 |
The Technology Supporting the Vision
By David Pescovitz
|
Eric
Brewer established TIER, Technology and Infrastructure for
Emerging Regions, to develop the technology that will form
the foundation of the work done by the ICT4B effort.
BART NAGEL PHOTO |
“If you don’t change the way technology is spread,
the gap between the rich and poor will just continue to widen,”
says Eric Brewer, professor of electrical engineering and computer
sciences (EECS) and principal investigator for the ICT4B project.
To that end, Brewer established Technology and Infrastructure
for Emerging Regions (TIER), a research sub-group within the vast,
interdisciplinary ICT4B effort. TIER is tasked with developing
the technology that will drive ICT4B to success. Ultimately though,
Brewer hopes ICT4B will become just one “TIER customer”
side by side with various non-governmental organizations and companies
that also find value in the Berkeley developments.
Below is a sampling of the research projects proceeding under
the auspices of ICT4B and, in the case of the medical technologies,
the Center for Information Technology Research in the Interest
of Society (CITRIS). The aim of this use-inspired basic research
is to help citizens of emerging nations solve their toughest problems
through a wide and creative range of innovative engineering solutions.
TIER is not technological imperialism, Brewer says. It’s
about providing appropriate technologies for emerging regions
to help themselves.
“In the last 100 years, technology has been the single factor
that caused the most change in the first world,” Brewer
says. “Now, technology is the best source of hope in developing
regions.”
Wireless Networking
The Problem: People are unable to communicate
or share information quickly and easily beyond their own village.
A robust wireless network must be developed to link rural villages
in ways that are cheap and easy to manage. This infrastructure
will be the foundation for most ICT4B efforts.
The Team: Professor Eric Brewer, EECS; Kevin
Fall, Intel Research; EECS graduate students Sergiu Nedevschi,
Rabin Patra, Michael Demmer, Sonesh Surana; Jordan Hayes, a Berkeley-based
software engineer volunteering on the project.
The What: A robust wireless network is ICT4B’s
foundation in low-income developing countries where telecommunication
infrastructure, not to mention electricity, may be nonexistent
or prohibitively expensive to use. Novel hardware and software
must be developed to link rural villages in ways that are cheap
and easy to manage.
The How: One networking project focuses on dramatically
extending the range of inexpensive 802.11 Wi-Fi chipsets, which
use directional antennae to transmit 50-80 kilometers from village
to village until it reaches a cellular base station or network
hub in an urban location. However, users of wireless services
in rural regions can’t count on the instantaneous two-way
communication of traditional networks. Everything from bad weather
to a power outage can wreak havoc on a rural network. As a result,
the ICT4B team is designing “delay-tolerant” systems
that intelligently store data within the network, and route it
in hops from point to point to its intended location during moments
of connectivity.
The Status: This summer, the team visited several
small villages in southern India where a variety of wireless solutions
donated by industry made it possible for residents to exchange
electronic information and videoconference with other villages
about commerce and political issues. By studying the “logs,”
essentially the meeting notes, of how the villages use their technology,
the researchers hope to better understand what kind of services
the citizens are likely to use and what ingredients are necessary
to build a more sustainable network infrastructure.
The Reason: “With networks that go into
unusual or remote areas, it’s either impossible or prohibitively
expensive to have an always-on network,” Fall says. “Before
now, nobody has specifically tried to design a network for environments
that have those problems.”
Bio-Chips for Disease Detection
The Problem: Dengue fever, a tropical disease,
incapacitates as many as 100 million people each year. Medical
facilities capable of testing for this disease and other diseases
are often far away from villages.
The Team: Professor Bernhard E. Boser, EECS;
Professor P. Robert Beatty, Molecular and Cell Biology; Professor
Eva Harris, School of Public Health; EECS graduate student Turgut
Aytur; BioE undergraduate student Jonathan Foley; Department of
Molecular and Cell Biology undergraduate student Wilfredo Lim.
The What: The two millimeter-square ImmunoSensor
chip provides a quick, inexpensive test for the dengue virus.
The How: Melding microbiology with microcircuitry,
the ImmunoSensor puts a laboratory on a chip at a cost that should
approach less than $1 each. Plugged into a conventional laptop
computer, the chip can analyze a drop of blood serum for antibodies
that indicate if the patient is infected with the dengue virus.
The entire test takes approximately 15 minutes.
The Status: The research team plans a summer
2005 field study for the ImmunoSensor bio-chips in Nicaragua.
Meanwhile, they’re also developing an HIV test that would
run on the same platform.
The Reason: “In the third world, there
aren’t very many means outside of specialized labs to test
blood samples,” Boser says. “Many regions don’t
even have the quality of water you need to do traditional tests.
But you could imagine buckets of these chips, all coated with
different antibodies or antigens, so not only can we detect on-the-spot
when someone is ill, we can also find out exactly what illness
they have.”
Speech Recognition
The Problem: Different languages and levels of
education make traditional alphanumeric keyboard interfaces unsuited
for most applications in developing nations. If you can’t
read, you can’t type.
The Team: Professor Eric Brewer, EECS; EECS graduate
students Sergiu Nedevschi and Rabin Patra; International Computer
Science Institute researcher Chuck Wooters.
The What: An inexpensive all-in-one chip will
enable users in developing regions to talk to next-generation
information and communication devices in various languages, from
Hindi to Tamil.
The How: Even though reliable recognition of
large-vocabularies is still a challenge in computer science, a
single chip that can handle a very limited number of words is
a reasonable goal. The researchers have built a single device
that can accurately recognize 30 to 100 words spoken by any speaker.
The chip can also be programmed dynamically during a particular
application. For example, if a user is filling out a long online
form, the chip may recognize specific words depending on the content
of each page.
The Status: The device is currently being redesigned
as a custom Application Specific Integrated Circuit (ASIC) to
keep costs and power requirements to a minimum. With the aid of
the Berkeley Tamil Studies Program, the researchers are collecting
samples of the Indian language to program the chip.
The Reason: “We can’t assume that
the users of devices are literate, so spoken language input and
output plays a major part in the design of user interfaces for
emerging regions,” Brewer says.
Cheap Printable Electronic Displays
The Problem: Current costs for displays make
it difficult to provide affordable information and communication
devices for villagers.
The Team: Professor Vivek Subramanian, EECS;
EECS graduate student Alejandro de la Fuente Vornbrock; EECS undergraduate
student Teymur Bakhishev.
The What: A four- to six-inch computer display
fabricated using an inkjet printer at a cost of under $5 dollars
each.
The How: A reel-to-reel fabrication process negates
the need for high-vacuum processes and clean room lithography
that are the dominant costs in traditional displays. “Organic
electronics” to drive the display are literally printed
on a flexible substrate using organic and nano-particle inks.
The display itself relies on polymer dispersed liquid crystals
(PDLC), a material that can be made to look black or white depending
on the application of an electrical field. Unlike traditional
liquid-crystal displays, the PDLC material does not require additional
layers of optical components to produce an image.
The Status: By summer’s end, the researchers
were scheduled to have demonstrated a 3-inch passive matrix display
with more than 100 pixels.
The Reason: “By developing a fully integrated,
all-printed, low-res display, we hope to deliver unprecedented
cost-reduction in ubiquitous information appliances," says
Subramanian.
Distributed Network Imaging
The Problem: Some of the most advanced surgical
techniques haven’t reached rural villages because the hardware
is too costly.
The Team: Professor Boris Rubinsky, ME and BioE;
David Otten, a recent Mechanical Engineering Ph.D. graduate; Dr.
Gary Onik, medical director of surgical imaging at Florida Hospital,
who works closely with Rubinsky.
The What: This novel system for medical imaging
provides doctors in remote villages with a real-time view inside
a patient’s body during minimally invasive cancer surgery.
The beauty is that the expensive equipment generating the images
can be located thousands of miles away from the patient and shared
by many remote physicians at once.
The How: Medical imaging systems convert data
collected by sensors near the body into an image of what’s
inside. Although the sensors are often inexpensive, the computing
power needed to translate the raw data into an accurate and detailed
3-D image is not. Distributed Network Imaging calls for the data
collection hardware to be installed at the remote site where the
actual surgery will take place. As the patient’s raw data
is generated, it is instantly digitized and transmitted via existing
communication conduits—telephone lines, satellite links,
or wireless networks, for example—to a state-of-the-art
image reconstruction server located in an urban hospital or university,
where it can be remotely accessed by doctors. The images are instantly
sent back to the remote sites for a physician to consult during
a surgical procedure or for diagnosis.
The Status: To prove their concept, the researchers
used electrical impedance tomography (EIT) to image in vitro cryosurgery
on a liver over a modem link. Commonly used in developing countries,
cryosurgery is a minimally invasive surgical technique Rubinsky
and Onik helped pioneer, in which a tiny tubular probe inserted
into the body kills cancer cells with blasts of intense cold.
EIT imaging helps surgeons guide the tube to the tumor site and
monitor the freezing to ensure that the tumor is completely engulfed
in ice.
The Reason: “We believe that the distributed
network concept will eventually provide diagnosis and treatment
of cancer and genetic diseases to parts of the world population
that have not been exposed to advanced medical technology in the
past,” Rubinsky says.
David Pescovitz writes Lab Notes, the
College of Engineering’s online research digest and contributes
to Popular Science, Small Times, and Business 2.0.
His writing on science and technology has been featured in Wired,
Scientific American, IEEE Spectrum, and the New York
Times.
|
