SANTA CLARA, Calif. If Andrew Berlin has his way, Intel Corp. could become known for advancements in medicine as well as microprocessors. The manager of a small but ambitious biotechnology lab was one of about 30 researchers who shared their work in areas that ranged from robotics to sensor networks as part of what was billed as a coming out party for the company's two-year-old research division.
Berlin's 12-person lab hopes to publish results in about six months of its work trying to identify individual molecules in a blood sample to help predict the onset of disease or track the effectiveness of certain drugs.
"We are developing types of electrical and optical probes to interact with molecules and identify them. If we could find the molecular or optical signatures of diseases that could be a huge benefit to medicine," said Berlin.
At an open house event here Wednesday (March 19), Berlin showed advancements in 3D micro-fluidic devices which could help isolate individual molecules from a blood sample as the first step to identifying them. "You need to be able to make sense out of what might be 100,000 molecules in a given sample," he said.
Berlin helped start research in microelectromechanical systems (MEMS) at Xerox PARC in the early 1990s and became Intel's first life sciences researcher two years ago. "We are looking at the intersection of biology, medicine and nanotechnology. This is a community that hasn't really formed yet," he said.
As part of a separate program, Intel has become the first commercial participant in a three-year-old U.S. government push to set software standards for robotics.
The Robotic Engineering Task Force, modeled after the Internet Engineering Task Force, joins government researchers from NASA, the Defense Advanced Research Projects Agency and the National Institute of Standards & Technology with university counterparts to establish standard software protocols and interfaces for robotics.
The RETF hopes to publish its work as open source code when it is complete in about two years.
"This has been driven by NASA robotics applications, but the intent is that it would have applicability to a larger set of robotics problems. Down the road we would like to see other commercial firms involved as users or participants in defining the software," said a government researcher close to the project.
Intel has defined a hardware platform based on its Xscale processor for the RETF effort which it is supplying as a low-cost base for future government and university robotics projects, said Jim Butler who manages the robotics research program at Intel.
The biotech and robotics programs are examples of the Intel research group's goal of exploring new growth fields for the PC giant. Perhaps the biggest overarching theme at the division is the move toward ubiquitous computing, an era beyond the PC in which people will interact more naturally with hundreds of computing devices more or less hidden in their everyday environments.
David Tennenhouse, Intel's director of research, said ubiquitous computing is becoming a reality with the rise of 802.11 networks, Web services based on XML and the variety of handheld devices people carry with them.
This vision is leading Intel to work in low-cost sensor networks that can monitor and report on changes in everyday environments. Such sensor nets will feed tomorrow's computers with real-time data on which they can make decisions automatically thanks to developments in machine learning.
"We really are at the point where a machine can give you a list of theories for a given collection of data. It still requires a human to determine which one is best and why. But the next challenge is to actually let computers make the decisions while we supervise them," he said, describing advances in a ten-year-old effort to create statistical models for machine learning.
"To some extent we have just started to apply information technology. These activities will fuel our business going forward," Tennenhouse said.