Low Cost, High Tech

The technology needs of developing nations differ greatly from those in the rest of the world. Lack of basic infrastructures like telephone
lines, low literacy rates, limited electricity, and a shortage of qualified IT workers are just some of the challenges researchers face in bridging the digital divide. Clearly it’s not enough to simply bring existing technology to the world’s poorest; new technology must be created.

This presents a Catch-22, as Eric Brewer, Director of Intel Research Berkeley and co-founder of the Technology and Infrastructure for Emerging Regions (TIER) project, explains. “It’s always harder to create something brand-new that’s low cost because the volume would be so low it would be more expensive,” says Brewer, who is on leave from his Computer Science professorship at UC Berkeley to run
the Intel tablet in Berkeley this year.

Instead of creating new devices and applications from scratch, CITRIS researchers like Brewer and John Canny are coming up with innovative high-tech solutions using inexpensive, off-the-shelf products. Others, such as UC Berkeley Associate Professor of EECS Vivek Subramanian, are pioneering revolutionary new manufacturing techniques that will lower the cost of existing hardware.

Prof. Subramanian bathes in the yellow light of the Printed Electronics Lab. The yellow light is used to avoid damage to photosensitive materials used in the lab.

WiFi That Goes the Distance

In Southern India, rural villagers no longer have to travel 20 kilometers to the Aravind Eye Hospital to speak with a doctor. They can
now receive their consultation via a videoconference link. It’s just one of the many benefits of a new, point-to-point, long-distance
wireless connection that is providing low-cost internet connectivity to a region that before lacked even the most basic telecommunications infrastructure.

A collaboration between TIER and Intel Research Berkeley (Intel is a CITRIS Founding Corporate Member), the project uses
existing technology to keep costs down. “We’re using off-the-shelf wifi cards as you might find in a laptop. Those aren’t made to go 20
kilometers. But by changing the software and the antennas and other things, you can make them go much further than they’re meant to,” says Brewer. TIER is currently setting up a test-bed in the Bay Area for meshes of these types of intermittent connections so that, eventually, entire networks of villages could be linked up.

The project has certainly caught the attention of another CITRIS Corporate Founding Member, Hewlett Packard. Says HP University Relations Manager and CITRIS liaison Patrick McGeer: “Imagine a world 10 or 15 years from now where you just get these self-forming networks. Suddenly you’re going to be able in the developing world to jump two or three generations of technology. Places that had poor landlines are going to jump past analog land, past analog cellular, past digital cellular to VOIP and pure digital communications, with the entire world, carried not from a central server but peer-to-peer over these mesh networks.”

McGeer believes that much of the work being done in collaboration with CITRIS researchers at UC Berkeley and other academic and corporate partners on the project PlanetLab, a next-generation Internet platform, will ultimately make these networks more robust and self-managing. That’s important because qualified IT professionals are in short supply in developing countries, a problem that the next project indirectly addresses.

More Than Talk

Prof. John Canny and other CITRIS researchers are pioneering ways to provide inexpensive technology to developing countries.

When UC Berkeley EECS professor John Canny began designing computer literacy courses for students in the impoverished Uttar Pradesh region of India, he quickly ran into a problem. “A large fraction of poorer people in India lack the English skills needed to use a computer effectively. For instance, 95% of the web content in India is in English,” says Canny, who receives support from CITRIS Founding Corporate Member Microsoft. Making matters worse, many of the nation’s poorest must work rather
than attend classes, further reducing the chance they’ll ever acquire the language skills needed to surf the Web or use most of the available software programs. And English literacy turns out to be a more important economic enabler than computer skills, even in India’s “new economy.”

To rectify the situation, Canny and his students are working with local NGOs to create speech-based, game-like English language instruction that can be run on a cell phone. Cell phones are very affordable in India. Because a speech-based program would require
neither the radio nor the display, it would require very little power. In addition, Canny points out, “The cell phone is hands-free, so it’s
compatible with a daily lifestyle that demands a lot of menial labor.” In other words, kids could learn while they worked.

The challenge lies in the software, which must support a continuous speech interface and India’s numerous regional languages and dialects, on a device approximately 10 times slower than your average desktop PC. Using several accelerating technologies, Canny says they’ve got the program working up to speed. His team has been collaborating with faculty from the Graduate School of Education to design instruction custom-tailored to this unique and challenging learning environment. Finally, they are working to improve recognizer accuracy using context information and using design patterns to develop more engaging, effective learning games.

Education isn’t the only possible use for a speech-based interface like the one Canny and his students
are creating. Brewer sees potential for speech as a new kind of user interface. “The most expensive part of the computer interface is the
screen. If you can make it smaller or replace it with a voice interface, it would make it much cheaper,” says Brewer, who also points
out it would give people with limited literacy the ability to use a computer. He and two students have produced a proposal for the
architecture of a speech-controlled hardware device.

Displays for Less

If a voice interface offers up one tantalizing possibility for a low-cost interface, a cheaper screen presents another. Imagine a laptop screen that costs only a few dollars to produce and consumes a fraction of the power. Such a screen may soon become a reality thanks to an alternative manufacturing technique developed by Vivek Subramanian and his Organic Electronics Group.

Behind every pixel on a typical laptop screen is a silicon transistor, which is produced using a costly and wasteful method called subtractive processing. In addition, because silicon transistors are opaque, they block part of the light, requiring more power to make the screen bright. That’s why as much as 60 percent of power consumption on a laptop computer comes from the display.

Using a novel manufacturing technique that prints transistors and chips using transparent materials, Subramanian’s group has successfully produced fully-printed low-cost liquid crystal display cells, and has also produced with transistors for driving these made
from transparent printed zinc oxide nanoparticles. Because these cells are transparent, they also require less power to brighten the screen. “Compared to amorphous silicon, our transistors are just as good. There are still a few problems to solve, but it’s potentially much cheaper with no real trade-off in performance,” says Subramanian.

Displays aren’t the only thing that Subramanian’s printing technique is making more affordable. It’s also being applied to make cheaper RFID tags and sensors.

Benefits for All

Low-cost displays, robust speech-interfaces are technologies, and long-distance wireless networks may have been created with the needs and constraints of the developing world in mind, but it’s not difficult to imagine how such innovations will make their way into devices used by people in wealthier nations as well.

For industry, the developing world promises new economic opportunities. “Companies like Intel and Microsoft have covered Western markets. To grow, they have to expand their markets. Working in developing regions is a way to understand these regions and needs and come up with strategies that are appropriate down the road,” says Brewer.

Hence, there is a very strong economic case to be made for investment in technology for the developing world. Intel, HP, Microsoft, Qualcomm all key CITRIS sponsors have recognized this and are enthusiastic about supporting deployments in countries like India, China, and Rwanda. The most immediate benefits of the work being done by CITRIS’s researchers and corporate partners are not for the developed world alone. We believe at CITRIS that if we can improve a lot of the poorest four billion people on the planet, we will improve society as a whole.