REALISTIC ROBOT
By Jean Thilmany

A quite realistic robot—which looks, in fact, like Albert Einstein—at the University of California, San Diego, has taught itself to smile and make facial expressions through self-guided learning.

The researchers used machine learning to help their robot learn to make realistic facial expressions, said Tingfan Wu, a computer science Ph.D. student at the university’s Jacobs School of Engineering.

The Einstein robot head has about 30 facial muscles, each moved by a tiny servo motor connected to the muscle by a string.

Initially, a human being manually sets up the muscles so that the servos pull in the right combinations to make specific facial expressions, Wu said. Then the researcher exposes the robot to about one million photographs of humans showing emotion.

A robot at the University of California, San Diego, forms facial expressions by
using servo motors as muscles. It has learned some expressions on its own.

To begin the learning process, the researchers directed the Einstein robot head to twist and turn its face in all directions. As it twisted and moved, the robot could see itself on a mirror and could analyze its own expression using facial expression detection software created at the university. This exercise provided the data necessary for machine learning algorithms to map the facial expressions against the movements of the muscle motors, Wu said.

After the robot determined the relationship between facial expressions and the muscle movements required to make them, it learned to make new facial expressions that it had only seen in photographs.

For example, the robot learned eyebrow narrowing, which requires the inner eyebrows to move together and the upper eyelids to close a bit to narrow the eye aperture, Wu said.

Although their preliminary results are promising, the researchers note that some of the learned facial expressions are still awkward. One potential explanation is that their model may be too simple to describe the coupled interactions between facial muscles and skin.

“Currently, we’re working on a more accurate facial expression generation model as well as systematic way to explore the model space efficiently,” Wu said.

FORECASTING NEED
By Jean Thilmany

Much as meteorologists predict the path and intensity of hurricanes, Alessandro Vespignani believes scientists will one day predict such things as the economic and social effects of billions of new Internet users in China and India. Or the exact location and number of airline flights that need to be canceled to halt the spread of a pandemic.

Vespignani, a professor of physics and statistics at Indiana University in Bloomington, wrote on modeling and predicting human needs in the July 24 “Perspectives” section of the journal Science. In that article, he proposed that advances in complex networks theory and modeling—along with access to new data—will enable scientists to predict the results of human behavior in areas never before imagined.

Researchers will be able to make predictions because the one wild card in the mix—the social behavior of large numbers of people taken in aggregate—is becoming more definable. Vespignani expects continued progress in data gathering, new informatics tools, and increases in computational power to help make the human group behavior definable.

Researchers have already shown they can track the movement of as many as 100,000 people over six months using mobile phones. They can also define human mobility by looking at currency traffic worldwide, Vespignani said.

This new reality mining, as Vespignani called it, should allow researchers and scientists to forecast the effects of phenomena like catastrophic events, mass population movements, or invasions of new organisms into ecosystems.

“It’s analogous to what happened in physics when we saw the shift from the study of atomic and molecular physics to the study of the physics of matter,” he wrote. “Here we see a movement from the study of a small number of elements, or small social groups, to the study of the behavior of large-scale social systems consisting of millions of people that can be characterized in space—both social and geographic—and in time.”   

A WIDER SHADE OF BROWNIAN
By Harry Hutchinson

More than a hundred years after Einstein did it, a research team has taken a close look at Brownian motion. As you’d expect, given a century of advances in instrumentation, they found something that Einstein missed. According the leader of the research team, a clearer understanding of Brownian motion has implications for applications ranging from barrier coatings to the design of electrolytes in batteries.

Brownian motion is the phenomenon that accounts for the random movement of particles in a fluid. Robert Brown, a Scottish botanist, first wrote about it in the 1820s.

Einstein observed Brownian motion in 1905 and described it as the result of variables including particle mass and the fluid’s coefficient of friction. He determined that the random movements of a particle would fall within a bell curve.

More recently, Steve Granick, Founder Professor of Engineering at the University of Illinois at Urbana-Champaign, and several associates decided to revisit Brownian motion because today’s instruments would give them a closer look than Einstein had.

The researchers conducted two series of experiments in which they used a fluorescence microscope to track the movement of 100-nanometer beads in fluids. In one series, Brownian motion moved the beads up and down tiny tubes of lipid molecules. In the second series of experiments, Brownian motion diffused beads through a porous membrane of entangled macromolecular filaments, naturally occurring large polymer chains found in living cells.

In both sets of experiments, much of the data agreed with Einstein and fell within the bell-shaped curve, but the team also observed unexpected movements of beads outside the prediction of the curve.

According to Granick, who is also a professor of materials science and engineering, of chemistry, of chemical and biomolecular engineering, and of physics at the university, “These large displacements happen less often, but when they do occur, they are much bigger than we previously thought possible.”

They reported their findings in a paper published in the Online Early Edition of the Proceedings of the National Academy of Sciences in the last week of July. Co-authors are Bo Wang, a graduate research assistant who is the lead author; Stephen M. Anthony, a graduate research assistant, and Sung Chul Bae, a research scientist.

“Now that we know the bell-shaped curve isn’t always the right way to think about a particular problem, process, or operation, we can begin to design around it, and maybe take advantage of it,” Granick said. “And, we can correct the textbooks.”

SAND SWIMMING
By Jeffrey Winters

Some 10 percent of the Earth’s surface is covered with desert sand. Animals have adapted to sand’s peculiar properties—a solid that also can flow—in a number of different ways. The sidewinder snake, for instance, moves across the loose, shifting surface with a characteristic S-shaped motion. And certain species of lizards can burrow so quickly that they seem to disappear under the sand.

But for a species of skink native to North Africa and Southwest Asia, sand isn’t something to walk on—it’s something to swim through.

A research team a Georgia Tech has observed that the sandfish, shown here, moves through sand by swimming with a wave motion of its body.

Biophysicists at the Georgia Institute of Technology in Atlanta have studied a skink, known as a sandfish, to uncover just how it moves under the surface. When startled, the four-inch sandfish dives headfirst into the sand, covering itself in an instant. But as a means of evading predators, the sandfish doesn’t just stay in place under the sand, it moves away from danger.

To get a sense of just how it does this, the researchers, led by physicist Daniel Goldman, built a bead-filled glass container. Using X-ray images, the researchers tracked the sandfish as it moved under the sand. Instead of crawling or using its front and hind legs to pull itself through the sand, the sandfish tucked its forelimbs and hindlimbs close to its body. The lizard’s entire body undulated through the sand in a swimming motion, contorting into a very large amplitude wave shape.

What’s more, the speed at which the wave moved along the sandfish’s body determined the speed at which the creature moved through the sand—as fast as six inches a second.

The team’s work may find some application in robotics. Researchers may one day build a robot that can mimic the motion of a skink through the sand, enabling it to move through debris to look for earthquake survivors, say, or though through sand or loose soil for a military reconnaissance mission.

LINKING THE NUMBERS
By Harry Hutchinson

The author of the Mathcad software program, Allen Razdow, has formed a company that is marketing a new technology for embedding numbers in electronic text with metadata.

According to Razdow, people using the technology can elect to highlight and store any number in a database along with associated information including units, unit conversion, and appropriate formatting. A reader of the document would be able to click on the number to access the stored data. Razdow calls the formatted numbers truenumbers and compares them to the hypertext link of a URL or of an e-mail address in text.

The system could be used to store and retrieve engineering numbers, part specifications, or other numerical information. It can also be used so that, when numbers are copied from one source to another, attribution goes with them, allowing for traceability and reuse.

Razdow has named his new company True Engineering Technology. He was one of the founders of Mathsoft, which was formed to sell Mathcad software. Mathsoft is now owned by PTC Corp., and Razdow is no longer affiliated with the company.

True Engineering Technology is selling the system for in-house enterprise use as a dedicated server appliance with client software in a package that costs about $40,000. The price includes licenses for 100 user seats. According to Razdow, there will be a nominal annual licensing fee for additional users that, as of this writing, had not been determined. Razdow said his company’s first customer was the U.S. Navy.

According to Ricardo Garcia, an aircraft structural analyst for the Naval Air Systems Command, “With truenumbers, we will be able to manage and share critical data including design parameters and materials properties, ensuring that anyone on the team can easily get the most up-to-date information available at any time, from anywhere.”

People can try the system for free on a public database. HTML-enabled applications such as e-mail, PowerPoint, Word, or Excel are compatible with the free application, Create a Number at www.truenum.com. Registered users can access and store data in a free, public repository.

GULF STREAM POWER PLAN
By Jeffrey Winters

A scheme to turn the Gulf Stream into a source of power made national news in July.

The program, headed by ASME Past President Sue Skemp, is exploring the idea of putting sea turbines into the Gulf Stream to generate electricity. It was featured by CNN on July 27.

Skemp, the executive director of the Center for Ocean Energy Technology at Florida Atlantic University in Boca Raton, was quoted as saying that turbines positioned in the Gulf Stream could generate between 4 and 10 GW of power. This could provide as much as 30 percent of Florida’s electricity consumption. 

According to the report, the group plans to develop a 20 kW underwater turbine by spring 2010. The turbine will be erected to intercept the water of the Gulf Stream and will be designed so that flowing water will turn the blades to generate electricity, much the way flowing air turns the blades of a wind turbine.

The state of Florida has allocated more than $13 million in grants toward the development of the
pilot project.

Researchers at Florida Atlantic are also exploring whether the electricity generated by sea turbines could be used to produce hydrogen through electrolysis. Hydrogen produced in this way could, in theory, be used to fuel vehicles, replacing gasoline and diesel, which must be brought into Florida from out of state.   
 
The practical use of tides and currents to generate electricity was discussed in the May feature “Renewable Sea Power” by Michael McCormick of the U.S. Naval Academy, and Cengiz Ertekin of the University of Hawaii, and in an online exclusive companion article that month, “To Harness the Seas.”      

BRIEFLY NOTED

Pointwise of Fort Worth, Texas, has upgraded its Pointwise software to version 16.02, revision three. The package now includes an interface to the AcuSolve computational fluid dynamics software package from Acusim of Mountain View, Calif., for CFD mesh checking. Pointwise generates mesh for CFD applications. /// Lattice Technology of San Francisco, developer of digital manufacturing applications using the XVL format, has released version 3.0 of Lattice3D Reporter, for delivering interactive 3-D data in Excel spreadsheets. /// The American Association of State Highway and Transportation Officials has launched a Web site, www.transportation1.org/RealSolutions, to highlight strategies to reduce greenhouse gas emissions from transportation. The site includes best practices, state examples, research findings, and links to other climate change information sites.