New Sensor Network Sniffs Cities for Bioweapons

Here's the scene: Terrorists release deadly chemicals into the air near a major metropolitan center, creating a killer plume that is drifting with the wind toward millions of unsuspecting citizens. But within minutes, authorities from the president down to the local firefighters know what the chemicals are, where the plume is heading, and how to best cope with the unfolding crisis.

At the same time hospitals throughout the area receive the same information, and nearby jurisdictions are alerted to the probable need for assistance.

And it all happens automatically.

Sound too good to be true? Not to the people at Oak Ridge National Laboratory in Tennessee. They've already developed just such a system locally.

Work in Progress

"It works," says John Strand, a telecommunications expert who is heading up the lab's efforts to promote what it calls "SensorNet" on a national scale.

Chemical, biological, and radiological sensors have been installed in three eastern Tennessee cities — Nashville, Knoxville and Oak Ridge, as well as in New York City, Washington, D.C., and at Fort Bragg, N.C. The sensors are tied into a telecommunications network that collects data from all sorts of places, thus giving users the ability to compare various types of information and limit the false positives that inevitably occur on any early warning system.

The program was unveiled last week for Frank Libutti, a top official with the U.S. Department of Homeland Security, who told Tennessee authorities he would direct his staff to look into the feasibility of establishing a national network patterned after the Oak Ridge program.

Strand is a "hands-on" kind of executive who directed Sprint's development of a broadband communications program, and he is clearly optimistic about the future of SensorNet. He estimates it would cost from $5 billion to $6 billion to expand the program nationally.

But this is clearly a work in progress, so it is being designed in such a way that it can evolve over time as new technology emerges. Not all chemicals, or biological materials, can be detected by remote sensors. Sometimes it takes days, or even weeks, to positively identify a specific substance.

The development of new types of sensors that can, for example, literally sniff out hazardous materials on the basis of smell, is a hot-button subject these days, and labs all over the world are working in that area. Some are furnishing sensors free to Oak Ridge just to test out their effectiveness.

So the goal of the lab is to provide the infrastructure "so if you have a better sensor you just unplug the old one and you plug in the new one," Strand says. "You don't have to change anything."

Piggybacking on Cell Phone Tech

During tests conducted over the last few months, the lab has found a number of bugs in the system, many of which have already been fixed.

"One of the things that bothered me the most was we had a lot of sensor information that was going to people, but it wasn't going to the right people," Strand says. "The first responders were sometimes the last to know."

So SensorNet has been designed to integrate the flow of data with existing networks that serve the needs of people who really need to know. The person who monitors fire alarms in Nashville, for example, now also gets data from SensorNet, so if there's a problem, that person, sitting in a local fire station, is among the first to know.

The idea for the network actually began prior to the tragic consequences of 9/11, when two men at the lab put their heads together and came up with an interesting idea. Jim Kulesz and Dick Reid figured there ought to be another use for the cell phone towers that were springing up all over the place.

Oak Ridge was already developing a new generation of sensors, as were many other institutions, but to be useful they needed to be integrated into a communications network. Cell phone towers seemed a logical choice because the data could be immediately transmitted to wherever it was needed, and the towers tended to be concentrated near population centers.

After 9/11, the idea moved into high gear when it caught the attention of Congressman Zack Wamp, R-Tenn., who wrestled enough money out of Washington to establish a "proof of concept" network. That led to the current program of sensors spread across a significant geographical area, and Strand says it did, indeed, prove the concept.

The project evolved beyond cell towers, putting sensors in the heart of cities, where they are most needed.

"We actually have a system that works and could be deployed," Strand says. That's a bit of a change of pace, by the way, for a laboratory that usually finds itself on the cutting edge of science and technology. Most of this program is based on off-the-shelf technology.

Breath Mint Alert

In a sense it's an idea that is ahead of its time. Strand admits there are a lot of false positives in the system, because even the most advanced sensors can sometimes be fooled by a harmless substance. Even the "essence of wintergreen," which provides the flavor in some breath mints, has set off sensors during trial runs, he says.

That introduces the risk of shouting wolf too often, but Strand argues that even the emergency phone number of 911 gets a fair share of false reports, and it's still very useful.

A network of sensors, however, should make it possible to reduce that problem because the data from multiple sensors should reinforce each other if the threat is real.

"What we are learning is how to pattern the information so that when you see certain kinds of spikes, along with other kinds of information, you can limit the false positives," he says. Does another nearby sensor also detect the presence of the same chemical? Is there a warehouse nearby where that chemical is routinely used? Does the weather sensor show the substance moving into the wind, instead of with it?

Such a program would have applications far beyond the threat of terrorism. Industrial accidents account for most exposures to dangerous chemicals in this country, and SensorNet could detect those as well.

So the folks at Oak Ridge will push ahead. Their immediate challenge, as Strand puts it, is pretty basic to the success of the project.

"How do you take existing capabilities, even though they may not be perfect, and build and design a system that could be rapidly deployed and supported throughout the United States to give us a level of protection that we don't currently have?"

Lee Dye’s column appears weekly on A former science writer for the Los Angeles Times, he now lives in Juneau, Alaska.