruptsR1lis1Task 6

R1 retrieves UGS2R1rS21Task 5

R2 goes to location AR2gA1Task 4

R1 goes to UGS2R1gS21Task 3

UGS5 takes measurementS5m1Task 2


UGS1 launches chemical alertu1Input 1

Descriptionnotationmission1

Table II- Mission 2-Task sequence

Mission 2 completedy2output

R1 docks the chargerR1dC2Task 5


R1 charges UGS3R1cS32Task 3

R1 goes to UGS3R1g S32Task 2


UGS3 batteries are lowu2input

DescriptionnotationMission2

Fast Programming of Missions

Multiple Missions

Node Deployment & Failure- Modify Fr

RESOURCE RESET LOGIC:

MISSION COMPLETE LOGIC:

Wireless Sensor Net

Sensor reading events

Tasks performed

Resources available

PerformanceMeasures

Targets or Events In

Matrix DE Controller

u

v

y

vs

rs

vs

dudurv uFuFrFvFx +++=xSv vs =

xSr rs =

xSy y=

Program DEC For WSN Applications

Program Missions- Selection of matrices

Select Resources- Priority modification of Fr NEXT TASK LOGIC:

Missions completed

Sensor readings

Tasks completed

Resources Idle

Missions completed

Task Commands

Resource Reset Commands

WSN logical Status information

y

Node Deployment & Failure- Modify Fr

RESOURCE RESET LOGIC:

xSr rs =

rs

Deadlock avoidance policy

Discrete Event Supervisory ControllerUS Patent

Supervisory Control of Mobile Wireless Sensor Networks

LabVIEW User Interface

Fast programming of multiple missionsReal-time event responseDynamic assignment of shared resources

NetEntry-inviteresponseponter

StartupNodeID nr.

Neighborinfo

I R P

Dist.toNeigh-bors

Comm link mesh info Position grid info

(x,y) coords.and OriginNode ID

Hier.routingnr.

extra

T frame

repeat nextTDMA frame

NetEntry-inviteresponseponter

StartupNodeID nr.

Neighborinfo

I R P

Dist.toNeigh-bors

Comm link mesh info Position grid info

(x,y) coords.and OriginNode ID

Hier.routingnr.

extra

T frame

repeat nextTDMA frame

TDMA frame for both communication protocols and relative positioning

Node Relative Positioning & LocalizationAd hoc network- scattered nodes

Nodes must self organize

Calibrated network-Each node knows its relative position

1 2d12 x

a. Two nodes- define x & y axes

y

b. 3 node closed kinematic chain-compute (x3, y3)

O 1 2d12 x

d23d13

3

y

x3

y3

θ213 θx23O1 2d12 x1 2d12 x

a. Two nodes- define x & y axes

y

b. 3 node closed kinematic chain-compute (x3, y3)

O 1 2d12 x

d23d13

3

y

x3

y3

θ213 θx23O 1 2d12 x

d23d13

3

y

x3

y3

θ213 θx23O

Integrating new nodes into relative positioning grid

1 2A12 x’

A23

A13

3

y’

O’

x

y

O

TOO’

A14

A34

A24

4

1 2A12 x’

A23

A13

3

y’

O’

x

y

O

TOO’

A14

A34

A24

4

Recursive closed-kinematic chain procedure for integrating new nodes

⎥⎦

⎤⎢⎣

⎡=

10ii

i

pRA

One can write the relative location in frame O of the new point 3 in two ways. The triangle shown in the figure is a closed kinematic chain of the sort studied in [Liu and Lewis 1993, 1994]. The solution is obtained by requiring that the two maps T13and T123 be exact at point 3.

Kinematicstransformation

Adaptive Sampling with Mobile Sensor Nodes Dan Popa, Sreenath, Mysorewala, F.L. LewisICCA Budapest 2005

kkkkk wuxhxx ++=+ ),(1

kkk xfy ξ+= )(

kkkk axgz ν+= ),(

kTkk QwwE =][

kTkk RE 1][ =ξξ

).(...)(11 kmmkok xgaxgaaz +++=

Mobile node dynamics

Mobile node position measurement

Distributed field measurement model

Sum of Gaussian model (RBF neural network)

,1

kkkk

kk

AGzAA

ν+==+

))()...(...1( kmkik XgXgG = RE Tkk =][ νν

).(

,

),,(~

^

111

11

^

1

^1

11

111

_

0

kkkTkkkk

kTkkk

AoAoo

AGZRGPAA

GRGPP

PPPAAo

++−

+++

+−

+−−

+

−+=

+=

=

A. Estimation of Field Without Localization Uncertainty

B. Estimation of Field With Localization Uncertainty

( ) ( ) ,10

01

,00

3

3

1

1

kk

kTkk

k

kTk

k

k

k

kkk

kkk

k

k

k

k

AX

XI

AXX

ZY

BUAXw

UI

AX

AX

λνξ

ϑ

+⎟⎟⎠

⎞⎜⎜⎝

⎛⎟⎟⎠

⎞⎜⎜⎝

⎛=⎟⎟

⎠

⎞⎜⎜⎝

⎛+⎟⎟

⎠

⎞⎜⎜⎝

⎛=⎟⎟

⎠

⎞⎜⎜⎝

⎛

++⎟⎟⎠

⎞⎜⎜⎝

⎛=⎟⎟

⎠

⎞⎜⎜⎝

⎛+⎟⎟

⎠

⎞⎜⎜⎝

⎛+⎟⎟

⎠

⎞⎜⎜⎝

⎛=⎟⎟

⎠

⎞⎜⎜⎝

⎛

+

+

( )

).(

,10

0

))((

,

1

11

1

1111

1

1^

1

^

1

1

13

11

11

111

^

^

1

11

⎟⎟⎠

⎞⎜⎜⎝

⎛−⎟⎟

⎠

⎞⎜⎜⎝

⎛+⎟⎟

⎠

⎞⎜⎜⎝

⎛=⎟

⎟

⎠

⎞

⎜⎜

⎝

⎛

+⎟⎟⎠

⎞⎜⎜⎝

⎛=⎟⎟

⎠

⎞⎜⎜⎝

⎛⎟⎟⎠

⎞⎜⎜⎝

⎛=

+=

+⎟⎟

⎠

⎞

⎜⎜

⎝

⎛=⎟⎟

⎠

⎞⎜⎜⎝

⎛+=

−+

−+

+−+

+−++−

+

−+

+

+

−+

−+

−+

−+

−−++

−+

−+−

+

k

kk

k

kTkk

k

k

k

k

kk

k

k

kTk

k

kTkkk

k

k

k

k

kkk

AX

GZY

RGPAX

A

X

BUAX

AX

XI

G

GRGPP

BUA

XAX

QPP

Cooperative Mapping and Localization

Distributed Network Architecture

Multi-sensor Fusion for Distributed Fields

Select next sample point to minimize uncertainty Kalman Filter

Implementation at ARRI’s Distributed Intelligence & Autonomy Lab (DIAL)

Measured Field is a color map. Mobile robots have color sensors. ybxbbBygxggGyrxrrR 210210210 ,, ++=++=++=

0.001,0.00078,0.10.0018,,0.0002,00.00048,,0.0012,0.2307

2102

10210

−=−=====−===

bbbgggrrr

Mobile sensorsBuilt at DIAL LabBy Dan Popa

Raster Scan Adaptive Sampling

Dan Popa

Greedy Adaptive Sampling Algorithm

876

54

321

876

54

321

Current Sampled Location

Select next sample point to minimize covarianceonly among neighboring cells

Cross-Layer Navigation Using Potential Fields Dan Popa

)(rU r−∇=F attractive forces to the goals, repulsive forces among the robots and obstacles

)(),( ijijrestore uji rrF −= Restoring force to avoid getting out of communication range

iiiii vm Frr =+ &&& Mobile node eqs. of motion

⎟⎟⎠

⎞⎜⎜⎝

⎛+= αd

PWN

KWC t

o

1log2Shannon Link communication capacity

with internode distance d

rrPk

∂∂

−=||))((||

infF )(rPk is the adaptive sampling error covariance calculated via the EKF

∫=+=t

o

iiiiioi dtEtEkt τττνν ν )()()()),(1()( rF & Conserve energy by making damping increase withmotion energy expended

Information potential

)ˆ()ˆ(ˆ111 kk

Tkkk XXWXXM −−= −

+−++ Work to go to next predicted state for adaptive sampling

CFc −∇=

Energy cons.

Initial configurationNode 20 at (0,0) is a sink

Final configuration after(7,8) is selected as a target point

Nodes 3, 12, 14 go to (7,8) to sense informationOther nodes move to maintain comm. links

Dan Popa

Dynamic Localization of Mobile WSN Dang, F. Lewis, D. Popa

[ ]Tiii yxX =

⎥⎥⎦

⎤

⎢⎢⎣

⎡⎥⎦

⎤⎢⎣

⎡+⎥

⎦

⎤⎢⎣

⎡⎥⎦

⎤⎢⎣

⎡=⎥

⎦

⎤⎢⎣

⎡y

i

xi

i

i

i

i

ff

IO

XX

OOIO

XX

2

2

22

22&&&

&

Node position

Estimator for position

∑∑=

≠=

−=N

i

N

jij

ijijijugs rrKV1 1

2)(21

iv

N

j ji

jiijijiji XK

XX

XXrrKf &

r−

−

−−−= ∑

=1

)()(

∑=

+=N

ii

Tiugs XXVL

1 21 &&

Potential fn.

Theorem. Let virtual force be given by

Then the position estimates reach steady-state values that provide optimal estimates of the actual relative localization of the nodes in the sense that is minimized

Proof:

1. Relative Localization

mpX aip

,...,2,1; =

∑∑∑== =

+−=m

p

ai

ai

m

p

N

j

aji

aji

ajiuav ppppp

eKrrKV1

2

1 1

2

21)(

21 [ ] 2

122 )()( a

ia

ia

ia

ia

i pppppyxxxe −+−=

∑ ∑

∑∑∑

+= =

= ==

−+

−+=

N

mp

N

jjijiji

m

p

N

j

aji

aji

aji

m

p

ai

aip

ppp

ppppp

rrK

rrKeKV

1 1

2

1 1

2

1

2

)(21

)(21

21

p

p

p

pppp iv

N

j ji

jijijijii XK

XX

XXrrKf &−

−

−−=∑

=1

)()(

ai

av

N

j ja

i

ja

iaji

aji

aji

ai

ai

ai

ai p

p

p

pppppppXK

XX

XXrrKXXKf &−

−

−−−−−= ∑

=1

)()()(

2. Absolute Localizationm nodes with GPS

abs. loc. pot. fn. with

Theorem. Let virtual force be given by

nodes with no GPS

nodes with GPS

Proof:

Range-Free Localization of Mobile WSN Koushil Sreenath, F.L. Lewis, Dan PopaCIS/RAM Bangkok 2006

ik

ik

ik

ik

ik

ik

ik wGuBxAx ++=+1

ik

ik

ik

ik vxHz +=

⎥⎦

⎤⎢⎣

⎡=⎥

⎦

⎤⎢⎣

⎡=⎥

⎦

⎤⎢⎣

⎡=⎥

⎦

⎤⎢⎣

⎡=

1001

,1001

,0000

,1001 i

kik

ik

ik HGBA

,0

0,

⎥⎥⎦

⎤

⎢⎢⎣

⎡== Bot

y

BotxBotBot

kRσ

σσσ

const

BotyBot

yconst

BotxBot

xRangeRangeσ

σσ

σ == ,

Algorithm 1 : Static sensor node localization algorithm1At each discrete time instant,2if robot broadcast received by sensor3then4 Update sensor state and uncertainty estimates using KF5else6 Propagate estimates using time updates

7end if

1. Localization of Stationary Nodes

uncertainty in comm. range

The first reading localizes the node to a projection on the robot’s path Kalman Filter on UGS

( ) ( )wtGtuXaX += ,,&

[ ] gpskk

gpsgpsk vktXhZ += ),(

( ) ( )⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢

⎣

⎡

=

⎥⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢⎢

⎣

⎡

=

⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢

⎣

⎡

=

0000000000100001

,sin

sincoscoscos

, tGL

vvv

yx

tXat

t

t

αωα

φαφα

αφ&

&&

&

[ ] [ ] ⎥⎦

⎤⎢⎣

⎡=⎥

⎦

⎤⎢⎣

⎡=

yx

ktXhyx

ktXh kugs

kgps ),(,),(

const

iyi

yconst

ixi

x

iy

ixiiiugs

k

RangeRange

PR

σσ

σσ

σσ

σσ

==

⎥⎥⎦

⎤

⎢⎢⎣

⎡=⎥⎦

⎤⎢⎣⎡ +=

,

00

,22

Includes uncertainty in position and in comm. range

Algorithm 2: Mobile robot localization algorithm.1Navigate robot along desired path.2Broadcast location information at discrete intervals.3if broadcast from GPS received4 Update robot state and uncertainty estimates using measurement Eq. (20).5end if6if broadcast from sensor received7 Update robot state and uncertainty estimates using measurement Eq. (21).

8end if

2. Simultaneous Localization of Mobile Robot & Stationary Nodes

GPS update when available

Update from UGS position when available

[ ] ugskk

ugsugsk vktXhZ += ),(

Mobile robot localization turned off With Mobile robot localization

Kalman Filters on UGS and Mobile Robot

3. Adaptive Localization Mobile robot moves to localize the un-localized sensors

Network communication connectivity is exploitedInitiation of the navigation request "NAV-REQ" packet from the robot

Badly localized sensors reply back with a localization request "LOC-REQ" packet. Already localized adjacent receiving nodes add their location and forward the request.

Algorithm 3 : Adaptive localization algorithm.1Broadcast Navigation request, NAV-REQ, packet.2Wait to receive Localization request, LOC-REQ, packets.3for all LOC-REQ with the same friendly neighbor4 Combine uncertainty scalars of the requesting sensors.5end for6Pick friendly neighbor with maximum combined uncertainty scalar of the requesting sensors.7if multiple maximas arise8 Among the maxima, pick the most localized friendly neighbor.9end if10Navigate around the picked friendly neighbor executing the simultaneous localization algorithm, on the sensors and on the mobile robot.

11Repeat Steps 1-10 as required.

Problem- how does it know where to go to localize nodes with unknown positions?

Network layer

Sensor Protocols for Information via Negotiation

RoutingData fusion

SPIN protocol

Directed diffusion

Interest disseminationResponsive actionEvent detection

publish / subscribe

advertise interest

Akyildiniz, Su, et al. 2002

RoutingMinimum energyMinimum hopMax. min power available

usersensor

User/

III. Issues and Tools in WSN

Election of cluster headsEvent-basedApplication-basedLEACH

Hierarchical Routing Allows Multicast – Efficient Routing

5 links

18links

source node destination

Standard peer-to-peer routing Multicast routing

1. Source to leader 2. Leader to destination

Taken from Chen et al. (2000)

group leader

15 linkstotal

7 links

23 linkstotal

8 links

LimitedRangePowerProcessing power / memoryCost

Large number of nodesProne to failuresEasy to be compromisedChanging topologyLack of global ID

Sensor Management Protocol (SMP)Attribute-based namingLocation-based addressingData-centric routingUser broadcast interest

DisseminateSensor dataInformationUser Interest

Long-term reliability

Figure courtesyAkyildiniz, Su, et al. 2002

Tasking sensor node

Extra Connectivity routing sensor nodes

Sensor Placement and Lifetime EstimationJain and Qilian Liang, 2005

Failure of two nodes causesloss of sensor coverage

Square grid

Hex grid

Reliability TheorySurvivor Function = prob. that a unit is still functioning at time t

s(t)= 1- cdf

Reliability block diagram of square grid

32

)1)(1(1

sss

ssss cbablock

−+=

−−−=

Nblocknet ss )(=

Reliability block diagram of hex grid

)1)(1(1 3sssblock −−−=2/)( N

blocknet ss =

∑−

=ii

thresholdinitnode Pw

EET

Node lifetime

power consumed in mode i

fraction of time spent in mode i

Assume only 2 modes, then binomial pdf xTxx ppcxwP −−== )1(}{ 1

Pr node is idle

Pr node is active (defined by net protocol)

Finding node lifetime pdf

Results T= nr. of time units

Can predict net lifeMore accurately thanNode life

Energy Conserving Sensor Coverage

Sample time #1 Sample time #2

Selected Sensors

Extra nodes selectedfor connectivity

Selected Sensors

Extra nodes selectedfor connectivity

Grey= area not covered

Entire area covered in 2 sample timeslatency (delay) = 2

Formal algorithms for specifying QoS% coverage of sensorsmax latency

Choi and S. Das, Mar 2005

Math Basis

Select min. nr. K of sensors s.t.U

k

iiSRQDSC

1=

∩⊂

Circular sensing region SRi of radius r

and entire region is covered within desired latency T

TtN

ii ≤∑

=1Assume:

sensors are uniformly distributedlocation info not available

Find probability that a point (x,y) is not covered by randomly selected sensor ),( yxqPq

Then, min. number of sensors needed to cover DSC is

⎟⎟⎠

⎞⎜⎜⎝

⎛

+++

−=

22

2

44log

)1log(

raraara

DSCk

π

a

DSC= probability of coverage of point (x,y)1. Find Required Number of Sensors for DSC

2. Add Extra Routing Nodes for Comm. Connectivity

Test probable connectivity of k sensors in k-1 steps, adding nodes when needed

3. Construct Data Gathering Tree (DGT)For routing and sensor scheduling

Data sink sends flood messageEach sensor keeps a forwarding record

with best upstream candidateSensors broadcast join request setup msgs.

Localized sensor scheduling algorithm

Λ= /inodeofrangeradioPis

iterationnumber

Connectedset

Other nodes

Distributed Greedy Algorithm for Connected Sensor Cover

Find minimum connected sensor cover (MCSC)

Def. MCSC1. Monitored area contained in Union of node sensor regions2. Induced communication graph is connected via multihop

Energy conserving sensor coverage

Problem of finding MCSC is NP-hard [Garey and Johnson 1991]

Ghosh and S. Das, June 2005

Communication radius Rc

Induced comm. graph Gc=(V,ERc)edge i,j exists if d(si,sj)< Rc

Induced sensing graph Gs=(V,ERs)edge i,j exists if d(si,sj)< 2Rs

Graph = (nodes, edges)

Sensing radius RsAssume sc RR 2≥

Def. Independent SetA subset of vertices such that no two vertices has an edge in G.

Def. MISAn IS that is not contained in any other IS

Finding MIS for a general graph is NP-hard

Phase 1 – Find Maximal Independent Set (MIS)

Phase 1 – Find Maximal Independent Set (MIS)

Use greedy approach looking only at 1-hop nearest neighbors

Def. Eligible next node given node si1. sj not yet included in the connected MIS2. sj a one-hop neighbor of si3. sensing circle of sj does not overlap any selected sensing circles

Suboptimal MCSC using greedy approach

Phase 2- Select extra nodes to get full sensor coverage

Construct Voronoi Diagram for nodes selected in Phase 1

Voronoi Diagram divides the plane into convex polygons whose edges are equidistant from two nodes

Algorithm 2- Choose best 1-hop neighbor that maximally covers holes in its polygon

Voronoi structure allows efficient formal algorithm for doing this

Result of Phase I MIS

has holes

Time complexity of first algorithm is

Time complexity of second algorithm is

Math Analysis

Let N nodes be uniformly randomly distributed over area A. Density is

Then number of nodes in Phase I MIS is bounded by 225 ss R

NR

Nρπ

ζρπ

≤≤

AN /=ρ

)( NO ζ

)log( ζζO

Network SecurityGroup Key Distribution via Local Collaboration

Chadha, Y. Liu, and S. Das, Sept. 2005

1. (n,t) Threshold cryptography via polynomials

Random secret polynomial 1110)( −−++= t

t xaxaaxf where secret key is D= f(0)

f(x) Can be reconstructed from t points from the set {f(1), f(2), …, f(n)}, with n= number of nodes

Select masking polynomial h(x) and securely predeploy personal secrets h(i) on each node i.

The Sink broadcasts w(x)= f(x)g(x)+h(x)

Where revocation polynomial is ))...()(()( 21 wrxrxrxxg −−−=

with the set of compromised nodes },...,{ 1 wrr

Then each node i can evaluate its personal key)(

)()()(ig

ihiwif −=

Compromised nodes have g(i)=0 and cannot find personal key

Now, t nodes can collaborate to exchange personal keys f(i) and so compute f(x), and hence find the secret group key D= f(0)

Since h(i) is securely predeployed and f(x) is random, the scheme can be shown to be unconditionally secure

2. Enhancement to avoid disclosing personal key f(i)

Select a random concealing polynomial L(x) of degree t-1 and securely predeploy concealing secret L(i) on each node i

Instead of exchanging personal keys f(i), the t collaborating nodes exchange the concealed personal key s(i)= f(i) + L(i)

Since s(x) has degree t-1, it can be reconstructed using t concealed personal keys from t nodes

Then, the group secret is D= s(0)

The sink selects a random t-1 degree polynomial f(x) such that the secret isD= f(0) + L(0)

The concealing key allows improved defense against compromised nodes, who now do not know the personal keys f(i)

This also allows self-healing strategies, i.e. in the presence of lost broadcast messages from the Sink

Theorem 1. Assume that the local exchange of concealed secrets is secure.Then the scheme is unconditionally secure, and has t-revocation capability.

Multi-Layer Approach to Defense Against Compromised Nodes

Y. Liu and S. DasCross-Layer Design

Distributed Energy-Efficient Self-Organization Zhao, Hong, Qilian Liang, Nov. 2004

LEACH selects cluster heads based on energy availableSelects randomly and does not give evenly spaced headsRequires global info – total number of nodes, and total energy available in all nodes

Expellant Self-Organization (ESO)

Based on cluster radius Rc ,energy available, and number of neighbors

)()1()(max 11 iEmeaniEENiNi

th∈∈

−+= γγ

)()1()(max 11 inmeaninnNiNi

th∈∈

−+= γγ

Energy threshold

E(i)= energy of node IN= set of neighbors

Number of neighbors threshold

n(i)= number of neighbors of node i

Fault-Tolerant & Energy Efficient Cross-Layer Routing Qilian Liang, Oct. 2005

Select node for next hop transmission using: 1. Distance of next node (NN) to destination- should be small 2. Remaining battery capacity of next node- should be large3. Mobility of next node- should be small

Fuzzy Logic

The required information is periodically locally broadcast via beacons

Rules: If (NN is near dest.) and (NN has large remaining energy) and (NN is stationary)THEN (NN is a strong candidate)

When a node wishes to transmit, it sends a ROUTE NOTIFICATION, and the receiving nodes send a REPLY packet

If a node fails, the previous node broadcasts a ROUTE DELETION packet

Comm. Packet Structure & Comm. protocol

Cross-Layer Routing in Sensor Networks

Luo, Yonghe Liu, Sajal Das, 2006

Transmission cost t(e)= w(e) c(e), e= an edge

Graph (nodes, edges) G= (V,E) Find Data Gathering Tree

w(e)= amount of datac(e)= transmission cost – congestion, distance, latency, energy used

Fusion cost f(e)

Fusion Cost for L bits = 2L x 5 nano Joulesthe resulting data has L(1+η) bits

re αη −−=1r = node separation, good for field measurements

Data correlation model

Maximize the cost with link cost factors given as

remaining energy at next node

(delay to reach next node) X (dist. from next node to destination )

Fonda, Zawodniok, S. Jagannathan, ISIC Munich 06

Dijkstra’s Algorithm can be used

Define cross-layer cost

Minimize the Total Cost ∑∈

+routee

etef )]()([

Transmission Costs

Transmission cost per hop 42, ≤≤∝ γγd

Total energy per packet using N hops

∑+=hops

idNE γβα

Trans. costSetup cost

Sensor / Data Fusion Luo, Lin, Scherp 1988

)}}({{),( 121 LCEfPPd =x1 x2 x

P1= P1(x/x1) P2= P2(x/x2)

)/()/()(

22

11

xxPxxPxL =

General distance measure

J-divergence

}log)1{(),( 121 LLEPPJ −=

Likelihood ratio

2/12121 })1{({),( −= LEPPd

Matsusita distance

For data representing the same property:

)1log(),( 221 dPPB −−=

Bhattacharyya’s distance

f(.) an increasing fn.C(.) a convex fn.

AdxxPxxPdj

i

x

xiiiiij 2)()/(2 == ∫

BdxxPxxPdi

j

x

xjjjjji 2)()/(2 == ∫

Distance is not symmetric

Confidence distance measures

For multiple sensors, use matrix

⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢

⎣

⎡

=

mmm

m

dd

ddddd

D

1

2221

11211

OM

L

Draw a digraph having edge (i,j) if

thresholddij ≤

1 2

34

Sensor 2 supportsSensor 3

Outlier-Correct or discard

Luo, Lin, Scherp 1988

Information Fusion & Sensor Selection in WSN “MIDFUSION,” Alex, Mohan Kumar, Behrooz Shirazi

Select the best set of sensors that meet the goals of the applications with guaranteed QoS

Middleware for customized services and resource assignment

Expected utility given evidence

∑=i

ninnin SGUSSGPSUE }))({(})/{})({(})({

})({ ni SG is the goal state reached as a result of the selected sensor set }{ nS

Utility of sensors can be written }{})({( nni uSGU =

Where utility factor for sensor sn is)(cost

1)1(n

nn sDu αα −+=

and Dn is a measure of the definitiveness or accuracy of sensor sn

Use Bayesian Networksgiven by user constructed by

middlewaregoal

Internal states

observables

sensors

Security Threat Example

BN showing conditional probabilities

Utility is maximized by using sensor set {RFID2, Video3}This gives 70% threshold

Jianyong Lin, Wendong Xiao, Lihua Xie, 2006

Adaptive Multi-Sensor Scheduling for Target Tracking in Wireless Sensor Networks

{ } ( )Pr 1 1 MM α= − −

Probability of successful detection by at least one sensor

α= Pr detection by 1 sensorin detection range of M sensors

( )( )

log 1log 1

dMθα

− −≥− −

1. Required Sensor Coverage

Sensor density needed is

θd= desired detection probability Kalman Filtering

2. Sampling Time

( ) ( )( )( )( )( ) ( ) ( ) ( )( )

( ) ( ) ( ) 2 311 33 12 34 22 44

1 1

1

,

22 23k k

T

T T Tk k k

k

k k trace k k

trace L P k k L

trace L F t P k k F t L L Q k t L

t t q tσ σ σ σ σ σ

Φ + = Σ +

= +

= Δ Δ + Δ

= + + + Δ + + Δ + Δ

Predicted position uncertainty

( ) 01k kΦ + = Φ

Maximum sampling interval can be solved from the equation

0 0Φ >Allowable uncertainty threshold

target

Tracking trajectory by multi-sensor scheduling scheme

Target loss time of two tracking schemes

⎣ ⎦∑ −−−=j

Njcjf sdTcjTtwts )()( /δ

where w(t) is the basic pulse of duration approx. 1ns, often a wavelet or a Gaussian monocycle, and Tf is the frame or pulse repetition time. In a multi-node environment, catastrophic collisions are avoided by using a pseudorandom sequence cj to shift pulses within the frame to different compartments, and the compartment size is Tcsec. Data is transmitted using digital pulse position modulation (PPM), where if the data bit is 0 the pulse is not shifted, and if the data bit is 1 the pulse is shifted by d. The same data bit is transmitted Ns times, allowing for very reliable communications with low probability of error.

Ultra Wideband Sensor WebUWB

Precise time of flight measurement is possible.Use UWB for all three:

CommunicationsNode Relative positioningTarget localization

1 2

3

T

d d2

d3

x

y y

1 2

3

T

d d2

d3

x

x’y’

θ213

a. Target, transmitter node, and 2 receiving nodes b. Ellipsoid solution for multi-static target localizing

1 2

3

T

d d2

d3

x

y

1 2

3

T

d d2

d3

x

y y

1 2

3

T

d d2

d3

x

x’y’

θ213

1 2

3

T

d d2

d3

x

x’y’

θ213

a. Target, transmitter node, and 2 receiving nodes b. Ellipsoid solution for multi-static target localizing

Multi-Static Radar Target Localization

22122 ,2/,2/)( sabdsdda −==+=

Intersection of two ellipses with semimajor and semiminor axes

Simultaneous solution of two quadratic equations, one for each ellipse

11

=

=

BXXAXX

T

T

Uses time of flight

gives position of target.

References

\item W. Choi and S. K. Das,``A Novel Framework for Energy-Conserving Data Gathering inWireless Sensor Networks,"{\em Proceedings of IEEE INFOCOM}, Miami, Florida, Mar 2005.

\item A. Ghosh and S. K. Das,``A Distributed Greedy Algorithm for Connected Sensor Coverin Dense Sensor Networks,"{\em Proceeeings of IEEE International Conference on DistributedComputing in Sensor Systems} (DCOSS), Marina del Ray, CA,pp. 340-353, June 2005.

\item A. Chadha, Y. Liu and S. K. Das,``Group Key Distribution viaLocal Collaboration in Wireless Sensor Networks,"{\em Second IEEE International Conference on Sensorand Ad Hoc Communications and Networks} (SECON),Santa Clara, Sept 2005.

\itemW. Zhang, S. K. Das and Y. Liu, ``Security in Sensor Networks,"{\em Security in Wireless Sensor Networks: A Survey} (Ed. Y. Xiao), CRC Press, 2006.

\item H. Luo, J. Luo, Y. Liu and S. K. Das,``Routing Correlated Data with Fusion Cost in Wireless Sensor Networks,"{\em IEEE Transactions on Mobile Computing}, to appear, 2006.

Ekta Jain, Qilian Liang, “Sensor placement and lifetime of wireless sensor networks: theory and performance analysis,”Sensor Network Operations, edited by S. Phoha, T. F. La Porta, and C. Griffin, IEEE Press, 2005.

Qilian Liang, “Fault-Tolerant and Energy Efficient Wireless Sensor Networks: A Cross-Layer Approach,” accepted by IEEE Military Communication Conference, Atlantic City, NJ, Oct 2005.

Liang Zhao, Xiang Hong, Qilian Liang, “Energy-Efficient Self-Organization for Wireless Sensor Networks: A Fully Distributed Approach,” IEEE Globecom, Nov 2004, Dallas, TX.

I.F. Akyildiniz, W. Su, Y. Sankarasubramanian, and R. Cayirci, “A survey on sensor networks,” pp. 102-114, IEEE Comm. Mag., Aug. 2002.

R.C. Luo, M.-H. Lin, R.S. Scherp, “Dynamic multi-sensor data fusion system for intelligent robots,” IEEE J. Robotics & Automation, vol. 4, pp. 386-396, Aug. 1988

F.L. Lewis, B.G. Horne, and C.T. Abdallah, "Computational complexity of determining resource loops in reentrant flow lines," IEEE Trans. Systems, Man, and Cybernetics, vol. 30, no. 2, pp. 222-229, 2000.

Jianyong Lin, Frank Lewis, Wendong Xiao, Lihua Xie, “Adaptive Multi-Sensor Scheduling for Target Tracking in Wireless Sensor Networks,” 2006, submitted

1.G. Vachtsevanos, F.L. Lewis, M. Roemer, A. Hess, B. Wu, Intelligent Fault Diagnosis and Prognosis for Engineering Systems, John Wiley, New York, 2006, to appear.

2.J. Mireles, F.L. Lewis, A. Gurel, and S. Bogdan, “Deadlock Avoidance Algorithms and Implementation, a Matrix-Based Approach,” in Deadlock Resolution in Computer-Integrated Systems, chapter 7, ed. Mengchu Zhou, Marcel Dekker, New York, 2004.

3.F.L. Lewis, “Wireless Sensor Networks,” in Smart Environments: Technologies, Protocols, Applications, ed. D.J. Cook and S.K. Das, Wiley, New York, 2004.

4.V. Giordano, F.L. Lewis, P. Ballal, and B. Turchiano, “Supervisory control for task assignment and resource dispatching in mobile wireless sensor networks,” in Cutting Edge Robotics, ed. V. Kordic, p. 133-152, 2005.

5.D.O. Popa and F.L. Lewis, “Algorithms for robotic deployment of WSN in adaptive sampling applications,” in Wireless Sensor Networks and Applications, ed. Y. Li, M. Thai, and W. Wu, Springer-Verlag, Berlin, 2005.

6.N. Swamy, O. Kuljaca, and F.L. Lewis, “Internet-based educational control systems lab using NetMeeting,” IEEE Trans. Education, vol. 45, no. 2, pp. 145-151, May 2002.

7.V. Giordano, P.Ballal, F.L. Lewis, B. Turchiano, J.B. Zhang, “Supervisory control of mobile sensor networks: Matrix formulation, simulation and implementation,” IEEE Trans. Systems, Man, Cybernetics, Part B, to appear, 2006.

8.B. Harris, D. Cook, and F.L. Lewis, "Automatically generating plans for manufacturing," J. Intelligent Systems, vol. 10, no. 3, pp. 279-319, 2000.

9.N. Swamy and F.L. Lewis, “Routing algorithms in a novel hierarchical mesh network,” Proc. IEEE 12th Symp. Mobile Computing, Bangalore, Nov. 2003.

10.A. Tiwari, F.L. Lewis, and S.S. Ge, “Wireless Sensor Networks for Machine Condition Based Monitoring,” Proc. Int. Conf. Control, Automation, Robotics, and Vision, pp. 461-467, invited paper, Kunming, China, Dec 2004.

11.V. Giordano, F.L. Lewis, J. Mireles, B. Turchiano, “Coordination control policy for mobile sensor networks with shared heterogeneous resources,” Proc. Int. Conf. Control & Automation, pp. 191-196, Budapest, June, 2005.

12.V. Giordano, F.L. Lewis, B. Turchaino, P. Ballal, V. Yeshala, “Matrix computational framework for discrete event control of wireless sensor networks with some mobile agents,” Proc. Mediterranean Conf. Control & Automation, Limassol, Cyprus, June 2005. This paper won an award at MED 05.

13.O. Kuljaca, N. Swamy, J. Gadewadikar, F.L. Lewis, “Transfer Function Illustration With Simple Electronic Circuits”, Proc. XXVII Int. Meeting MIPRO 2005, CE, Conference on Computers in Education, 2005.

14.D.O. Popa, K. Sreenath, and F.L. Lewis, “Robotic deployment for environmental sampling applications,” Proc. Int. Conf. Control and Applics., pp. 197-202, Budapest, June 2005.

Automation & Robotics Research Institute The University of Texas at Arlington

F.L. LewisMoncrief-O’Donnell Endowed Chair

Head, Controls & Sensors Group

Talk available online athttp://ARRI.uta.edu/acs

Wireless Sensor Networks

Documents

Automation & Robotics Research Institute The University of