A team of Ohio State University researchers is part of a consortium of universities developing large-scale applications for a material called "smart dust". On Monday, the researchers demonstrated new technology with a network of smart dust sensors. A car rolls slowly between two lines of orange flags scattered across the lawn of the chemical abstracts service building just north of Ohio State University. The crowd of engineers and journalists gathered to watch then turns from the live display on the lawn towards a large screen set up in the shade of a tree, where a cartoon image of a car has appeared between two lines of red points.
Each of these points - or "motes" - represents a small, wireless sensor integrated with a "smart dust" device. The array of these sensors is designed not only to detect movement, but also to distinguish between the movement of metallic objects - such as vehicles and people carrying weapons - and nonmetallic objects, such as civilians without weapons.
The OSU team, led by professor Arora, is part of a consortium of universities developing large-scale applications for "smart dust", a playful name for an innovative technology that uses tiny, inexpensive sensors that relay information to a central computer.
The potential applications for smart dust technology are numerous; researchers have suggested that it can be used to monitor changes in light, temperature, and humidity, to track pollutants, pesticides and gas leaks, and to detect nerve agents and other biological weapons.
While other universities focus on shrinking hardware and developing software, the OSU group is trying to test the limits of networking and communication among the sensors. Arora's team uses a network of "motes" on the chemical abstracts lawn to demonstrate how the sensors can detect two separate "intruders with guns" played by graduate students carrying metal pipes. Determining that an object is a car, rather than a soldier with a gun, requires the combined efforts of more motes.
Prabal Dutta, a graduate student in OSU's department of electrical engineering and a member of Arora's team, says that the motes become a cooperative network, able to detect and track objects moving among them. At that point, Dutta says, a computer a few hundred feet away, or even a satellite link halfway around the world, can detect different types of targets, from people to soldiers to cars. Arora adds that the next step is to give the motes some sense of localization, to know where they are without being told. Telling each sensor where it is could become prohibitively difficult as the number of motes in an array reaches thousands or tens of thousands.