20
CSIRO Sensors & Sensor Networks 8 June 2009 Dr. Darrell Williamson Deputy Director, CSIRO ICT Centre Marsfield, Sydney AUSTRALIA
CSIRO
Sensors & Sensor Networks
8 June 2009
Dr. Darrell Williamson
Deputy Director, CSIRO ICT Centre
Marsfield, Sydney AUSTRALIA
CSIRO Goal: Sensor & Sensor Networks
To provide a transformational data driven approach
to scientific discovery through the creation of sensor
network technologies that will transform our
understanding and management of the environment and
resources.
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
Significant Activities - since 2003
HARDWARE/SOFTWARE
• FLECKTM sensor node – wireless range
>1000m; nano FLECKTM <20m
• Operating Systems: TinyOS & FOS
APPLICATIONS
• Water quality
• Rainforest re-vegetation
• Virtual fencing
• Pipeline monitoring & localisation
• Coastal monitoring
• Sensor informatics
• Salinity intrusion
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
Receiving
models
Environment
impact assessment
Monitoring +
assessment
Adaptive
coastal
communities
WfO– ORCA
Climate
projections
Climate
adaptation optionsCAF– SSC
Coastal
vulnerability
Wave +
sea-level
modelling
Monitoring +
Observation– IMOS, TERN
Catchment
models
Material
flux
Catchment
restoration
Urban
water
WfHC– HWE
– Urban
Water Quality / Catchments to CoastHealthy Water Ecosystems, Our Resilient Coastal Australia
Agricultural
Sustainability
Wealth from Oceans
Flagship
Water for a Healthy
Country Flagship
Information management
Nodes and networks
Novel sensors
CSIRO
Organisation of Sensor & Sensor Network Research
Capability
Water Quality
Monitoring
Catchments to
Coast
Terrestrial
MonitoringProjects
Low cost sensors
Energy constrained
communications
Analysis tools
Organic and pesticide
contamination
Underwater networks
Query abstraction
On node AV processing
Energy harvesting;
Underground comms.
Security and privacy
Achieving sustainable
multiple use of estuaries
and coastal zones
Informed management of
catchments, aquifers,
rivers and recycled water
Ecosystem monitoring
and precision farming for
sustainable agriculture
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
Deployments: 1
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
Deployment Area;
#Nodes
Sensors Notes
Belmont, QLD
Dec05
CSIRO Livestock
industries
2,500,000m2
50 static;
50 mobile
Soil; multispectral;
VF collars; GPS
1 min @ 50msec rates
Virtual Fencing,
Pasture Monitoring,
Environmental Impact
Reduction
Lake Wivenhoe,
Nov08
SeqWater,
QLD Gov
27,000,000m2
50 static;
Water temp @ 6 depths;
1 min rate
Detection of algal
blooms and other
abnormal events
Springbrook,
QLD
May08
DERM
135,000m2
615,000m2 by
2011
10 static
50 by 2010;
200 by 2011
6 per node; soil
moisture & temp, leaf
wetness, wind speed &
direction, air temp,
humidity
Future Sensors:
barometric press,
rainfall, tree thickness,
sap flow, audio / video
fauna monitoring
Monitoring rainforest
regeneration,
Energy adaptation for
long life, low power,
minimal information
loss
On-node processing of
audio/video signals for
biodiversity monitoring
Deployments: 2
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
Deployment Area;
#Nodes
Sensors Notes
Burdekin Water
Board QLD
Dec05
CSIRO Livestock
industries
600,000 m2
9 static;
Water level; salinity;
pH; flow rate
1 min rate
Water resource and crop
irrigation management
Lansdown
2010
CSIRO Livestock
Industries
3,000,000m2
600 static (to be
deployed)
Microclimate / weather
stations
Will form part of the new
site infrastructure
Elliot Research
Farm TAS
Dec07
Tasmanian
Institute for
Agricultural
Research
Dairy Australia
500,000m2
72 static
Soil Moisture @ 3 levels
1 x Automatic Weather
Station;
15 min rate
Deficit irrigation
management in dairy
farming
Capabilities
Audio and visual nodes for species recognition
and identificationOn-node and in-network signal processing
Advanced water quality sensors:
(cheap, drift-free, resistant to bio-fouling)New nano-materials and fabrication processes
Wider node spacing, networks that are robust to
dynamic environmental conditionsEnhanced radio communications
Networks which are robust under changing
physical configurationsNetwork protocols
Under water or underground networksAcoustic and/or optical communications
Generic nodes, self-healing networksNode operating system
Energy neutral networksEnergy harvesting and prediction,
energy-cognisant operation
Tools for users to query networkSemantic querying
Tools for data analysis and visualisation
accessible to end-usersData analysis and visualisation
No
ve
l
Se
ns
ors
No
de
s &
Ne
two
rks
Info
rma
tio
n
Ma
na
gem
en
t
Projects
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
Long Term Focus
• Information management: Query, analysis, security, privacy
• Novel Sensors: Low cost sensors, water quality, AV processing
• Power management: Energy harvesting, protocols
• Network technologies: Communication, hardware nodes, interfaces
0
20
40
60
80
100
120
08/09 09/10 10/11 11/12 12/13 13/14
Network technologies
Novel sensors
Power management
Information management
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
Benefits of Sensor Networks
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
• Ability to gather data about the natural and built environment at a
spatial scale and intensity not easy to achieve by other means
(manual collection, data logging, remote sensing).
• Real-time or near real-time data availability.
• Wireless sensor networks (WSNs) can be “infrastructureless”,
extracting energy from the environment and establishing their own
wireless communications network to concentrate the data to a
single gateway/backhaul point.
Drivers & Obstacles for Innovation: 1
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
• Lack of a set of clear “killer applications”. One potential “killer
application” may come in domestic, office and industrial energy
monitoring as an enabler for energy efficiencies driven by the
needs for carbon accounting/carbon trading/carbon tax regimes.
• Standards to allow embedded/commodity devices to have IPv6
connectivity (6LoWPAN)
• No developed market. Competition in the built environment
from conventional SCADA systems.
• Scalability issues – difficult to deploy large, long-lifetime
systems.
• Relatively high cost of sensor devices with respect to the cost
of the sensor nodes themselves especially for environmental
modelling, and purpose-built sensor networks.
Drivers & Obstacles for Innovation: 2
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
• Software is mostly bespoke, per deployment programming in
a difficult programming environment, increasing costs of
deployment.
• High cost of deployment and maintenance in environmental
applications
• Battery life/energy extraction from the environment is difficult
in some environments.
• Room for improvement in the energy/information cost
tradeoffs.
Infrastructure Issues: 1
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
• Internet side standards still in development; e.g. Open
Geospatial Consortium (OGC), Sensor Web Enablement
(SWE).
• Bandwidth (and energy use by the radio) is an issue within
WSNs for nodes close to the gateway, because of the higher
radio traffic carried by those nodes. Using more powerful
backbone nodes to provide in-network backhaul to the
gateway is feasible, but complicates the system.
• 6LoWPAN has the potential to make any sensor node
visible on the Internet.
Infrastructure Issues: 2
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
• Management of sensor data and allowing both real-time
access and efficient access to historical data can be
challenging.
• Improved data registry & search technology needed.
• Data and processing provenance, data quality assurance,
quality control and “ground truthing” data is a continuing
challenge.
Information Security: 1
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
• Security is difficult to maintain against a reasonably well-
resourced adversary. It must usually be assumed that an
adversary has physical access to any chosen nodes.
• Wired networks and networks in physically more secure
environments can be secured more easily.
• Energy costs of the processing for security, especially
when using Public Key cryptography can be high.
• Data availability is often a problem in realistic WSNs, often
with less than 70% of theoretically available data being
extracted from the system, often much less data.
Information Security: 2
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
• Data integrity is usually ensured within WSNs by CRC-
based, rather than cryptographic, message digests and is
more vulnerable to message insertion or message
modification.
• Care must be taken to preserve provenance and the
association of data with its metadata.
• Data availability can be adversely affected by natural
events: floods, fires, electrical storms.
Privacy
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
• Privacy of data is a problem within WSNs because of
issues above under “security”.
• The intensity of data collection could become a problem for
networks capable of inferring, for example, peoples’
movements or business use.
• There is a risk that privacy issues associated with intense
data collection may only become apparent after privacy is
compromised.
• The association of data with the device generating it for
provenance has associated privacy risks
Interoperability
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
• ZigBee, IEEE 802.15.4-2006 address interoperability at the
MAC layer.
• 6LoWPAN addresses interoperability at link and network
layer.
• Interoperability standards at the gateway are in development
(OGC SWE, W3C SSN-XG).
• No data format standards within the sensor network.
W3C-OGC Goal: Sensor & Sensor Networks
CSIRO. Sensors and Sensor Networks OECD Portugal June 2009
• W3C has announced the creation of the Semantic
Sensor Network Incubator Group, sponsored by W3C
Members: CSIRO, Wright State University, and OGC.
• The group's mission is to begin the formal process of
producing ontologies that define the capabilities of sensors
& sensor networks, & to develop semantic annotations of
a key language used by services based sensor networks.
• "As networks of sensors become more commonplace
there is a greater need for the management & querying of
these sensor networks to be assisted by standards &
computer reasoning.
• The OGC's Sensor Web Enablement activities have
produced a services-based architecture & standards,
including four languages for describing sensors, their
capabilities and measurements, & other relevant aspects of
environments involving multiple heterogeneous sensors.
Dr Michael Brünig (Program Leader, SSNTCP)
Dr Darrell Williamson
Contact CSIRO
Phone: 1300 363 400 or +61 3 9545 2176
Email: [email protected] Web: www.csiro.au
