Antijamming and GPS for Critical Military Applications
Anthony Abbott
The Department of Defense is working hard to enhance the jam resistance of its GPS-based systems. Recent research at Aerospace has yielded promising results.
The Global Positioning System (GPS) has become an essential part of the military infrastructure. For that reason, it presents a target for adversaries wishing to undermine the ability of the United States and its allies to conduct military operations. Although the GPS spread-spectrum signal offers some inherent antijam protection, an adversary who is determined to negate a GPS system need only generate a jamming signal with enough power and suitable temporal/spectral signature to deny the use of GPS throughout a given threat area. The reason for this problem is clear: GPS satellites produce low-power signals that must travel great distances to reach the receiver. A jammer, on the other hand, can produce a stronger signal much closer to the receiver, and since signal power diminishes as the square of the distance traveled, the jammer has a distinct advantage.
This vulnerability has been identified as a high priority within the Department of Defense (DOD), and numerous programs have been established to develop near-term solutions for today's potential threats and more extensive long-term solutions for projected future threats. The Aerospace Corporation has been spearheading many of these development efforts.
A generic adaptive-array processing scheme. Signals from the antenna array are prioritized or "weighted" before being combined and processed by the GPS receiver. (View larger image.) |
Traditional Approaches
The first system developed to increase GPS antijam capability for users on the ground or in the air was the controlled reception pattern antenna. This device consists of an array of six antenna elements arranged in a hexagon around a central reference element. The elements are all connected to an electronics box that controls the phase and gain (or complex weights) of each element's output and combines the seven elements into a single output. This signal processing produces an adaptive gain pattern that can be manipulated to place a null in the direction of an undesired signal source. The underlying principle is fairly straightforward: Received GPS signals are rather weak and cannot be detected or measured without a signal-correlation process; therefore, the processing algorithm assumes that any measurable energy above the ambient noise must be a jamming signal, and so it computes the necessary weights to null the source.
Aerospace has been at the forefront of improving the performance and robustness of the adaptive processing algorithms for three decades. Still, certain factors limit the usefulness of these antennas for some vehicles. Controlled reception pattern antenna arrays are physically quite large (on the order of 35 centimeters in diameter) and generally cannot be used, for example, on small missiles that lack the necessary mounting space. In addition, a controlled reception pattern antenna can only counter a limited number of jammers, as it eventually runs out of "degrees of freedom" or antijamming options when the number of spatially distributed jammers grows too great. This is because the array must use at least two elements to null one jammer. Hence, as a rule of thumb, n elements can null n – 1 jammers. Moreover, the antenna must devote a degree of freedom to a jammer regardless of the jammer type (broadband or narrowband). This approach is less effective than other, more advanced processing techniques that can attack a broadband jammer with spatial resources and a narrowband jammer with time/frequency resources.
Various alternatives are being researched as part of the GPS Modernization and Navwar programs. Aerospace is working closely with the GPS Joint Program Office, other federally funded research and development organizations, and the various DOD laboratories to identify several mutually synergistic antijam techniques to meet current and projected threats. The most obvious approach to increase antijam performance is to increase the transmitted power from the GPS satellites. Although the GPS Modernization program will increase satellite power, this approach alone will not provide the entire antijam performance that is required. It is therefore necessary to provide additional antijam capability from the user equipment. Basically, these user equipment techniques fall into two categories: those that reduce the jammer power while retaining or amplifying the GPS signal and those that increase the signal-to-noise ratio through advanced signal processing in the receiver (i.e., processing gain).
Space-time adaptive processing can be effective in combatting multiple jammers. In this technique, the output of each element in a phased array is delayed slightly longer than the one that preceded it. The output of each is available as a separate signal, and each can be processed with a unique weight and combined into a composite signal. |
No one method is right for all circumstances because each application presents its own unique requirements and constraints. Moreover, a given technique may be effective against a particular class of threats, but may not necessarily address all threats. For example, an adaptive narrowband filter is effective against a jammer that has some repetitive or predictable signal structure, but is ineffective against a broadband noise jammer, whose signal cannot be predicted from previous samples. Likewise, spatial adaptive antenna arrays are effective against a limited number of broadband noise and structured signal jammers, but eventually run out of degrees of freedom as the number of jammers increases.
Jammer Signal Power Reduction
Among the advanced techniques for reducing jammer power, the most promising employs a technology that was originally developed for radar, called space-time adaptive processing. With this technique, the output of each antenna array element is delayed using a series of tapped delay lines, each stage of which outputs a version of the input signal slightly later than the previous stage. The output of each tap is available as a separate signal, and each can be processed with a unique complex weight and combined into a composite signal. A close variant of this technique, called space-frequency adaptive processing, performs equivalent processing in the frequency domain.
These techniques show promise because they optimally attack multiple jammers with a coordinated use of spatial and temporal resources. Although space-time adaptive processing and space-frequency adaptive processing can also run out of degrees of freedom, they can counteract many more jammers of various types before reaching their limits because there are n x m choices of weights, where n is the number of elements and m is the number of taps on each element.
Structured interference signals can be removed via time- or frequency-domain processing techniques. The top figure shows the input power spectrum of a GPS signal with four continuous-wave jammers present. The bottom figure shows the output power spectrum of a frequency excision filter developed at MITRE. This processing can be implemented in real time. (The MITRE Corporation) |
A very similar antijamming technique—actually a subset of space-time and space-frequency adaptive processing—is known as adaptive narrowband filtering. Adaptive narrowband filters work with a single antenna element, so they are typically used in applications that lack sufficient space for a spatial antenna array. They are effective against structured interference signals, such as continuous (e.g., sine) waves or pulsed signals, but they are ineffective against broadband interference, which does not have an identifying signature that can be tracked and eliminated. Adaptive narrowband filters can operate in the frequency domain, time domain, or amplitude domain.
As with the controlled reception pattern antenna, conventional space-time and space-frequency adaptive processing systems attempt to minimize measured power under the assumption that any measured power must be a jamming signal. The weakness in that strategy is that the GPS signal may also be attenuated if the processing algorithm does not consider the direction from which the GPS signal arrives. This weakness can be overcome through additional beam steering or beamforming. Although these two techniques attempt to accomplish the same result, they do so by completely different strategies.
Beam steering uses the direction to the desired satellite as an additional constraint on the complex weight applied to each tap output. To perform these calculations, the processor needs to know the direction to the desired GPS satellite and the position and attitude of the host vehicle.
Beam steering is a "precorrelation" technique, meaning it does not require GPS signal detection to compute the phase and gain for each tap on each array element. Beamforming, on the other hand, is a "postcorrelation" technique, meaning it attempts to maximize the signal-to-noise ratio after signal capture. Both techniques maximize the GPS signal while simultaneously minimizing the jammer power for multiple jammers of various types.
Beamforming improves antenna gain in the direction of the GPS satellite. The image to the left shows how adaptive nulling can neutralize jamming signals. The image to the right shows how beamforming works with the nuller to neutralize jammer signals while strengthening satellite signals. (Raytheon Company) |
Processing Gain
The second major antijamming strategy involves processing gain improvement. The GPS spread-spectrum signal derives some inherent jam protection from the "despreading" process, which converts it from a 20-megahertz bandwidth to a narrower bandwidth. Signal power grows stronger as bandwidth is reduced, so for maximum antijam performance, the narrowest possible bandwidth should be used in the despreading process.
Just how narrow the bandwidth can be depends in part on the design of the code and carrier tracking loops used by the GPS receiver and the dynamic operating environment. Recall that a GPS receiver gets a signal from a satellite, generates a local copy, and compares the two to derive range and range-rate measurements. The tracking loops try to maintain a "lock" on the satellite signal by driving the difference in the signals (as measured by the signal correlator) to zero.
In general, greater antijam performance can be achieved by narrowing the bandwidth of these code and carrier tracking loops. Unfortunately, narrow tracking-loop bandwidths imply sluggish response time, and if a vehicle is undergoing high acceleration, the narrow-bandwidth tracking loop cannot keep pace. If the tracking-loop bandwidth were widened, it would be more responsive to high acceleration, but it would not filter the noise as effectively.
In a power-inversion array antenna, the individual elements are geometrically arranged with an interelement spacing of one-half a GPS carrier wavelength. This arrangement is useful for applications where the desired signal is weak and the interference is strong. |
One solution is to aid the tracking loops by supplying information about the vehicle's acceleration and the motion of the satellite to be tracked. This information could be supplied, for example, by an inertial navigation system and the GPS satellite almanac. With this supplemental information, the receiver's tracking loops can anticipate the dynamics along the line-of-sight to the satellite and use a narrow- bandwidth filter to process the fresh outputs from the signal correlators. If the aiding information is reasonably accurate, the bandwidth of the tracking loop can be narrowed because it will only need to track the errors in the aiding information (which vary slowly over time), rather than the absolute motion of the antenna.
The aided tracking loop, with its narrower bandwidth, provides more processing gain and more protection against jamming; however, it's still not enough to thwart a very strong jammer that may be close to the GPS navigation set. The limitations of aided tracking loops are more practical than theoretical: In actual implementation, the aiding information will contain numerous errors.
The most notable errors arise from two sources: imperfect implementation of the aiding data interface, and the inconsistency of the motion between the aiding sensor and the GPS antenna or "lever arm." (In most vehicles, the antenna and the aiding sensors are in different locations, and "lever-arm" compensation must be provided because the GPS antenna is not sensing the same motion as the aiding sensors.)
The first error source, the data interface, exists because traditional receivers are designed to use whatever inertial measurement unit is present on the host vehicle. (An inertial measurement unit—or IMU—is a set of gyros and accelerometers that feed the inertial navigation system in an aircraft or missile.) The GPS receiver and the host vehicle communicate over an asynchronous serial bus, and the designer of the GPS receiver usually does not accept the IMU data without "deweighting" it in some manner. This deweighting process can limit the achieved bandwidth reduction below theoretical levels and hence limit the antijam performance.
The second error source, lever-arm compensation, is unavoidable if the GPS antenna is not located with the IMU. Unfortunately, many factors—such as vehicle attitude, vehicle rotation, and body flexure—prevent perfect lever-arm compensation, even when the IMU is situated in the same box as the receiver. Hence, the bandwidth of the tracking loops must be wide enough to maintain GPS signal lock despite these factors—and this limits the antijam performance. In some applications, such as small weapons, the antenna is naturally close to the IMU and the body is rigid, so the lever-arm compensation is not as significant an error source as it is in avionics applications.
New Approaches
To meet the future challenge of GPS applications that must operate in projected jamming environments, the GPS Joint Program Office is pursuing several promising technologies and a future GPS set architecture that will yield further improvements in antijam performance. Aerospace is actively involved in defining advanced architectures and technologies that will economically provide better antijam performance. Two approaches in particular are generating considerable interest in the field.
Microelectromechanics
With the recent advances in microelectromechanical systems, new architecture concepts that were unimaginable five years ago have now come within reach. One such technology, the microelectromechanical IMU, will have a significant impact on the future design of user navigation sets.
As noted, the best way to reduce the bandwidth of the tracking loops (and thus improve antijam performance) is to keep the GPS antenna and the IMU together, thereby forcing the lever arm to zero. This placement eliminates the need for the lever-arm correction and its associated errors. Of course, when IMUs were first invented, they were very large, and although they've become smaller over the years, they remain large enough to require special attention concerning their placement in a host vehicle or missile. The ability to place an IMU in the same box with the GPS receiver was viewed as a significant step forward. But until recently, no one considered the possibility of embedding the IMU in the antenna itself.
A jam-resistant GPS antenna undergoes testing at the Air Force Research Laboratory. |
That is precisely the thinking now being pursued under the leadership of Aerospace. The cost, size, and performance of microelectromechanical IMUs are improving to the point where they'll soon be good enough to embed in a GPS antenna. This new architecture overcomes many of the factors that prevented the narrowing of tracking-loop bandwidths in older systems. For example, because the IMU would be dedicated to the GPS set, a synchronous interface between the two could be designed with proper attention to interface errors and data latency. In addition, the placement of the IMU with the GPS antenna would make both sensors experience the same motion, so there would be no need for lever-arm compensation with its associated errors.
Although the accuracy of microelectromechanical IMUs cannot compete with more traditional technologies (such as those that use ring-laser gyros), accuracy is reaching a level that is adequate for aiding GPS. Extremely high accuracy is not required if the IMU error sources are reasonably stable because the navigation processing algorithm constantly estimates these low-bandwidth error sources and compensates accordingly.
In other words, it's the short-term stability of these instrument error sources that's important for aiding GPS. And although short-term stability errors can be sensitive to temperature and acceleration, compensation models whose coefficients are calibrated prior to operation can usually mitigate their effects. So, for short periods of time, errors in the microelectromechanical IMU approach acceptable levels for aiding GPS.
It should be noted that the microelectromechanical IMU is not meant to replace the IMU that may be present in the host vehicle. If there is a need for inertial navigation accuracy without GPS, then the microelectromechanical IMU would probably not satisfy that requirement. The microelectromechanical IMU is intended as part of the GPS navigation set (notice that the word "receiver" has not been used), and is present in the GPS antenna regardless of whether there is a need for an IMU by the host vehicle.
Ultratight GPS/Inertial Coupling
Another technology has recently emerged to address the need for antijam performance. This new technique, called ultratight GPS/inertial coupling, is a different method to jointly process GPS and IMU data (see sidebar, GPS/Inertial Coupling). Several organizations throughout the United States have been performing research in this area, either through independent research and development funds or DOD research contracts. Although each approach is unique in its implementation, they all share certain common traits. For example, they all eliminate the code and carrier tracking operations, which are susceptible to jamming even when aided. All use estimated navigation parameters to generate the local replica signal needed to track the satellite signal. All directly use the correlator outputs (i.e., comparisons of the local and satellite signals) to compute the range and range-rate errors for the navigation processing algorithm.
This graph shows the effects of jamming on unprotected GPS performance. For example, at a jammer-to-signal ratio of about 55 decibels, a jammer located about 100 nautical miles from the receiver could jam the GPS signal through a 1-kilowatt signal. At 1000 nautical miles, 100 kilowatts would be required. (View larger image.) |
Aerospace is an industry leader in ultratight coupling. Four years ago, Aerospace began to develop its formulation of ultratight coupling and filed for a U.S. patent. About the same time, Aerospace became aware of similar research being conducted at other companies and other patents that were pending. When the antijam potential of this processing approach was determined, Aerospace was instrumental in obtaining interest at the various DOD research laboratories to fund development programs.
Today, virtually all GPS vendors to DOD have contracts to pursue some sort of ultratight coupling. A milestone was reached in November 2001 when the first official government-sponsored test of an ultratight coupling formulation was conducted at Eglin Air Force Base. The antijam performance was slightly better than predicted. The test results essentially confirmed the performance that had been predicted at Aerospace using simulations. Currently, the Aerospace formulation is being implemented in a real-time computer. One GPS vendor has asked to license the Aerospace formulation, and many other companies are using it for studies.
Summary
Future GPS systems—particularly for weapon delivery—will benefit from the optimal integration of GPS receivers with inertial measurement units and the use of adaptive processing algorithms and antennas that reject unwanted signal interference while maximizing the power of the desired satellite signal. The combination of all these technologies and the associated system architecture will be the blueprint for DOD GPS sets for the next several decades.
Many of the GPS antijam techniques and architectures that will be used in future equipment have roots at Aerospace, which has been the technical conscience of the program since its inception.
To Summer 2002 Table of Contents
