Intel Research, working
with the academic community and industry, is addressing many of the
significant challenges for ad hoc sensor networks to become a
reality.
Already, a broad spectrum of sensor network pilot
applications have been demonstrated. As sensor network technology
emerges from research laboratories, the ability to instrument the
world is likely to transform every facet of our lives.
New Uses, New Users:
Intel and BP, one of the world's largest
petroleum and petrochemicals companies, are collaborating on a
joint research project using a wireless sensor network to provide
continuous vibration monitoring of the engines on one of BP's oil
tankers off the Shetland Islands in northern Scotland. View Video (WMV file, 10MB; requires Media
Player).
Smart Surrogates, by Terry Knott, BP Frontiers
magazine, Issue 9, April 2004 BP is at the forefront
of applying the latest sensory network digital technology to a
broad spectrum of its businesses. Learn
more.
Aging Boomers: Technology to the
Rescue? The "age wave" is coming and there's nowhere
to run. In the next 25 years, the 65-and-over population in
America will double. The first baby-boomers will reach retirement
age just six years from now. Already, healthcare is America's
biggest cost, and fastest growing too. It's a huge challenge - but
technology, healthcare and education leaders say it is also a huge
opportunity. Reporter Rick Lockridge provides insight from CAST
and Intel Sensor Net Open House events held March 2004 in
Washington, D.C. View Video (WMV file, 6.93MB; requires Media
Player).
For more details about this new class
of technology:
In today's model of computing, we
interact directly, one-on-one, with our desktop PCs, mobile phones,
and personal digital assistants. In the near future though, the
majority of computers will be embedded deep in the world around us,
hidden inside our homes, roads, farms, hospitals, and factories.
When we are in control of hundreds or thousands of computers each,
it will be impossible for us to interact directly with each one. The
time has come to transition from interactive to proactive computing.
These proactive computers will anticipate our needs and sometimes
act on our behalf. Sensor networks represent this paradigm shift in
computing.
Events held in March 2004 in Washington, DC
showcased various wireless sensor technology and real-world
applications.
Wireless sensor networks have
value in industrial applications only if they can demonstrate return
on investment. Intel is currently conducting a pilot deployment of a
wireless sensor network to monitor the health of semiconductor
fabrication equipment. In the same way that a car's engine sounds
"right" when it is well tuned, heavy equipment has a characteristic
vibration signature in normal operation. Intel currently uses manual
monitoring to predict failures and schedule maintenance or
replacement to avoid costly manufacturing downtime. By deploying
wireless sensor networks, Intel intends to demonstrate reduced
equipment failures through continuous monitoring, low installation
cost, and elimination of costly manual equipment monitoring.
A return on investment of this technology can also be
realized in other industrial and government sectors. Using building
sensors that intelligently target the application of fire
suppression sprinklers within a building, businesses can recover
from small fires more quickly and insurance costs can be reduced. By
embedding sensors into bridges that cooperate to automatically
verify the safety of these structures after an earthquake, city
infrastructures can quickly return to normal operation. Over time,
such applications of wireless sensor networking can reduce the cost
of doing business. Learn
more.
IrisNet is an architecture
and system for a worldwide sensor web. The sensor web combines a
variety of sensor types, including video, and permits globally
distributed data collection, actuation, and data mining. In this
demo, we show how IrisNet can be used to help drivers locate
available parking spaces near their destination.
Overview of IrisNet: Today's common
computing hardware - Internet connected desktop PCs and inexpensive,
commodity off-the-shelf sensors such as Webcams - is an ideal
platform for a worldwide sensor web. IrisNet provides a software
infrastructure for this platform that lets users query globally
distributed collections of high-bit-rate sensors powerfully and
efficiently.
Authoring and deploying sensing services
requires addressing a number of challenges related to data
acquisition and processing, placement and replication of the sensor
data for availability and performance, query processing on the
widely distributed sensor data, data integrity and privacy, etc.
IrisNet aims to provide these generic functionalities as well as a
simple programming interface such that the service author can easily
write their new sensing services.
Example wide-area
sensing applications:
Parking space finder
Watch-my-child: After your child departs for school
in the morning, a few clicks of the mouse allow you to ask whether
your child has arrived safely. When your child doesn't return home
at the expected hour, a few queries to a web browser allow you to
determine the last location where your child was seen, and to ask
if your child has deviated from her usual route home.
Watch-my-aging-parent: Enable the elderly to live at
home while being monitored as needed.
Epidemic early warning systems: Discover
concentrations of sneezing, fevers, etc.
Homeland security
Computer network monitoring
Planet-wide sensor observatories, e.g., for
near-shore oceanography
Palo Alto Research Center: "Distributed
Attention: Enabling Sensing in Complex, Urban
Environments"
As sensing technology becomes cheaper
and more effective, large networks of intelligent sensors are moving
into heavily populated environments in applications such as traffic
safety, factory monitoring, urban surveillance, and battlefield
situational awareness. In these complex environments, the system
needs to sense a small number of "misbehaving" people or
vehicles--mixed in with countless of their counterparts simply going
about their everyday business. Monitoring all activities in these
environments is both overly intrusive and impractical, even with the
ever-greater computing power available each year.
Our system
is designed to juggle many sensing tasks according to their
importance, while simultaneously looking out for new, surprising
behaviors. An analogy can be made to attention in humans, where the
eye is attracted to interesting stimuli in the environment, and the
brain in turn focuses on interpreting the most important stimuli in
greater detail. For example, we sense people sneaking up on us "out
of the corner of our eyes" and focus attention on them to make sure
they are not a threat. In our demonstration, we will show video of
our system tracking small numbers of real military vehicles using
sound alone, and of our more recent work tracking "interesting"
vehicles among larger numbers of normal vehicles using a network of
steerable cameras. Learn
more.
Intel Research: Proactive Agriculture, Sensor
Networks for Understanding Site Variability
Agriculture is a domain that
we believe is one of the areas most likely to include early adopters
of wireless sensor networks. Our early work with agriculturalists
showed that sensor networks could play a role in preserving the
environment by reducing water and pesticide usage. They can also
provide early alerts for frost damage and help in precision
harvesting to maximize quality. These theoretical benefits needed to
be tested in the field. Since then, we have worked with agricultural
scientists on a long-term deployment of a wireless sensor network in
a wine grape vineyard. By densely monitoring climatic conditions we
were able to show that differences within a site are substantial and
often cannot be predicted by statistical models. Some of the
benefits of the development of this technology include improving
crop quality and, thereby, enhancing the value of the crop. The same
infrastructure can also be used to give performance support for use
in precision agriculture. Perhaps more significantly, this
technology can also help farmers to feed the expanding population of
the nations of the world by increasing the viability of semi-arable
lands.
This technology brings computation outside the
predictable domains of office automation, and allows researchers and
practitioners to address other types of knowledge work, in this
particular case - farming. This project begins to address
fundamental research issues outside the boundaries of its own core
technology. Learn
more.
Intel Research: Sensing Social
Health, Assisting Seniors with Cognitive Decline
The
mission of Intel's Proactive Health
Research lab is to catalyze and conduct pioneering research into
home healthcare and aging-in-place technologies that can improve
peoples' health and quality of life while reducing the nation's
overwhelming healthcare bill. As the U.S. population of seniors
doubles over the next fifteen years, it will become increasingly
necessary to shift the locus of care from formal institutions to the
home/workplace, from professional providers to the friends and
family members who already provide billions of dollars of care
annually, and from crisis-driven, reactive care to preventive,
proactive care.
After a year-long field study of families
dealing with Alzheimer's, Mild Cognitive Impairment (MCI), and other
cognitive disorders, our researchers have built various prototypes
to help people with MCI and their caregivers to better cope with
this often devastating condition. Our recent focus has been on tools
for "social health monitoring and support," which we will trial with
MCI households in Portland and Las Vegas this summer. Today's
demonstration shows prototypes of a wireless sensor network that
looks for sudden declines in social contact, provides visualization
of one's social health, and employs a screen phone that uses the
sensor data to provide rich contextual cues (e.g., who is calling
me, when we last spoke, what we discussed). Our long-term goals are
to use home-based technologies to aid in the early detection of
cognitive decline, to embed cognitive assessment metrics into
everyday activities, and to help those with decline stay socially
active and engaged for as long as possible.
Intel
Research Seattle University of Washington:
Caregiver's Assistant and CareNet Display: Making Eldercare
Easier
Caregiver's Assistant. Monitoring the
activities of elders is an important aspect of eldercare. However,
it can be intrusive for the elder and exhausting for the caregiver.
Can technology help? We have built a prototype system that can make
eldercare easier by detecting the activities of an elder without
requiring direct observation by a caregiver.
Our system
collects data from small, wireless, battery-less sensors called
Radio Frequency Identification (RFID) tags that are stuck on
household objects. These sensors tell us which objects are touched
and when. We then use statistical methods on this data to detect
high-level activities.
Our proof-of-concept prototype shows
how this approach can benefit caregivers. The Caregiver's Assistant
helps fill out the standard Activities of Daily Living (ADL) form,
allowing the caregiver to focus on the elder's quality of care
rather than tedious tasks.
CareNet Display. The CareNet
Display is an interactive, digital picture frame which augments a
photograph of an elder with information about her daily life. It is
used by the many people who provide the elder with care, including
family and friends, to coordinate the elder's care. Learn
more.
Sensor network application development and
deployment presents daunting software design challenges. Yet, the
ultimate users of sensor network technology, ranging from plant
biologists examining micro-climates in a giant redwood trees to
facility managers monitoring vibration signatures of their
equipment, are most likely not sophisticated software developers.
At Intel Research in Berkeley, we have built a suite of
tools called the Tiny Application Sensor Kit (TASK) that break down
the barrier to entry for non-sophisticated users to develop and
deploy their own sensor network applications.
TASK consists
of the following components:
TinyDB - A software component that allows
programs to interact with the sensor network through a declarative
SQL-like interface.
TASK Server - A server process running on a
sensor network gateway that acts as a proxy for the sensor network
on the Internet.
TASK DBMS - A relational database that stores
sensor readings, sensor network health statistics, sensor
locations and calibration coefficients, etc.
TASK Client Tools - These include a TASK
Deployment Tool which helps users record sensor node metadata, a
TASK Configuration Tool that helps users choose data collection
intervals and data filtering and aggregation criteria, and TASK
Visualization Tool that helps users monitor the network health and
sensor readings.
TASK Field Tool - This tool runs on a PDA and
permits in-situ diagnosis and resolution of problems in a deployed
sensor network.
TASK also integrates easily with most
popular data analysis tools, e.g., MS Excel, Matlab, ArcGIS, etc. At
present, TASK runs on the Mica2 and Mica2Dot sensor network
platforms with weather station sensor boards, the Intel XScale®
technology based StarGate sensor gateway, and most x86 based PCs
running Microsoft Windows or Linux operating systems. Learn
more.
Figure 1. NIMS
creates new mobile sensing devices that explore the
environment on suspended infrastructure. By integrating
networked embedded sensing with mobility, a sustainable,
and self-aware sensing system is
created.
Sensor
networking capabilities are urgently required for some of our most
important scientific and societal problems in understanding the
international carbon budget, monitoring water resources, and
safeguarding public health. This is a daunting research challenge
requiring distributed sensor systems operate in complex environments
while providing assurance of reliable and accurate sensing.
Networked Infomechanical Systems (NIMS) (see Figure 1) is a
program within the Center for Embedded Networked Sensing (CENS) that
adds essential new architectural tiers to the sensing system
ecology. By combining fixed and mobile nodes with infrastructure,
the remote sensing system may be sustainable; the sensor network may
now collect and distribute energy, introduce new sensors, reposition
communication devices, and also calibrate sensing systems. A
particularly important new attribute enabled by NIMS is
self-awareness that will provide sensor networks with the ability to
probe their own comprehensive sensing performance and ultimately
adjust physical configuration to optimize and maintain sensing
performance.
Figure 2. NIMS
node and cable deployed at a height of 50 m at the
Wind.
This
demonstration will display the NIMS system in operation as it has
been deployed in the field (for example, at the Wind River Canopy
Crane Research Facility in the Wind River Experimental Forest in
Washington (Figure 2)). with horizontal and vertical transport
capability, embedded computing, wireless networking, and sensor
systems. Sensor systems include articulated imaging as the
"microclimate" sensors required for the probing the interaction
between the forest canopy and atmosphere. System development is
underway for characterizing forest, river and stream, and in the
future, marine environments.
NIMS Systems have been
developed using the Emstar software development environment enabling
rapid prototyping, testing, and simulation across heterogeneous
fixed and mobile platforms.
Ohio State University: A line in the
Sand
Our demonstration is based on a field
experiment involving a sensor network classification and tracking
system called A Line in the Sand, held in August 2003. Our
experiment supported the objective of "putting tripwires anywhere",
including deserts and other areas where physical terrain does not
constrain dismount or vehicle movement. A smart dust sensor network
of 90 nodes, empowered with distributed middleware services was
used. These nodes self-formed into a network and used magnetometer
and micro-power impulse radar (MIR) sensors as a basis for locally
detecting metallic and nonmetallic objects moving through the smart
dust network. As objects moved through the network, the nodes that
detected them then cooperated to classify and track them. An
important challenge in our work was to make these distributed
middleware services robust and tolerant to a host of faults like
node failures, message losses, etc. Classification and tracking of
objects with significant metallic content (such as soldiers and
cars) and objects without significant metallic content (such as
civilians) was demonstrated at various speeds of motions (ranging
from 3mph to 25mph). Such a system is usable in several homeland
security related scenarios like surveillance, protection of
important assets, border patrol, etc.
In the demonstration,
we will showcase a miniaturized version of A Line in the Sand
wherein we will classify and track different types of objects with
varying magnetic content using magnetometer sensors. We will
illustrate how our system can be used in a realistic setting through
videos and posters from the actual outdoor demonstration. Learn
more.
University
of Virginia: An Energy-Efficient Surveillance System Using
Wireless Sensor Networks
Surveillance missions often
involve a high element of risk for human personnel and require a
high degree of stealthiness. Hence, the ability to deploy unmanned
surveillance missions, using wireless sensor networks, is of
practical importance for the military. Unlike the traditional sensor
devices used by the military, the wireless sensor devices we use are
much smaller and less expensive. That allows them to be rapidly
deployed behind enemy lines and in remote areas, such as mountain
passes, by dropping them from airplanes. Once deployed, the
middleware and application software we have developed in this
research project, self-organizes the sensor network into a
surveillance and communication system. The software also allows the
sensors to cooperatively detect, track, and identify different
targets of interest. The sensors can also activate more powerful
sensors, such as those that can capture video and audio data, on
demand and the aggregated data can be delivered to command and
control locations with the help of long-range communication devices.
Another key contribution of our software is that it extends the
lifetime of the system by a careful power management service, called
a sentry service. This service allows most of the nodes in the
network to hibernate until an interesting event occurs. Key
ingredients of this project were successfully demonstrated last
summer to researchers, government agencies, and the US military
personnel. Work is continuing to improve the middleware and to
integrate with other sensor systems.
UC Berkeley Industrial and Social
Applications of Self-Powered Wireless Sensor Nets
Ubiquitous networks of
wireless sensor and communication nodes have the potential to
significantly impact industry and society as a whole. In this
example, we are creating wearable computing systems that are
integrated into standard fire-fighting masks. We refer to them as
"Heads Up Display (HUD)" units that can show the wearer a
postage-stamp-size "You Are Here" map of the building-floor. The
same map can be seen on the Fire Chiefs laptop as he or she
coordinates the fire with the deployed fire crew. Using wireless
sensor platforms as the research base, we have also created
rudimentary 'beacons' for each fire fighter to wear. This allows
each fire fighter to be tracked on the HUD-maps as a moving
'red-dot.' Such tracking allows further coordination with the
Chief's laptop that monitors the main location of fire and smoke.
Improved designs of wireless smoke and CO alarms are also integral
to the project.
To achieve the full potential of these
wireless networks, practical solutions for self-powering these
autonomous electronic devices and smoke detectors need to be
developed. To address this potential, we have also developed a
miniature power source that uses ambient vibration as an energy
source for wireless electronics. Learn
more.
Links to learn more about sensor
network technologies: