Wireless Network Embedded Systems: Sensor Nets …...Systems: Sensor Nets and Beyond Edited and...

1

Chess ReviewMay 11, 2005Berkeley, CA

Wireless Network Embedded Systems:Sensor Nets and Beyond

Edited and Presented byShankar SastryDept of EECS,UC Berkeley

Review May 11th, 2005 2

Moore’s Law – 2x stuff per 1-2 yr

2


Bell’s Law –new computer class per 10 years

year

log

(peo

ple

per c

ompu

ter)

streaming informationto/from physical world

Number CrunchingData Storage

productivityinteractive

• Enabled by technological opportunities

• Smaller, more numerous and more intimately connected

• Ushers in a new kind of application

• Ultimately used in many ways not previously imagined


Building Comfort,Smart Alarms

Great Duck Island

Elder Care

Fire Response

Factories

Wind ResponseOf Golden Gate Bridge

Vineyards

Redwoods

Instrumenting the world

Soil monitoring

3


The Sensor Network Challenge

applications

service

network

system

architecture

data mgmt

Monitoring & Managing Spaces and Things

technology

MEMSsensing Power

Comm. uRobotsactuate

Miniature, low-power connections to the physical world

ProcStore


Miniaturization – Pister (SmartDust)

• The Goal: Autonomous millimeter-scale robots

– Sensing– Computation– Communication– Power– Motors– Mechanisms

SENSORS ADC FSMRECEIVER

TRANSMITTER

SOLAR POWER1V 1V 1V 2V3-8V

PHOTO 8-bits

375 kbps

175 bps

1-2V

OPTICAL IN

OPTICAL OUT

Solar Cell Array CCR

CMOSIC

XL

4


Low Power RF – Rabaey (PicoRadio)

• CMOS– Cheap, Integrated

• mW -> sub mW• Simple• Advantage in

Numbers

RF Filter LNA

fclock

RF Filter EnvDet ∫

fclock


RF Filter LNA

fclock


fclock


RXOn: 3 mWOff: 0 mW

RXOn: 3 mWOff: 0 mW

Receiver

RF Amp Test

LNATest

Diff Osc

PA TestTX1

TX2Env DetTest

Passive Test Structures4 m

m

MatchingNetwork

MOD1

MOD2

OSC1

OSC2 Preamp PAMatchingNetwork

MOD1

MOD2

OSC1

OSC2 Preamp PATX

On: 4 mWStby: 1 mWOff: 0 mW

TXOn: 4 mW

Stby: 1 mWOff: 0 mW

BWRC


System/Networking/Prog. – Culler (TinyOS)

Small microcontroller

- 8 kb code, - 512 B data

Simple, low-power radio

- 10 kbEEPROM (32 KB)Simple sensors

WeC 99“Smart Rock”

Mica 1/02

NEST open exp. platform

128 KB code, 4 KB data

50 KB radio

512 KB Flash

comm accelerators

- DARPA NEST

Dot 9/01

Demonstrate scale

- Intel

Rene 11/00

Designed for experimentation

-sensor boards

-power boards

DARPA SENSIT, Expeditions

TinyOS www.tinyos.net

Networking

Services

Crossbow

5


Traditional Systems

• Well established layers of abstractions

• Strict boundaries• Ample resources• Independent Applications at endpoints communicate pt-pt through routers

• Well attended

User

System

Physical LayerData LinkNetwork

TransportNetwork Stack

Threads

Address Space

Drivers

Files

Application

Application

Routers


by comparison ...

• Highly Constrained resources– processing, storage, bandwidth, power

• Applications spread over many small nodes– self-organizing Collectives – highly integrated with changing environment and

network– communication is fundamental

• Concurrency intensive in bursts– streams of sensor data and

network traffic• Robust

– inaccessible, critical operation

• Unclear where the boundaries belong

– even HW/SW will move

=> Provide a framework for:

• Resource-constrained concurrency

• Defining boundaries

• Appl’n-specific processing and power management

allow abstractions to emerge

6


Vast Networks of Tiny Devices• Past 25 years of internet technology built up

around powerful dedicated devices that are carefully configured and very stable

– local high-power wireless subnets at the edges– 1-1 communication between named computers

• Here, ...

• every little node is potentially a router• work together in application specific ways• collections of data defined by attributes• connectivity is highly variable• must self-organize to manage topology, routing,

etc• and for power savings, radios may be off 99%

of the time


Mote Evolution

7


Berkeley Open Experimental PlatformKey Areas of Progress

• OEP3 Hardware– Telos 802.25.4 mote (& MicaZ)– Mica2 testbed

• TinyOS Advances– Cross-platform 802.15.4 support– EXScal XSM support– Structured Release– BMAC– Large Scale Simulation

• Embedded Networking– Reliable, low-power data collection– Dissemination: trickle and drip– Reliable Bulk Communication– Scheduled Communication

• Network Programming– Deluge (really works at Scale)– Towards a tight lower bound– Incremental Updates

• MacroProgramming– Mate II application specific VM

• Security– 802.15.4

• Localization– Robust methodology - ultrasound– Making RSSI actually work

• Long Lived Applications– Analysis of GDI– Calibration in the redwoods

• Capstone Demo– Management infrastructure– Telos/XSM integration– Complete Simulation– Multiobject Tracking


OEP3 Platform

• Focused on low power • Sleep - Majority of the time

– Telos: 2.4µA– MicaZ: 30µA

• Wakeup– As quickly as possible to

process and return to sleep– Telos: 290ns typical, 6µs

max– MicaZ: 60µs max internal

oscillator, 4ms external• Process

– Get your work done and get back to sleep

– Telos: 4MHz 16-bit– MicaZ: 8MHz 8-bit

• TI MSP430– Ultra low power

» 1.6µA sleep» 460µA active» 1.8V operation

• Standards Based– IEEE 802.15.4, USB

• IEEE 802.15.4– CC2420 radio– 250kbps– 2.4GHz ISM band

• TinyOS support– New suite of radio stacks– Pushing hardware abstraction– Must conform to std link

• Ease of development and Test– Program over USB– Std connector header

• Interoperability– Telos / MicaZ / ChipCon dev

• 97% Yield on first 200– Remaining reparable

UCB Telos

Xbow MicaZ

8


Deluge Network Programming

• Reliable Pipelined Epidemic Distribution of series of pages

– Constrained storage hierarchy– Robust to lossy or asymmetric links– Very low maintenance bandwidth

• Page Advertise, Request/Fix, Xfer– Density-aware suppression and snoop on each

• Packet CRC + Page CRC• 159 Byte memory footprint• Packed image (no 64k xnp limit)• Multiple Program images• Extensive simulation of dissemination Alg.

and many variants• Tested extensively on 77 nodes• Simulated on >1,000 nodes in EXSCAL

configuration with multiple sources• 4 mins for 1-hop 33k image to 33 nodes• Command line host tools

flash

…

Maintain

Request

Transmit


Simulated Dissemination (10x10)

• Smooth pipelined wavefront of multiple pages

announcement

requests

9


Simulation at Larger Scale (16x16)

Wavefront propagates more rapidly along the edges due to lower density

- long linear structures are fast

Behavior exacerbated by conservative interference model in TOSsim


Localization Methodology

1. Environmental characterization – Measure accuracy and loss rate of many

pairs and range of distances (pseudo-random pattern)

– 45 environments empirically characterized– 660 empirical measurements taken within

.3m of every distance between 0-30 m2. System design

• Design filters, schedule, model, localization alg. against observed error and loss distribution

• Select simplest that works3. Pre-deployment analysis

• Determine a adequate system to deploy• Trace-based simulation make realistic

predictions• Use data from 1.

• Model-based simulation makes optimistic predictions

−200 0 200 400 600 800 1000 1200−200

0

200

400

600

800

1000

1200

1

2

3

4

5

6

7

8

9 10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

Localization with Empirical and Model−based Ranging

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Data Collection Topology

Distances are in meters

A

B

C

D

E

F

G

H

I

J

K

L

M

N

O

P Q

R

S

T

U

V

W

X

Y

0 15 30 45 61 76 91 106

0

200

400

600

800

1000

1200

1400

1600

1800

2000

RSSI (mV)

Tru

e D

ista

nce

(cm

)

Fit of Model to Raw Data

raw datalinear modeltheoretical model

Lower Signal Strength

30 35 40 45 50 55 600

50

100

150

200

250

300

350

400

450

500

Node Density (#nodes)

Err

or (

cm)

Mean FilterMedianTube FilterConsistency CheckingTheoretical

−50 0 50 100 150 200 25010

20

30

40

50

60

70Effect of Uniformity

Stdev on Placement of Nodes (cm)

Err

or (

cm)

Consistency CheckingTheoretical

10


Ultrasound Results

• Non-gaussian error• Probabilistic loss• 50cm median error in deployment,

as predicted– 49 node deployment– 8 hop network

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1Probability of Ranging Success

Distance (m)

Pro

babi

lity

grassfree spacecarpetpavement

0 50 100 150 200 2500

0.02

0.04Empirical Error Distribution (kernel smoothing)

Error (cm)

Pro

babi

lity

200 400 600 800 1000 1200 1400 1600200

400

600

800

1000

1200

1400

1600

Dis

tanc

es in

Cen

timet

ers

31

21

1

9

3

26

20

39

4

2

47

28

19

29

33

38

30

35

11

44

40

23

36

10

22

46

32

14

7

5

48

15

34

18

16

41

24

27

8

12 6

25

45

43

13

37

42

17

49

−0.5 0 0.5 1 1.5 2

0.001

0.003

0.01 0.02

0.05

0.10

0.25

0.50

0.75

0.90

0.95

0.98 0.99

0.997

0.999

Ranging Error (m)

Pro

babi

lity

Normal Probability Plot of Ranging Error


Signal Strength Results

0 5 10 15 20 25 30 35 40 45 50

0

5

10

15

20

25

30

35

40

45

50

24

31

8

42

21

25

15

27

30

39

26

2

29

49

22

44

16

19

36

48

10

33 7

35

4

46 38

28

3

11

40

43

47

14

32

6

12

23

9

45

41

37

5

20

18

13

34

17

1

X Axis (m)

GPS errors on a 49 node network

Y A

xis

(m)

0 5 10 15 20 25 30 35 40 45 50

0

5

10

15

20

25

30

35

40

45

50

24

31

8

42

21

25

15

27

30

39

26

2

29

49

22

44

16

19

36

48

10

33 7

35

4

46 38

28

3

11

40

43

47

14

32

6

12

23

9

45

41

37

5

20

18

13

34

17

1

X Axis (m)

Y A

xis

(m)

RSS errors on a 49 node network

1.9m / 7.2m

4.1m / 8.9m

0 5 10 15 20 25 30 35

0

5

10

15

20

25

30

35

6

1

2

19

4

23

20

18

8

12

21

5

25

11

15

17

3

7

22

16

14

24

10

9

13

X Axis (m)

Y A

xis

(m)

RSS errors on a 25 node network

0 5 10 15 20 25 30 35−5

0

5

10

15

20

25

30

35

X Axis (m)

Y A

xis

(m)

GPS errors on 25 node network

23

7

20

24

14

25

1

21

15

18

9

6

3

2

10

13

5

11

22

8

4

19

17

12

16

4.2m / 12.6m

6.1m / 10.1m

•Results vary greatly with each environment•Pre-deployment analysis accurately predicted effects ofenvironment, topology, and algorithmic parameters

•RSSI-based accuracy can be comparable to GPS

GPS RSSI

11


Vanderbilt Radio Interferometric RangingChampion to Date

• COTS radio chip (CC1000 on MICA2)– transmit frequency: 400-460 MHz– wave length: 65 cm < λ < 75 cm– adjustable in 64 Hz steps

• Two senders (A and B) transmit simultaneously

– frequency separation: 100-800 Hz– duration of transmission: 32 ms

• Several receivers (C, D and E) measure interference

– sample radio signal strength at 8.9 kHz– beat frequency: 100-800 Hz– samples per beat: 10-80– beats per transmission: 3-25– use time synchronization with 1 µs

precision to correlate phase offsets– result is (dAD-dBD+dBC-dAC ) modulo λ

» dXY is distance between X and Y » λ is wave length of carrier frequency

– expected error is less than 5 cm• Perform multiple measurements with

different frequencies to obtain dAD-dBD+dBC-dAC

© 2005 Vanderbilt University

A B

C D

EdACdBC

dADdBD

relative phase offset of beat frequency= (dAD-dBD+dBC-dAC ) modulo λ

where 65cm < λ < 75cmAkos Ledeczi, Miklos Maroti, et al


Technology Push matched by Application Pull

12


Structural Monitoring – Glaser, Fenves

25 Motes onDamaged sidewall

30 Motes onGlue-lam beam

0 2 4 6 8 10 12 14 16 18

-1

-0.5

0

0.5

1

Low resolution Sensor, Test4, Increasing frequency

Time (sec)

Acc

eler

atio

n (g

)

Wind ResponseOf Golden Gate Bridge

• Dense Instrumentation of Full Structure– Cost is all in the wires

• Leads to in situ monitoring• Self-inspection and Diagnosis

Liquifaction, Tokashi Port


Forest Ecophysiology - Dawson

•• How TREES shape How TREES shape the hydrological the hydrological cycle?cycle?– 2/3 of fresh H2O recycled through forests

• Microclimatic Drivers of Plant Dynamics

• Influence climate2003,

unpublished

Temperature vs. Time

8

13

18

23

28

33

7/7/039:40

7/7/0313:41

7/7/0317:43

7/7/0321:45

8/7/031:47

8/7/035:49

8/7/039:51

8/7/0313:53

8/7/0317:55

8/7/0321:57

9/7/031:59

9/7/036:01

9/7/0310:03

Date

Humidity vs. Time

35

45

55

65

75

85

95

Rel H

um

idity (

%)

101 104 109 110 111

Bottom Top36m

34m30m

20m

10m

13


Built Environments - Arens

• 2/3 of US energy used to maintain our bldg environment

– 40% lighting• Inefficient,

unhealthy, uncomfortable due to lack of sensing and control

Lighting, temperature, sound, air quality…

Electricity, gas, water, weather…

Occupancy, comfort, productivity

Building

EnergyPeople

Light ballast

VAV actuator

Reflective vane actuator

Occupancy sensor

Desk climate sensor

Window switch

Climate sensor

Base station

BACnet

Comfort stat


Shooter Localization Using SensorWebs

Berkeley motes and Vanderbilt algorithm

14


And many more

• Sitar - Firebug• Agogino - Occupancy• Varaiya – Traffic• Wright – emergency

personnel• Brewer – ITC4B• Student solar car• Habitat Monitoring• Intel fab app

Condition-BasedMaintenance

Intel Research

Great Duck Island

< 100 m

Intersection

mote13

mote14

receiver


Protection – Sastry, Culler, Brewer, Wagner

• Detect vehicle entering sensitive area, track using magnetics, pursue and capture by UGV.

• Components– 10x10 array of robust wireless, self-localizing

sensors over 400 m2 area– Low cost, robust ‘mote’ device– Evader: human controlled Rover– Pursuer: autonomous rover with mote, embedded

PC, GPS• Operation

– Nodes inter-range (Ultrasonic) and self localize from few anchors, correct for earth mag, go into low-power ‘sentry’ state

– Detect entry and track evader» Local mag signal processing determines event and

announces to neighbors» Neighborhood aggregates and estimates position» Network routes estimate from leader to tracker

(multihop)– Pursuer enters and navigates to intercede

» Motes detect and estimate multiple events» Route to mobile Pursuer node» Disambiguates events to form map» Closed inner-loop navigation control» Closed information-driven pursuit control» Capture when within one meter

dot

mag ultrasound

acoustic

ultrasoundmain

mag sensepower

battery

reflector

pursuer

evader

15


Overall Demo Operation (2):Object Detection

• Evader enters field• Pursuers enter field• Each detecting mote

announces its readings• All readings in a

neighborhood are aggregated

• Aggregate packet is routed through the root mote


OEP2 Hardware Platform

• Main board: Mica2 dot – ATmega microcontroller, Flash, clock– CC1000 frequency agile FSK radio– Small form-factor– Xbow based on Mica

• Sensor Board: magnetometer– Honeywell mag with 2-stage amplification– Set/reset circuit (5v)– 4-port digital Pot for biasing and filtering

• Ranging boards– Ultrasound– Acoustic

• Enclosure

dot

mag ultrasound

acoustic

16


PEG Software Overview

• New routing protocol to relax dependency on localization service

• Remote configuration interface– Solution to a problem, not an original goal

• Network reprogramming• System layer for remotely invoking disparate services• Standard services

– Sleep, ping, RF power, blink, network reprogram• MATLAB command-line interface to network• Strong decoupling between sensor networks and clients


PEG Demo from July 03

17


Mote Drop Experiences:Last 2 of 6 motes are dropped from MAV 29

Palms March 2001


Bald Camel Drop: China Lake 02/19/04

MIR integrated with Mica-2 Packaging to allow antennas to land pointing up

Explosive DispenserAerial Drop from 500 ft.

18


Bald Camel Drop Feb 2004

-80

-60

-40

-20

0

20

40

60

80

0 50 100 150 200 250 300 350 400

Down Range Dispersion [meters]

Clockwise from top left: Actual Sensor Lay Down Pattern, Phoebus Chen walking through sensor field, Tracking of Phoebus through sensor field


Big July 2005 Demo Philosophy

• Midterm demos focused on building real applications

– “Will this work at all?”• ExScal focuses on scaling real applications

– “How big can it get?”• Capstone will focus on evaluating real

applications– “Can we describe, characterize, and predict how well

it works?” Metrics, Metrics, Metrics– This is our opportunity and need to leverage

experience toward good science– Deployment over a period of time (2-3 weeks) at

Richmond Field Station not just in field but spread out over the station

19


Capstone Goals

• Establish metrics to evaluate PEG– Metrics vary at each abstraction

level– Pursuer Algorithm cares about:

» Sensor Accuracy, Latency, Missing Data, False Alarm Rate, etc.

– “Quality” of data stream from the Base Station depends on:

» Network Congestion, Node Density, Power Consumption Constraints, Physical Distance, etc.

• Simulations• Experiments• Characterize and predict

SensorNetwork

(Mobile)Base Station

Pursuer Algorithm


Towards Final Demo

• Redeployed substantial sensor field to collection data and develop metrics

• Worked with EXSCAL design to ensure can reuse XSM as sensor platform for final demo– Attach Telos as standalone or as bridge

• Replace/ augment acoustic localization with RSSI

• Airborne pursuer, airborne drop simulated• Multiple evaders• Focus on middleware architecture, science,

measurement, ease of deployment, and reliability– Developed a baseline management architecture

20


Script for Final Demo: Aug 2005

• 1000+ XSM + Telos motes spread out over urban area (Richmond Field Station: area approximately 3 sq. km. Including several buildings, varying topography, trees, …). Experiment will be up and running for 2 weeks: fully instrumented.

• Base line demo components developed by UCB team. Technology Developers can swap in their components: localization, time synchronization, tracking, network management, network programming.

• Scenario: Evaders are humans/robots. Pursuers are humans/robots: some aerial and some ground robots. Distinguish between red forces, blue forces and green forces in tracking algorithms to minimize capture time or localization

• UAVs with motes used to localize sensor field• Sensor fields to cue higher bandwidth sensors (cameras)• Metrics: time to capture, energy expenditure, ease of

re-programmability


Features of Final Demo

• Pursuit evasion games with red, blue and green forces, a combination of pre-deployed sensor webs, MAV dropped sensor webs, MAV/UGV camera platforms and other higher bandwidth assets

• Persistence/Reprogrammability• Network Management• MOBILE Sensor Webs• Determination of Pursuit and Evasion policies and tactics

21


Implementation in an Instrumented Testbed

• 9 x 5 grid of 45 motes

• 10 ft by 20 ft area, 2.5 ft spacing

• Ethernet backend for monitoring/data collection

• Multiple robots running underneath mesh of sensors

• Camera for “ground truth” position of robots


Soda Hall Routing Environment

22


Conceptual Issues

• Given a certain WSN, can we successfully design a particular application?

• How does the application impose constraints on the network?

• Can we derive important metrics from those constraints?

• How do we measure network parameters?

We need analytical tools and experimental data


What can you do with a sensor network?

• Literature provides key asymptotic results

• We are interested in answer different semantic questions, e.g.:

• At the algorithmic level:– How much packet loss can a tracking

algorithm tolerate?• At the network level:

– How many objects can a particular sensor network reliably track?

23


Expanding the Vision


Legal / Privacy – Samuelson, Mulligan

• TOWARD A LEGAL FRAMEWORK FOR SENSOR NETWORKS

– existing laws?– new laws or regulations

• Modes of Regulation and Policy• Evolving notions of Privacy• FAIR INFORMATION PRACTICES

– Limitations on collection of data – specific purpose – Notice and choice– Verifiability– Accuracy, Security– Accountability

24


Small Technology, Broad Agenda, Unique Confluence

• Social factors– security, privacy, information sharing

• Applications– long lived, self-maintaining, dense instrumentation of

previously unobservable phenomena– interacting with a computational environment

• Programming the Ensemble– describe global behavior, synthesis local rules that have

correct, predictable global behavior• Distributed services

– localization, time synchronization, resilient aggregation• Networking

– self-organizing multihop, resilient, energy efficient routing– despite limited storage and tremendous noise

• Operating system– extensive resource-constrained concurrency, modularity– framework for defining boundaries

• Architecture– rich interfaces and simple primitives allowing cross-layer

optimizationl ADC di i ti ti


Secure SCADA and beyond

We think that there is a great deal to be done in terms of operationalizing secure versions of SCADA (Supervisory Control And Data Acquisition) and DCS (Digital Control Systems) for the infrastructures considered, especially power, natural gas, chemical and process control, etc. However, the sense was that this infrastructure was going to be gradually replaced by networked embedded devices (possibly wireless) as computing and communication devices become more ubiquitous and prevalent. Thus, the major research recommendations were for an area that we named Secure Networked Embedded Systems (SENSE).

25


SCADA of the Future

• Current SCADA– Closed systems, limited coordination, unprotected

cyber-infrastructure– Local, limited adaptation (parametric), manual control– Static, centralized structure

• Future requirements – Decentralized, secure open systems (peer-to-peer,

mutable hierarchies of operation)– Direct support for coordinated control, authority

restriction– Trusted, automated reconfiguration

» Isolate drop-outs, limit cascading failure, manage regions under attack

» Enable re-entry upon recovery to normal operation» Coordinate degraded, recovery modes


Embedded Software prevalent in all critical infrastructures. Critical to high confidence embedded software are open source techniques for

• Automated Design, Verification and Validation– Verified design in a formal, mathematical sense– Validated design in an engineering sense– Certifiable design to allow for regulatory and certification input

• High Confidence Systems– Narrow waisted middleware

» Trusted abstractions, limited interfaces» Algorithms and protocols for secure, distributed coordination and

control– Security and composable operating systems– Tamper Proof Software

• Generative Programming• Intelligent Microsystems: infrastructure of the future with

security codesign with hardware and software.

Secure Network Embedded Systems

26


Adaptive Networked InfrastructureCore partners: Berkeley (lead), Cornell, Vanderbilt

Outreach partners: San Jose State, Smith, Tennessee Tech, UC Davis, UC Merced.Principal investigator: Edward A. Lee, Professor, EECS, UC Berkeley, [email protected]

•Enabling technologies: wireless networked embedded systems with sensors and actuators

• Deliverables: Engineering Methods, Models, and Toolkits for: • design and analysis of systems with embedded computing• computation integrated with the physical world• analysis of control dynamics with software and network

behavior• programming the ensemble, not the computer• computer integrated systems oriented engineering curricula

•Approach: Engineering methods for integrating computer-controlled, networked sensors and actuators in societal-scale infrastructure systems.

•Resource management test beds:• electric power• transportation• water

•Target: efficient, robust, scalable adaptive networked

infrastructure.

The ANI ERC


Strategic Framewor

k

Wireless Network Embedded Systems: Sensor Nets …...Systems: Sensor Nets and Beyond Edited and...

Documents

Transcript of Wireless Network Embedded Systems: Sensor Nets …...Systems: Sensor Nets and Beyond Edited and...