Remember the first time you spun Google Earth’s globe and zoomed in to get a bird’s-eye view of your house? Imagine the additional wow factor if you could zoom in even closer to see a 3D view of your neighborhood, including video of small figures walking up the street or a bus pulling up to the stop a few blocks away.

Researchers at the University of Southern California’s Integrated Media Systems Center are developing software that could one day make access to that level of real-time data a reality. The GeoDec (Geospatial Decision Making) system integrates multiple sources of online data, including satellite imagery and video streams, to create a highly detailed 3D model. The system is designed to aid in decision making, allowing for complex queries that only taking into account multiple levels of layered data can answer.

“I would like to think of GeoDec as a set of tools that one can use to rapidly and accurately build a virtual geographic location (for example, a city) linked with all its related information,” says Cyrus Shahabi, the lead researcher on the project.

First, GeoDec rapidly constructs a 3D model from two aerial or satellite images. Then geospatial data that’s gathered from both public and private databases enhances the model, integrating road vector data and maps, as well as data associated with indexed geographic points. Finally, the system maps textures to the buildings and landscape and integrates available video streams to complete the picture.

A haptic glove interface, a la “Minority Report,” can then manipulate the resulting model. The gloves lets a user select and display different pieces of information, which are integrated in the display, and then with gesture-based commands zoom, rotate, and move them from one part of the screen to another.

According to Shahabi, the GeoDec system has many potential applications in the areas of emergency response, military intelligence, real estate, tourism, simulation/training, and video gaming. Shahabi says, “With the right amount of investment, a preliminary version of GeoDec can be ready to use in a one-year timeframe.”

Shahabi and Craig Knoblock, another member of the USC research team, founded a company called Geosemble Technologies in order to commercialize some of GeoDec’s capabilities.


Big Blue Makes Big Leap

In the race to develop alternatives that will extend the shelf life of Moore’s Law, carbon nanotechnology is holding its own. Recently, IBM researchers created the first circuit built entirely on a single-walled carbon nanotube; it measures just 18 micrometers long, one-fifth the width of a human hair. While individual carbon nanotube transistors aren’t new, IBM’s circuit integrates 12 transistors made of palladium and aluminum into a ring oscillator designed to switch between two voltage levels.

The circuit reached a speed of 52MHz, significantly slower than today’s processors but 100,000 times faster than previous devices made with carbon nanotubes. Researchers predict that with further development, carbon nanotube transistors could one day achieve terahertz switching speed, making them a promising replacement for today’s silicon-based circuits.

“In general, one has to find the right circuit layout that avoids parasitic contributions since those are currently limiting the circuit performance,” says Zhihong Chen, researcher at IBM’s TJ Watson Research Center in Yorktown Heights, N.Y. “On the device level that can be done by reducing the overlap between the gate and the source/drain contacts. This will involve an improvement of lithographical techniques.”

Another important aspect of the circuit’s design is that it uses a CMOS-type architecture, making it potentially compatible with existing silicon circuit design. This provides an avenue for introducing new materials into current chip-making technology without throwing out the silicon baby with the bathwater.

Chen says, “Our successful integration demonstrates the compatibility of carbon nanotubes with conventional circuit architectures and emphasizes that the use of nano materials does not automatically imply that well-established circuit concepts have to be abandoned.” He adds that it, “is very hard at this point to predict the timeframe for commercializing nanotube electronics.”

Even though the nanotube circuit represents a significant step in the development of silicon chip alternatives, it’s likely it will be several years before the technology is ready for primetime. “It is still laboratory research,” admits Chen, “and is very hard at this point to predict the timeframe for commercializing nanotube electronics.”


Taste The Future

Many of us take for granted that sense organs are specialized: You see with your eyes, smell with your nose, and taste with your tongue. But that’s not actually accurate: Our sense organs are important collectors of data, but it’s our brain that interprets that data and does the work of sensing. Researchers have discovered that it’s possible to reroute sensory information through a substitute sensory channel, and they’ve created a device that could one day help humans gather sensory information the way snakes, fish, and other animals do—with the tongue.

It may sound like science fiction, but BrainPort is a device that encodes visual data as electric impulses and transmits them to the tongue through a narrow plastic strip that’s topped off with 144 microelectrodes. The brain learns to interpret the signals appropriately and uses the information to function as it would if it was receiving the signals through the natural sense organ.

Interestingly, the idea isn’t exactly new. Dr. Paul Bach-y-Rita at the University of Wisconsin pioneered the device over 30 years ago. Initially, his team transmitted the impulses through the skin on a user’s back, but they quickly discovered that the tongue was much better suited for the task due to the density and sensitivity of nerve fibers and the ability to sustain electrical contact.

What is new are the ways in which the technology is being applied. Wicab, the company Bach-y-Rita founded, is researching the therapeutic use of BrainPort in the treatment of difficulties with vision and balance, Parkinson's disease, and autism. Additionally, a team at the Florida Institute for Human and Machine Cognition led by Dr. Anil Raj is developing sensory augmentation applications of the technology for the U.S. Department of Defense, including working on a sonar system that will allow Navy and Marine Corps divers to navigate underwater with their hands free. Raj’s team is also working on an infrared night-vision system for Army Rangers.


Paint-On Laser Makes Faster Chips Possible

By 2010 experts predict that smaller, faster computer chips will exceed the capacity of the electrical interconnects that the chips currently use to transfer data, a problem known as “interconnect bottleneck.” In an attempt to overcome this speed barrier and extend the life of silicon-based circuits, researchers at the University of Toronto have created a laser that can be painted directly on a chip’s surface, introducing the potential that much faster fiber optic connections could one day replace electrical interconnects.

“The current interconnect architecture uses copper to transfer data. With data-transfer rates reaching 10GHz, this architecture is subject to large electrical losses,” says Sjoerd Hoogland, one of the researchers on the project. “Data transfer by light does not impose such a speed limit.”

The laser, developed by Department of Electrical and Computer Engineering professor Ted Sargent, is composed of nanometer-sized semiconductor particles suspended in a paint-like solvent. Sargent and his team created the particles to be just the right size and color to generate light at the infrared wavelength required for optical data transfer when exposed to electrical current.

In their tests, the scientists dipped a glass tube in the laser solution and dried it with a hairdryer, a process that took only about five minutes to complete. When pulsed with ordinary light, the tube emitted laser light. Eventually, the researchers hope to use electronics already found on microchips to power the laser, enabling higher processing speeds, potentially reaching the tens of gigahertz.

While the paint-on laser holds a great deal of promise for the future of the computer chip industry, there are still several open research questions to solve before the technology can be used to produce working chips.

“We are working on electrical pumping of these lasers, which is a necessity for successful implementation for optical interconnects . . . [and] we are aiming to operate the laser well above room temperature,” says Hoogland. “Once electrical pumping and room temperature operation is established, the laser will be ready for mass production.”

by Kristina Spencer