Tiny battery-powered
computing devices - embedded in the shoes of an Alzheimer's disease
patient, in the soil near the San Andreas Fault or on site at a
manufacturing plant - offer a new lens on the physical world and the
means to interact with it in real time
Living alone in
her quiet home, Harriet often feels lonely and depressed. Newly
widowed at 82, she has little interest in physical activity and no
zest for social contact. Eating is a chore. She doesn't feel like
going out for a meal, and she doesn't feel like fixing food for
herself. Consequently, Harriet's health is in serious jeopardy.
This combination of social isolation, inactivity and failing
nutrition is alarmingly common among today's aging population, but
information technology may offer the means to counteract a harmful
outcome. Imagine that there's a tiny sensor embedded in one of
Harriet's walking shoes and another on the peg where she hangs her
outdoor jacket. Whenever these and other miniature sensing devices
throughout Harriet's home notice her moving in the direction of her
front door, the sensors use a wireless network to alert Sonya,
across the street, and Roland, down the block, that Harriet is about
to go for a morning walk. Sonya gets the suggestion on her TV, and
Roland hears it by phone: "Call Harriet now if you want to go along
for some exercise." Later the same day, Harriet learns much the same
way that cabinet doors are opening in her neighbor Ruth's kitchen.
Acting on a verbal cue that she hasn't eaten yet today, Harriet
calls Ruth to suggest that they prepare and eat lunch together. The
two end up sharing food, activity and conversation.
Such
technology scenarios may seem far-fetched, even science
fiction-like, but they are in fact close to becoming real. The key
is wireless
sensor networks, an intriguing new technology model in which
behavioral and biological data is collected and analyzed for
customized proactive computing applications. As the name implies,
proactive computing aims to anticipate people's needs and take
action to meet those needs on their behalf, relieving people of
tedious data entry.
Intel has taken the initiative to invest
in research and
development of sensor networks, recognizing this technology as
crucial to addressing the pending global age wave and public health
crisis. Experts predict that the worldwide elderly population will
grow dramatically as the post-World War II baby boom generation
reaches retirement age. In the United States alone, the cost of
caring for older adults is expected to escalate sharply in 2010,
when 76 million baby boomers begin to reach age 65¹. Cognitive and
physical decline are also concerns. The Alzheimer's
Association, for example, reports that more than 4 million
Americans have the disease - a number that's projected to triple to
14 million by 2050 as the elderly population continues to increase.
Sensor networks hold promise for meeting these challenges
because they have the potential to revolutionize healthcare, but
Intel also sees the technology as having broad application in
virtually all aspects of life and industry. Intel envisions a proactive
computing world in which a multitude of unseen, connected
computing nodes automatically acquire and act on real-time data
about a physical environment, helping to improve lives, promoting a
better understanding of the world and enabling people to become more
productive.
¹ Center for Aging
Services Technologies (CAST), 2002.
Wireless sensor networks represent an
entirely new way of looking at computing. In a sensor network,
dozens, hundreds or even thousands of tiny, battery-powered
computers, often called "motes," are scattered throughout a physical
environment. Silently and wirelessly, each mote in this ad hoc
network collects data, for instance, monitoring light, temperature,
humidity, vibration or other environmental factors. The mote relays
the collected data to its neighboring motes and then to a specified
destination where it is processed. This sensory input, when gathered
from all the motes and analyzed by more traditional computers,
paints a comprehensive, high-resolution picture of the surroundings
in real time.
What's in a mote?
Essentially, the ability to sense, compute and communicate. Despite
their complex functionality, motes have just three key hardware
components: microprocessors, tiny microelectromechanical systems
(MEMS) and low-power radios (also called transceivers). The
microprocessors process the data, the MEMS sensors provide a broad
array of sensory inputs, and the radios enable the motes to
wirelessly transmit their sensor readings throughout the network.
New software enables the raw data collected by the sensors
to be analyzed in various ways before it leaves the network,
ensuring a proactive stream of information that can be acted upon in
real time. Standard consumer AA or coin-style batteries keep motes
"alive" for six months to a year, and researchers are exploring
other sources of power to further shrink size and extend longevity.
Although the size, type and configuration of motes in a
sensor network depend largely on the application, common design
constraints include power conservation, compact form factor and
limited memory and storage capacity. Moreover, motes must be
reasonably economical to be suitable for practical applications.
Fortunately, microprocessors, sensors and RF
transceivers can be inexpensively produced in large quantities
using conventional semiconductor manufacturing techniques. Several
species of motes based on prototypes developed by Intel and the
University of California at Berkeley have recently become
commercially available at $50 to $100 (U.S.) each. Researchers at
Intel expect that, with re-engineering, Moore's Law² and volume
production, motes could drop in price to less than $5 each over the
next several years.
The combination of small size, low cost
and wireless networking functionality makes sensor network
technology exceptionally scalable. As price points become more
attractive, scientists will be able to deploy many sensors
simultaneously, with better proximity to the physical phenomena
being monitored and more detailed tracking than has heretofore been
possible, leading to ubiquitous computing.
² According to Moore's
Law (posited by Intel founder Gordon Moore in 1965), the number of
transistors on a chip roughly doubles every two years, resulting in
more features, increased performance and decreased cost per
transistor.
Research and development of wireless
sensor network technology has been a collective effort linking
Intel, university research centers, industry labs and government
agencies. At Intel, researchers have been working actively on the
technology since 2001. Their work has included extensive
investigation into practical applications, underscored by
ethnographic research and field studies, as well as developing
prototype sensor networks for testing and demonstrating the
technology. Meanwhile, other groups and labs at Intel work on core
technology elements, including silicon advances
and communications
network research.
The mote technology itself was
developed by Intel in collaboration with the UC Berkeley-based
Center for Information Technology Research in the Interest of
Society (CITRIS). UC Berkeley researchers also designed a
sophisticated operating system called TinyOS
specifically to manage mote hardware and software functionality.
Freely available and open source, TinyOS has become the de facto
industry-standard operating system for sensor network research and
applications.
TinyOS is written in nesC, a programming
language for deeply embedded systems. An extension of the popular C
programming language, nesC supports the event-driven processing that
is typical of motes, which remain "asleep" until sensors acquire
data or receive messages.
Researchers around the world have
developed numerous applications using nesC and TinyOS. However, if
heterogeneous
sensor networks are to be widely deployed, extracting meaningful
data from those networks must be simple. TinyDB,
a compatible database for sensor networks developed by researchers
at Intel and UC Berkeley, was built to do just that. Designed with a
graphical user interface and simple query structure, TinyDB
transforms diverse kinds of sensor networks into user-friendly
virtual databases rich with useful information about the real world.
The best example of Intel's efforts to
explore the human side of sensor network technology is the company's
Proactive
Health Research project. Since April 2002, a team based at
Intel's Hillsboro, Ore., campus has worked to develop technologies
that benefit today's aging population. The goal: catalyze research
on proactive computing applications that promote health and
wellness. Intel researchers expect their work will enhance quality
of life for elders, including the ability to "age in place" - to
stay home rather than move to institutions; improve healthcare
through prevention and early detection of disease; lower soaring
healthcare costs; and unburden family members and other caregivers.
Rooted in ethnographic
research such as in-home observations and interviews, the team's
work has focused largely on the needs of real people coping with
mild to severe cognitive impairments. By studying extreme conditions
such as Alzheimer's disease, researchers expect to discover core
technology and applications that people will find useful in other
health domains. For example, Intel researchers eventually hope to
extend their research to other physical ailments such as cancer and
cardiovascular disease. They see their work ultimately having
implications not just for monitoring and management of such
diseases, but potentially even for early detection and prevention,
such as spotting behavioral or physical problems that could indicate
the onset of Parkinson's disease.
Because of understandable
concerns around privacy and security, Intel has made these aspects a
core part of its research into sensor network technology.
Researchers have learned, however, that many of the people they
interview see the technology as a way to maintain privacy. This view
is most prevalent among frail elders who face being
institutionalized and losing the home they don't want to leave, or
having to come up with the money to pay for a 24-hour care nurse so
they can continue living at home. Seniors already living in a care
facility where they are subject to 24-hour camera monitoring also
respond positively to the technology as a way to regain the precious
privacy they've lost. In many cases, such elders will gladly share
nutritional and exercise data, and even confidential blood pressure
information, with friends or designated caregivers if it will help
them continue to live at home.
To help seniors age with more
dignity and independence, Intel is developing in-home technology
prototypes that assure private and secure data. Intel is also
building sensor network interfaces that will allow elders to make
decisions about the data they share, and who they share it with.
Interfaces are made intentionally simple and transparent, so seniors
won't need technical training to use them.
Intel is
encouraged by the fact that simple sensor networks can be built
today by adapting off-the-shelf technology, and home technologies
for health and wellness are beginning to appear on the market. Intel
is working with other companies, academic researchers and the
government to help ensure that more proactive, interactive
technologies will be commercially available to anyone who needs them
by the end of decade, when boomers begin to retire in huge numbers.
Because wireless sensor networks
promise to usher in a new model of computing, the repercussions are
certain to be felt beyond healthcare in society at large.
Researchers at Intel and elsewhere are actively exploring and
experimenting with innovative applications in a number of areas
today, including environmental monitoring, agriculture, wildlife
biology, public safety, structural engineering and manufacturing.
For example, in a pinot noir vineyard
in Oregon's Willamette Valley, embedded sensors monitor temperature.
Each mote takes one temperature reading per minute and stores the
results, also recording the highest and lowest readings for each
hour of the day. It's a small start, but someday researchers imagine
a "smart vineyard" in which sensors monitor nutrients, keep the
vines free from pests and mold, and track moisture - perhaps even
sending alerts to irrigate as needed. The project also incorporates
input from those who tend the vines, cultivate the soil and harvest
the grapes, as well as the needs of vineyard managers, winemakers
and wine sellers.
On Great
Duck Island off the coast of Maine, scientists are using a
prototype sensor network to monitor the nesting grounds of elusive
seabirds called Leach's storm petrels. Coin--sized motes buried in
burrows house sensors that continuously monitor light, temperature
and barometric pressure. Readings are relayed to laptops, and then
to an Intel lab
in Berkeley, in near real-time, for interpretation. The result:
Biologists get the information they need to observe the birds and
protect habitat, with minimal human disturbance.
In San
Francisco, 200 motes organized in an ad hoc sensor network track
stresses on the Golden Gate Bridge. The motes measure how much the
bridge sways from side to side, which can amount to several feet in
strong winds. Using RF, sensor readings are wirelessly sent to more
powerful computers for data analysis. Any anomalies might indicate a
weakness in structural integrity, alerting engineers to repairs that
can help keep the span safe in case of earthquake or other natural
disaster.
Intel is also testing sensor network technology by
installing motes that monitor cooling equipment in one of its
manufacturing plants. In this project, sensors watch for abnormal
vibrations that could indicate worn bearings or failing compressors
in the machines. Alerts are wirelessly transmitted to a central
control system, keeping small mechanical problems from becoming
costly production issues.
Still other teams of researchers
have field-tested sample sensor networks for monitoring the
micro-environment in a redwood grove, for tracking the rescue
operations of firefighters and for increasing crop yields through
automated irrigation and maintenance.
These and other
applications offer proof that sensing technology has enormous
potential in terms of delivering new benefits to society. They also
point to opportunities - in technology innovation, in increased
productivity and in strategic advantage for any nation that intends
to compete on a global scale.
Ultimately, Intel envisions sensors
built into roads, farms, hospitals, factories, office buildings,
clothing, swimming pools, baby cribs, vehicles and even bandages - a
ubiquitous computing infrastructure that consumers around the globe
can tap into and take for granted.
But research is still in
its infancy, and pursuing this goal will require long-term
collaboration among industry leaders, academia and government. By
underwriting sensor network R&D, these entities will fuel viable
and scalable solutions to the cresting age wave and multiple other
challenges faced by numerous industries and market segments.
**This link will take you off of the Intel
Web site. Intel does not control the content of these linked Web
sites.