MITRE | The MITRE Digest | Small But Powerful: Netted Sensors Fill the Gaps Between Larger Systems

You've probably heard the old saying: the whole is greater than the sum of its parts. Nowhere is this truer than with "netted sensors," an idea that might change the world of military surveillance and reconnaissance.

Imagine a set—perhaps two dozen or more—of small, inexpensive radars strategically scattered throughout hostile territory, perhaps dropped in by parachute or balloon from an unmanned aerial vehicle. Now imagine that the raw data generated by these devices (such as the location of enemy vehicles) could be relayed by internal radio transmitters to a single command center. The pieces would be connected through wireless technology, forming a virtual "net" of radar information. Using specialized software applications, all that information would be drawn together to form a coherent picture. What might we see? Without our enemies knowing we're watching, we can see where and how fast they are going.

In theory, the netted-sensor concept would work well, but will it work in reality? MITRE's Michael Otero, a radio frequency (RF) hardware engineer, has been exploring one aspect of a larger topic by trying to learn whether netted radio frequency RF sensors would be an effective addition to the platform-based surveillance tools available to the military. When might the military use netted RF sensors instead of large, all-in-one devices, such as airborne radars?

There's no dispute that the U.S. military has achieved many victories with its powerful mobile radar systems, such as the Air Force's Joint STARS (Surveillance Target Attack Radar System). These jet-based platforms can spot trucks and troop movements from miles away, as they proved many times during Operation Iraqi Freedom. But they have drawbacks: these resources are limited and expensive, and they can put their crews at risk during battle. Also, they can't be deployed continually, except during active conflicts.

Being on task all the time is an obvious benefit of netted sensors. While they, too, have some notable disadvantages (such as shorter sensing and transmitting ranges) they are also less expensive to purchase and use and they don't involve human operators. For those reasons, among others, the defense community has been interested in netted sensors for some time. However, although the concept is an exciting one and research has been ongoing for a few years, it hasn't really been tested in the field until recently.

"It's always been assumed that networking large numbers of small, inexpensive—possibly even disposable—sensors would work," Otero notes. "But although we knew that large sensors perform well, we still needed to learn if you could get comparable performance from small sensors that are disbursed over large regions for surveillance and tracking."

MITRE believes that netted sensors are so important to our sponsors that we have initiated a three-year innovative research program in the area. According to MITRE Fellow Garry Jacyna, "Wireless sensor networks are one of 10 emerging technology areas that will revolutionize the defense and commercial communities. Mike's work makes an important contribution to this problem by examining where netted sensors can provide the largest tactical payoff."

MITRE's investigation into the benefits of netted sensors began in 2001, and Otero became the principal investigator of the research in 2002. The goal: perform a trade study about netted sensors to learn whether adding small sensors to the military's current arsenal of intelligence, surveillance, and reconnaissance tools is practical and valuable.

This work required developing a sophisticated modeling and simulation testbed—a computer laboratory for investigating the pros and cons of netted sensors. Using MITRE-developed tracking algorithms in combination with government and commercial software applications, the testbed provides three-dimensional, real-time visualizations of netted sensors at work.

"Using an existing database, we created a complete simulation of a mountainous, desert-like region, such as the Army's National Training Center in Fort Irwin, California, with a very accurate depiction of the terrain, including topographical features," Otero explains. "We then developed scenarios of vehicles driving along those roads—even going off-road—and replicated positioning ground-based radar sensors all around. Simulating the radar returns from the moving targets, we processed them through a tracking algorithm and determined that we can track ground vehicles through a network of small radar sensors." Otero's team has successfully demonstrated the simulations at MITRE's annual Technology Symposium.

Otero and his colleagues haven't relied solely on computer simulations for their work, however. The team also purchased some low-cost, low-power RF motion sensors (such as those from door openers) and placed them around the roads and fields near MITRE's Bedford, Massachusetts, headquarters to see if they could duplicate the computer simulation on a small scale. By setting up a simple communications link to detect vehicles and track them as they moved, Otero found they were able to detect vehicles and measure their speeds.

Although he views the experiment as only a limited success, he is encouraged. "We were able to detect the vehicles in real terrain environments, in bad weather and through foliage," he says.

Next, Otero plans on expanding the scope and scale of the simulations his team runs, as well as seeing how many netted sensors constitute "enough."

"In tests, we've used up to 15 sensors in about a five-by-five kilometer region," he says. "So far, we're not running extremely dense sensor scenarios, but we can definitely scale it up. We've also been reducing the numbers and seeing how the tracker performs. If you have only three sensors operating, for instance, does the system still work? Sensor density is one of our main parameters."

A Simulation Tool with Real Benefits

An additional benefit to Otero's research is the simulation tool itself. Researchers from other MITRE programs have begun using the modeling and simulation testbed developed by Otero and his colleagues. One of these teams is working on an Air Force-sponsored project that simulates airborne radar data, such as those collected by unmanned aerial vehicles.

"We can run simulations in real time," says Otero. "It's a very powerful tool—a unique capability. For example, if someone needs a simulation for a specific terrain, as long as it's available in our database, we can run it. Right now, we're only using radar simulations, but this has plenty of potential to be upgraded for other kinds: seismic, acoustic, magnetic, infrared. We believe this tool has tremendous future potential."

"Without making sweeping conclusions, we've demonstrated with ground-based sensors that we can merge the information from multiple sensors and establish the tracks on vehicles," Otero says. "And this worked in some tough terrain conditions, so it seems practical to detect vehicles this way. Of course, a lot of elements are scenario dependent, including cost per sensor and other factors. That's one of the advantages of our testbed—we can simulate different circumstances."

Although it's highly unlikely that netted sensors will ever replace large surveillance platforms, the evidence of their benefit in certain situations is growing. The future of sensing for the military may involve a combination of many small sensors netted together and working in collaboration with a small number of large sensors. Awareness of the concept is building outside the military as well: the academic and commercial worlds have shown increasing interest, specifically for security and surveillance work.

"One sophisticated, powerful sensor, such as Joint STARS, can give you the whole picture—detect and track vehicles," Otero says. "When you use these smaller, less expensive sensors, the quality of information you get from each sensor is poorer, but the idea is that if you have a sufficient number of them you can use mathematical algorithms to extract more information than any individual sensor could give you.

"Netted sensors could be useful in a lot of scenarios. For instance, you can't always use a high-value asset like Joint STARS in a potentially hostile situation. But you could put a number of small netted sensors in place in a sleep mode. When something of interest comes by, they can 'wake up' and relay the information. But otherwise they can stay quiet and not alert anyone to their presence. They have a lot of possibilities."

	Solutions That Make a Difference.® Copyright © 1997-2013, The MITRE Corporation. All rights reserved. MITRE is a registered trademark of The MITRE Corporation. Material on this site may be copied and distributed with permission only.
	Privacy Policy \| Contact Us