wireless sensor networks and applications

21
May18, 2004 General Dynamics Meeting 1 Wireless Sensor Networks and Applications Cauligi S. Raghavendra Department of Electrical Engineering University of Southern California Los Angeles, CA 90089 [email protected] http://ceng.usc.edu/~raghu

Upload: jillian-hardin

Post on 03-Jan-2016

39 views

Category:

Documents


1 download

DESCRIPTION

Wireless Sensor Networks and Applications. Cauligi S. Raghavendra Department of Electrical Engineering University of Southern California Los Angeles, CA 90089 [email protected] http://ceng.usc.edu/~raghu. Wireless Sensor Networks. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 1

Wireless Sensor Networks and Applications

Cauligi S. Raghavendra

Department of Electrical Engineering

University of Southern California

Los Angeles, CA 90089

[email protected]

http://ceng.usc.edu/~raghu

Page 2: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 2

Wireless Sensor Networks

• Wireless Sensor Networks is one of the top 10 Technologies that will change the World in 21st Century– According to MIT Technology Review

• Researchers at USC and ISI Pioneered the field of Sensor Information Technology

• DARPA and NSF have Programs and Initiatives in Sensor Networks

Page 3: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 3

Centers, Activities, and Projects

• USC is a Partner in UCLA’s Center for Embedded Networked Sensing (CENS)– NSF Science and Technology Center– Director: Prof. Deborah Estrin, UCLA

• USC SoE Sensing Systems Cluster– http://www.usc.edu/dept/engineering/research/sensornetworks.html

• USC Center for Robotics and Embedded Systems (CRES)– Director: Maja Mataric, USC

• Several Ongoing Projects in EE and CS

Page 4: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 4

Some Sensor Nodes

Modern Sensor Nodes

UC Berkeley: COTS Dust

UC Berkeley: COTS DustUC Berkeley: Smart Dust

UCLA: WINS Rockwell: WINS JPL: Sensor Webs

Page 5: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 5

Characteristics of Sensor Networks

• Sensor Nodes – Limited Battery Life

• Capacity – CPU, Memory

• Limited Software Support

• Ad Hoc Networks with Low Bandwidth

• No Network Infrastructure

• Link Quality – Fading and Interference

• Deployment – Less Control

Page 6: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 6

Research Challenges

• Interdisciplinary Research • Application Aware Research• New Networking Paradigms and Protocols• Self Organization and Localization• Incomplete and Inaccurate Field Data• Energy Efficient Algorithms and Protocols• Embedded Environments and Deployment

Page 7: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 7

Research in School of Engineering

• EE, CSCI, CE, BME Departments, and ISI

• Profs. Helmy, Krishnamachari, Kumar, Lee, Mitra, Mendel, Narayanan, Ortega, Prasanna, Raghavendra

• Profs. Govindan, Sukhatme, Requicha• Profs. J. Caffrey, E. Johnson, S. Masri• Dr. Heidemann and others at ISI

Page 8: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 8

Sensor Nets for Search and Rescue

Inactive Sensor

Page 9: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 9

Sensor Nets for Search and Rescue

Page 10: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 10

Sensor Nets for Search and Rescue

Active Sensor

Page 11: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 11

Situation Awareness

Periodic Voice Traffic

Internet Connectivityto Mission HQ

Squad

Energy Efficient one-to-all and All-to-all Broadcasting Algorithms

100-300% Improvement for Situation Awareness in Ad hoc Networks

Energy Efficient one-to-all and All-to-all Broadcasting Algorithms

100-300% Improvement for Situation Awareness in Ad hoc Networks

Page 12: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 12

Power Aware Computing and Commn.

Multi-Level Power Management in Distributed Battlesite/Sensor Network (PAC/C Program)

Task

Algorithm

Protocol

Physical

Page 13: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 13

Multiple nodes Sensedata and Coordinate

Multiple nodes Sensedata and Coordinate

Distributed Computationin Sensor Networks

Application:Target Detection

Application:Target Detection

•Power Aware Node Selection•Task Allocation•Collaboration•Signal Processing•Wireless Communication

•Power Aware Node Selection•Task Allocation•Collaboration•Signal Processing•Wireless Communication

Non-participating node

Selected node

Selection / Query

T1

T4

T3

T2

T5

Page 14: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 14

Acoustic Sensors

• Sensor database provided by the Army Research Laboratory (acoustic and seismic)

• Microphone arrays are typically 4 ft – 8 ft in diameter, not restricted to a specific geometry

• Inexpensive, passive and non-line of sight capabilities

Acoustic Sensor Array - RNADS

Courtesy of N. Srour, Army Research Lab

Page 15: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 15

Spatio-Temporal Correlation in Sensor Data

Compress data to reduce storage and communication bandwidth!

Page 16: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 16

A Distributed Algorithm for Waking-up in Heterogeneous Sensor Networks

2

1

3

5

4

{1,5}

{1}

{1}{1 ,2}

{1}

{1 ,5}

{4 ,5}

{3 ,4}

{3 ,4}

{2 ,3}

{2 ,3}

{i1,i2,...,iK}: set of chosen trackersTracker Off

Tripw ire Detect

Tripw ire Sense

Detection Area

2

1

3

5

4

{1,2}

{1 ,2}{1 ,2}

{1 ,2}

{1 ,2}

{1 ,2}

{1 ,2}

{1 ,2}

{1 ,2}

{1 ,2}

i Tracker On"w ake-up"

{1 ,2}

Event occurs

N tripwires detect event

Detecting tripwires run distributed algorithm => choose K optimal trackers

wake-up chosen trackers

6

15

2536

4964

0

50

100

150

200

250

total message

s

M

N

Message Complexity

200-250

150-200

100-150

50-100

0-50

Performance under link losses

05

10152025303540

0 5 10 15 20 25 30 35 40 45 50

link loss probability (%)

no

n-c

on

verg

ing

no

des

(%)

Convergence delay vs. link loss

0

20

40

60

80

100

120

0 5 10 15 20 25 30 35 40 45 50

link loss probability (%)

dela

y (

ms)

RESULTS

Page 17: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 17

False Alarm Detection in Wireless Sensor Network

Node Data

0

1000

2000

3000

1 186 371 556 741 926

Milliseconds

Am

plit

ud

es

False Alarm Data

1400

1500

1600

1700

1 184 367 550 733 916

Milliseconds

Am

plit

ud

es

False Alarm Data

-2500

-5001500

35005500

1 201 401 601 801 1001

Milliseconds

Am

plit

udes

Node Data

-2500-500150035005500

1 201 401 601 801 1001

Milliseconds

Am

plit

ud

es

Energy in subbands are evenly distributed (no variation)

Binary subband search

Linear search is slow

• In search iteration, pick the subband with larger energy and compare with a threshold

• Negative if greater than the threshold

• Report false alarm in the end

• It is possible detected signal is due to noise• Processing such noise data leads to unnecessary use of resources• False alarm detection help to save energy significantly

Page 18: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 18

Traffic shaping to mitigate Hotspots

Cluster based Traffic Shaping• Cluster

– Scheduled data collection within cluster• Data Combination (DC)

– Packet size vs. Packet number• Rate Adaptation (RA)

– Adapt rate according to latency bound• Traffic Dispersal (TD)

– Spread data traffic away from hot spot

2 3 4 5 6 7 8 9 100

50

100

150

200

250

Sensor Message Arrival Interval (second)

Ene

rgy

Eff

icie

ncy

(Pac

ket/

Joul

e)

CSMARARA+DCRA+DC+TD

2 3 4 5 6 7 8 9 100

0.2

0.4

0.6

0.8

1

Sensor Messsage Arrival Interval (second)

Del

iver

y R

atio

CSMARARA+DCRA+DC+TD

Page 19: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 19

Structural Health Monitoring

• Goal: Design sensor networks for improving the safety of structures (buildings, bridges, ships, aircraft, spacecraft)

• Research focuses:– Local excitation based

damage identification– System components for fine-

grain structural monitoring

• Multi-disciplinary effort:– John Caffrey (CE),

Ramesh Govindan (CS),Erik Johnson (CE),Bhaskar Krishnamachari (EE),Sami Masri (CE),Gaurav Sukhatme (CS)

Page 20: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 20

Courses and Conferences

• ACM Wireless Sensor Networks and Applications Workshop• IEEE SECON Conf. in October 2004• Edited Book on Wireless Sensor Networks

• Pertinent Courses Offered:– Advanced Topics in Computer Networks and Distributed

Systems (Prof. Ramesh Govindan, CS 694)– Wireless Sensor Networks (Prof. Bhaskar Krishnamachari, EE

652)– Intelligent Embedded Systems (Prof. Gaurav Sukhatme, CS

546)– Wireless Networking, Design and Analysis Laboratory (Prof.

Ahmed Helmy, EE 599)

Page 21: Wireless Sensor Networks and Applications

May18, 2004 General Dynamics Meeting 21

Faculty Contact Information

• A. Helmy, [email protected], (213) 821-1329• B. Krishnamachari, [email protected], (213) 821-2528• V. Kumar, [email protected], (213) 740-4668• D. Lee, [email protected], (213) 740-0882• U. Mitra, [email protected], (213) 740-4667• J. Mendel, [email protected], (213) 740-4445• S. Narayanan, [email protected], (213) 740-6432• A. Ortega, [email protected], (213)740-• V. Prasanna, [email protected], (213) 740-4483• C. Raghavendra, [email protected], (213) 740-9133• R.Govindan, [email protected], (213) 740-4509• G. Sukhatme, [email protected], (213) 740-0218u• A. Requicha, [email protected], (213) 740-4502• J. Caffrey, (213) 740-0603• E. Johnson, [email protected], (213) 740-0610• S. Masri, [email protected], (213) 740-0602• J. Heidemann, [email protected], (310) 448-8708