 |
 |
 |
 |
| |
|||||
|
|
|
Home > News & Events > MITRE Publications > The MITRE Digest > | |||||||||||||||||||
Keystoning Improves Radar Resolution for Tracking Targets March 2008
An algorithm invented eight years ago to identify moving targets on the ground is now being used in an advanced prototype radar system that tracks moving targets in urban environments. Known as the "Keystone algorithm," the MITRE-developed idea stands poised to become a crucial element of military radar systems used for information gathering or situational analysis. Dick Perry originally created the Keystone algorithm to sharpen radar images of moving vehicles such as antiaircraft missile launchers, troop-carrying trucks, and supply vehicles. At the time, radars that tracked moving targets required that estimates be made of the target's motion before any useful images could be obtained. Otherwise, without estimating the target's motion, the resulting radar image would be smeared like a moving subject in a photograph when you use a slow shutter speed. And smearing a radar image drastically reduces or obliterates its potential military use. The algorithm works as a building block in a system to detect low cross-section targets—targets that are difficult for radar to pick up. The "Keystone" name comes from the shape of the mathematical function, which looks like the wedge-shaped stone in an arch. Although the algorithm is basically simple and effective, radar manufacturers were slow in embracing it when Perry developed it eight years ago. "They didn't think it was needed for the kind of targets they were tracking back then," he notes. "At the time, only the Israelis expressed interest because they thought the algorithm might be useful for a radar system they were developing to detect mortars." Now, the Keystone algorithm is being referenced in newer radar text books, proposed for advanced radar systems, and implemented in radar test systems. MITRE has years of experience and success in the fields of synthetic aperture radar, signal processing, and ground moving target exploitation. Perry, a senior principal engineer, had been working on the challenge of improving target identification as MITRE-sponsored research. "I was fooling around with the problem and every time I did, I came out with the same answer," he says. "In layman's terms, the algorithm freezes the target so that it doesn't move. Before, you were always limited by the target's motion, which causes blurring. Freezing the target extends the radar's 'dwell time' so that the target can be easily examined to get good high resolution imagery." SAR and Doppler Radars The Keystone algorithm can be used by different types of radars such as synthetic aperture radar (SAR) and Doppler radars. SAR is a form of radar in which sophisticated post-processing of the data produces a radar image of the ground. It can only be used by moving instruments over relatively immobile targets. A Doppler radar produces a velocity measurement as one of its outputs. "An early problem with SAR and Doppler radars was that they couldn't automatically track a moving target for less than a second without causing blurred images," says Perry. "The movement causes the blurring. But the Keystone algorithm remaps the radar data to freeze the target's image so it doesn't move, and the radar can look as long as it wants to get very high Doppler resolution. "The Doppler resolution increases with dwell time while the range resolution can be increased by using more radar bandwidth. Because the target motion is frozen, there is no limit on increasing the range resolution to provide both high range and Doppler resolution simultaneously. This was not achievable prior to the Keystone algorithm—you either had one or the other, not both." Tracking Multiple Moving Targets "MITRE is currently applying Keystone processing techniques through simulations to track moving targets that are performing difficult maneuvers in dense and cluttered environments, such as urban areas," says David Zasada, a senior principal sensors systems engineer and MITRE's site leader in Rome, N.Y. Zasada is managing the Air Force-supported research project, which is in its third year and shows promise for implementation in the field. "The thing about the Keystoning," says Zasada, "is when you do a really good job of focusing on targets on the ground, you do an even better job of blurring the moving targets further, which is not normally desirable. Most people who try to form a precise SAR image view moving targets as something to be eliminated, rather than processed. So all of the techniques that people use to clean up SAR images and make them really sharp and clear basically don't help a bit. Another advantage of the MITRE process is that it uses only half the computational power of conventional radar systems because fewer mathematical transformations are required. "In 2007, we used Keystoning as the basis for designing a new suite of methods that rely heavily on advanced signal processing techniques," says Zasada. "These techniques can be used prior to target detection rather than in the post-detection data processing that's typical of state-of-the-art tracking programs. We expect this suite of methods to provide more efficent target-tracking and help avoid mis-identifying targets." Future research with the Keystoning algorithm will include range-Doppler interferometric phase tracking. This is where two antennas side-by-side send out and receive a signal back from a target. The time delay between the two reflected signals is used to give the precise location of the target.
—by David A. Van Cleave Related Information Articles and News
Technical Papers and Presentations
Websites |
||||||||||||||||||||
| Page last updated: March 11, 2008 | Top of page |
Solutions That Make a Difference.® |
|
|
