Signals, Instruments, and Systems – W11Environmental Monitoring –Sensor Nodes and Networks

in Real Deployments

1

Outline• Static sensor networks for

microclimate monitoring: the Sensorscope project– Motivation– System design– Practical issues

• Mobile sensor networks for air quality monitoring: the OpenSense project– Goals and research questions– System design– Lausanne and Zürich deployments

2

Motivation for Sensor Networks

What if we could monitor events which …

– have a large spatial and temporal distribution– require in-situ measurements– take place in hard to access places– generate data which need to be available in

real-time

3

Motivation for Sensor Networks

What would we need for that?A device which …

– is cheap – so we can distribute many of it – is reliable – so we can measure for a long time– uses little power – battery/solar cell powered– has a radio – so it can communicate– can potentially move – so it can potentially

relocate

4

Building a SensorNetwork: Key Concepts

5

Introduction

Temperature

Humidity

Light

6

Topology

?

7

Topology

8

Topology

9

Topology

GPRS

10

Topology

Pros• Very simple!• No restrictions in sensor locations

Cons• The server may be quite far from the stations• A long-range link may consume a lot of energy!

11

Topology (TinyNode Example)

0

200

400

600

800

Sensor MSP430 XE1205 GPRS

Power

Consum

ption[m

A]

0.5 350

700!

14xtim

estheXE

120 5

!

12

Topology (SensorScope Example)

0

200

400

600

800

Sensor MSP430 XE1205 GPRS

Power

Consum

ption[m

A]

0.5 350

700!

14xtim

estheXE

120 5

!

13

Topology

Assuming four AA batteries, 1.2 V, 2000 mAh

• Sensor: 167 days• MSP430: 28 days• Short range radio: 1.7 days• Long range radio: 8 hours

14

Topology

15

But makes sense on projects such OpenSense!

Topology

GPRS

Short range

Sink

16

TopologyRecall Friis law (week 10):

Example: To transmit over 5 Km we can using 868 MHz we can:

• One hop of 5 km: L = 106 dB• Two hops of 2.5 km: L = 99 dB• Five hops of 1 km: L = 92 dB

Energy is the main issue !!!

17

Multi-hop Sensor Network

GPRS

18

Multi-hop Sensor Network

Pros• Only one car battery in the network• The sensor network has extended monitoring coverage• Multiple routes for stations to communicate with the sink• Auto configurable network (robustness)

Cons• Significantly more complicated• Data rate is not increased• Unable to use directional antennas

19

Field Deployment of Static Sensor Networks

–The SensorScope Project

20

Microcontrollers are inside hermitically sealed boxes, attached on a mounting pole with up to seven external sensors.

Price: 4000-5000 CHF

SensorScope

21

SensorScope

22

Shockfish TinyNode with TinyOS 2.xMSP430 16-bit microcontroller @ 8MHz48KB ROM, 10KB RAM, and 512KB flash memorySemtech XE1205 radio transceiver @ 868MHz, 76Kbps

SensorScope

23

162x140mm solar panel12Ah NiMH rechargeable battery

SensorScope

24

25

SensorScopeMany previous successful deployments

97 stations deployed at EPFL (one year)

26

SensorScopeMany previous successful deployments

16 stations deployed at Le Génépi to monitor conditions leading to dangerous mudslides (two months)

27

Field Deployment of Mobile Sensor Networks:

–The OpenSense Project

Mobile sensors (parasitic, uncontrolled mobility) on public transportation vehicles

Static wireless sensing and communication infrastructure

OpenSenseCommunity-driven, large-scale air pollution

measurement in urban environments

Sensorscope

Permasense

• 2% of all deaths (1.2 million people)• 0.6% of burden of disease (DALY)

Urban air pollution

Global Health Risks, WHO 2009World Urbanization Prospects, U. N. 2008

Urban population will double in next decades

Motivation

• > 50% of world population already lives in cities• rural population expected to stagnate or drop

Fine resolution air quality data is needed!

Enabling research in:• Human exposure• Air Pollution Engineering• Urban Planning• Environmental Justice• Public Policy

Public service & education• enable private users to make

informed decisions• raising popular awareness

MotivationAir pollution is highly location-dependent• traffic chokepoints• urban canyons• industrial installations

Sparse networks of ground stationsExample: Switzerland’s NABEL (www.empa.ch/nabel)

• 16 stations

• specially selected sites

urban with traffic urban residential suburban rural, etc.

• resolution: high temporal low spatial

Mission: monitor air pollution on national level & gauge impact of environmental policies

Public data access: http://www.bafu.admin.ch/luft/luftbelastung/blick_zurueck/datenabfrage

Traditional Air Monitoring Systems

Ozone concentration

Station locations

Satellite-based remote sensingExamples:• Measurements of Pollution in the Troposphere (MOPITT on Terra satellite)• Ozone Measurement Instrument (OMI on Aura satellite)

Features:• daily scans• large coverage• homogeneous quality• sensitive to cloud coverage• low resolution

Traditional Air Monitoring Systems

• mobile sensor network• parasitic mobility: anchored to existing mobility sources

• low-cost, light-weight chemical (CO, CO2, NO2, O3) & ultrafine particle (UFP) sensors• intelligent integration & control to mitigate demanding constraints

- vehicle energy supply - predictable mobility- single point maintenance

public transport

OpenSense System

SENSING SYSTEMFrom many wireless, mobile,heterogeneous, unreliable rawmeasurements …

INFORMATION SYSTEM… to reliable, understandable and

Web-accessible real-time information

NA

NO

TE

RA

Nabel station Zürichstatic nodesmobile nodes

GPRSGPS

sensor network controloptimization of data acquisition

information dissemination

Proposed System

• Traditional approach Few stations Low resolution interpolated

estimates of pollutant concentrations across massive regions

• Recent results Massive deployment of stations (150)

at street-level (2008/2009 New York City Community Air Quality Survey)

Pollutants of interest heavily concentrated along roads with high traffic densities

Value of Dense Measurements

Global questions:• More data, more noise, but also more redundancy

Can we produce better quality data?• Case study for other environmental phenomena:

Radiation, noise, energy

Research directions:• Wireless Sensor Network control

When/Where to sample? What/To whom to transmit?

• Sensor Node design Sampling System Localization Software & hardware architecture Mechanical integration

Challenges

• Community sensing privacy protection trustworthiness of data, relevance of data gathered and

information produced• Modeling

sensor, device and mobility models air quality models privacy, trust & activity models

Gas Sampling System

Open sampling• sensors directly exposed to environmental

measurandBenefits:• simple & “slim” solution• continuous samplingDrawbacks:• no absolute concentration values• noisy signal (sensitive to environment

variations: pressure, humidity)Typical response:

Closed sampling• sensors exposed to measurand inside controlled

chamber• 3-phase strategyBenefits:• absolute measurements• noise due to environment filteredDrawbacks:• complex & bulky• non-continuous samplingTypical response:

[Lochmatter 2010] [Trincavelli 2010]

Idea: Combine these two approaches to get the benefits of both systems.

Problem:Chemical sensors have very slow dynamics (example: Telaire 6613 CO2 sensor step response <2min)

• Smart sampling module possibly hybrid single/multi-chamber wind sensing

controlled flow

uncleanair

cleanair

open

closed

passive active

[Lochmatter et al. 2010]

[Gonzalez-Jimenez et al. 2011]

Current deployment

Gas Sampling System

[Lochmatter et al. 2010]

Anemometer

Logger• GPRS link to back-end server• local storage on SD card

Robust localization – prerequisite for adaptive control• exploits commercial state of the art u-blox LEA-6R

GPS + dead reckoning (DR) module• augmentation with additional sensor modalities

GPS only

GPS + DR

Logging & Localization

logger

localization

Localizationdoors open Current stop: Sallaz

Next stop: Valmont

Next stop: Sallaz• large set of rich data:

location parameters (geographical coordinates, heading, odometer, speed, acceleration etc.)

vehicle context data

OpenSense Lausanne Node

Particle sampling module• Ultrafine particle

measurements using NaneosPartector

• Measures directly lung-deposited surface area

Gas sampling module• CO, NO2, O3, CO2,

temperature & relative humidity

• Hybrid active sniffer/closed chamber sampling operation

• Enables absolute concentration mobile measurements

Enhanced localization & logger• mounted inside bus• Fused GPS, gyro and vehicle

speedpulses• Accurate sample geolocation even in

difficult urban landscapes• GPRS communication

On top of 10 buses in Lausanne• CO, NO2, O3, CO2, UFP, temperature, humidity• Active sniffing & closed sampling system• Localization: Augmented GPS; communication: GPRS• Prototypes deployed in multiple stages since June 2011• Full deployment: Since November 2013

At NABEL roadside station in Lausanne• Calibration and sensor drift evaluation• Testing new sensors• Since June 2010

On top of C-Zero electric vehicle• 100% electric, clean platform• flexible mobility• system test bed• targeted investigation tool• intelligent network servicing

Lausanne Deployment

OpenSense Zurich Node

Inside the OpenSense Zurich node

OpenSense Zurich node

Installation on top of VBZ Cobra tram

Static OpenSense stationat NABEL station for sensor tests

Luftibus with OpenSense stationcovers whole Switzerland

10 streetcars in Zurich equippedwith OpenSense stations

At NABEL station in Dübendorf• Long-term sensor testing (e.g., O3)• Testing new sensors (combined CO/NO2)• Since April 2011

On top of 10 streetcars in Zurich• O3, CO, ultrafine particles, temperature, humidity• Localization: GPS; Communication: WLAN and GSM• Since September 2011

On top of “LuftiBus”• O3, ultrafine particles, temperature, humidity• Localization: GPS; Communication: GSM• Since March 2013, covers whole Switzerland

Zürich Deployment

CO concentration UFP concentration

Pollutant # of Measurements Sampling rate Time Period

UFP 56.000.000 5s 22 months

Ozone 8.900.000 20s 22 months

CO 5.300.000 20s 22 months

[Keller et al., SenseApp 2012]

Pollution Data – Zurich Deployment

Pollution Data – Lausanne Deployment

CO concentration UFP concentration

Pollutant # of Measurements Sampling rate Time Period

UFP 9.151.000 1s 3 months

Ozone 2.488.000 15s 5 months

[CO, NO2, CO2] 10.930.000 5s 5 months

[Arfire, unpublished, 2014]

Mobility Data – Lausanne Deployment

Coverage of Lausanne regionX-axis acceleration& vehicle context

Measurement # of Measurements Sampling rate Time Period

[GPS, gyroscope] 52.207.000 1s 5 months

[odometer, accelerometer] 211.767.000 0.25s 5 months

vehicle context info 751.000 event-driven 5 months

Gyro yaw rate

Next StopCurrent Stop

[Arfire, unpublished, 2014]

Martinoli, Thiele –Stations and mobility

Aberer, Faltings –Data, Models, Trust, Privacy

OpenSense

Krause – CrowdSensing for quakes

Emmenegger – Air quality measurement and modeling

1E5

2E5

3E5

4E5

particle

s/cm3

08:0009:00

10:0011:00

12:0013:00

14:0015:00

16:0017:00

Bochud, Riediker – Studies on health impact of air quality

OpenSense2

Phase 2: Crowdsourcing,dispersion modeling & “closing the loop”

Conclusion

50

Take Home Messages• Sensor networks enable environmental monitoring in remote locations

and of difficult-to-measure processes• Real-world deployments may be highly unpredictable!• Mobile sensor networks can increase coverage and spatial resolution of

measured data• Increasing the resolution of air pollution data is necessary for

understanding health impact.• Whether data extracted from poor quality measurements can be

processed to obtain useful data on air pollution is an important research question.

• Other questions: How to design the node? How to control the network?• Using existing mobility sources holds important benefits, but achieving

a reliable system integration is non-trivial.

51

Additional Literature – Week 11Environmental engineering applications• Static sensor networks

- Sensorscope: http://sensorscope.epfl.ch/- Swiss Experiment

http://www.swiss-experiment.ch/- Permasense: http://www.permasense.ch/- GITEWS: http://www.gitews.de- WiSARD network: http://wisardnet.nau.edu/- CENS: http://research.cens.ucla.edu/

• Mobile sensor networks- OpenSense: http://opensense.epfl.ch- CENS: http://research.cens.ucla.edu/urbansensing/ 52

Environmental engineering applications• Robotic sensor nodes and networks

– Aquatic microbic observing systems http://robotics.usc.edu/~namos/index.html

– Adapting sampling of oceans http://www.princeton.edu/~dcsl/asap/

– Robots and sensor networks systems for underwater monitoring http://groups.csail.mit.edu/drl/wiki/index.php/AMOUR

– http://research.cens.ucla.edu/mas/– http://research.cens.ucla.edu/aquatic/– IEEE Robotics and Automation Magazine, special issue on

Robotics for Environmental Monitoring, M. Dunbabin and L. Marques, editors, March 2012

Additional Literature – Week 11

53

Civil engineering applications• Static sensor networks

– Structural Health Monitoring http://enl.usc.edu/projects/shm/index.html

– Structural Health Monitoring http://www.empa.ch/plugin/template/empa/93/*/---/l=2

- Structural Health Monitoring

http://www.eecs.berkeley.edu/~binetude/ggb/- Structural Health Monitoring

http://research.cens.ucla.edu/projects/2007/Seismic/Tall_Special/

Additional Literature – Week 11

54

Additional Literature – Week 11• Static sensor & actuator networks

– Structural Controlhttp://people.ce.gatech.edu/~ywang/research.htm#_WirelessControl

– Structural Controlhttp://imacwww.epfl.ch/Common/research-en.jsp

– Structural Control for wind effect mitigationhttp://jahia-prod.epfl.ch/cms/site/disal2/op/edit/page-32507.html

– Structural Controlhttp://www-personal.umich.edu/~jerlynch/index.html

• Mobile sensor networks– Monitoring of water pipe networks

PipeProbe: http://mll.csie.ntu.edu.tw

Civil engineering applications

55