GPS-aided weapons such as the Joint Direct Attack Munition (JDAM) family of bombs have revolutionized warfare, dramatically improving accuracy and cost effectiveness, while significantly reducing collateral damage. Augmenting artillery rounds with GPS aiding can furnish similar improvements. Unfortunately, the rapid rotation on launch of many shells and rockets complicates GPS-aided guidance. This rotation can cause amplitude and phase modulation of the GPS signal, reducing navigation performance and effectiveness of traditional anti-jam systems. Further complicating matters, traditional attitude determination techniques, utilizing inertial sensors, tend to develop significant errors at high roll rates.
To address the special navigational needs of rotating vehicles, my colleagues and I have developed a suite of GPS technologies known as Advanced Spinning-Vehicle Navigation (ASVN). These technologies enable high-performance, interference-robust GPS navigation as well as roll-angle determination on both rapidly spinning and slowly rotating vehicles. Potential applications include artillery shells, rockets, missiles, space vehicles, and unmanned aerial vehicles (UAVs).
Bottom-Line Drivers. ASVN provides an innovative, low-cost, compact, and robust solution to both rotation angle determination and the navigation of rapidly rotating vehicles. ASVN can:
* Improve jamming immunity by using interference to aid navigation and guidance;
* Increase range by delaying deployment or reducing the required authority of control fins and actuators;
* Reduce cost, size, and weight by eliminating or decreasing the requirements of the inertial sensing system and simplifying antenna design; and
* Enhance accuracy by enabling early GPS acquisition for improved estimation of vehicle trajectory and inertial measurement unit errors.
ASVN Technology Suite
ASVN uses the amplitude and phase modulation of the signals received by a GPS antenna on a spinning vehicle to track both vehicle rotation and improve GPS jamming and interference immunity. Analyses and laboratory and field hardware simulations illustrate the viability of ASVN technologies to reduce costs and improve the performance and capabilities of a range of spinning and rotating applications. Four core solutions comprise ASVN technology:
GPS Roll Angle Determination. Using the GPS signal modulation with rotation to measure vehicle roll angle;
Interference-Aided Navigation. Leveraging a jamming signal or interference source to aid vehicle navigation;
Temporal Beam Forming. Improving jamming resistance with only a single rotating antenna element; and
Coriolis Pitch- and Yaw-Rate Sensing. Providing inertial aiding (including rotation rates) without the use of gyroscopes.
GPS Roll-Angle Determination
In rapidly rotating vehicles (such as artillery shells and missiles), high roll rates complicate attitude determination using conventional inertial sensing techniques. Gyroscope scale factor inaccuracies trigger small roll angle errors with every rotation. At high spin rates, these errors rapidly accumulate into large errors in the vehicle attitude estimate. Added sensors such as magnetic detectors can provide an absolute roll reference, but the performance of these systems depends on the geometry of the flight trajectory with respect to the Earth's magnetic field and the magnetic properties of the vehicle.
ASVN GPS roll-angle determination (GRAD) uses the GPS signal as a reference for roll determination. The diverse geometry of the GPS constellation provides a very trajectory-independent roll reference. Because a GPS receiver typically is present in guided munitions for position determination, no additional sensors are required. The GPS signal is very weak and buried in the background noise, rendering traditional radio frequency (RF) direction-finding techniques ineffective for GRAD. However, rotation demodulators controlled by a roll-angle estimate and GPS correlators help observe the signal.
Figure 1 shows a block diagram of the GRAD receiver with rotation demodulators preceding the GPS correlators. Figure 2 shows the amplitude and phase modulation measured for an example antenna, as well as the rotation demodulator amplitude and its phase correction centered on the rotation modulation. When the estimated roll angle aligns with the actual vehicle roll angle, as shown in the figure, the rotation demodulator corrects the phase modulation and passes the largest portion of the GPS signal amplitude on to the carrier demodulator.
Two more signal-processing channels provide feedback for rotation tracking. The roll-angle control for one channel is advanced ahead of the centered roll-angle estimate and the other is retarded behind it.
When the roll estimate is properly aligned with the true roll angle of the vehicle, the correlation magnitude from the centered channel will be maximized, and the advanced and retarded channels will be reduced in amplitude and approximately equal. The rotation-angle-tracking servo loop can accurately maintain alignment of the centered channel with the actual roll angle of the vehicle by controlling the angle estimate to keep the advanced and retarded magnitudes equal.
GRAD Demonstration. Creation of a benchtop RF rotation simulator allowed for modulation of the amplitude and phase of a GPS signal, received by a static rooftop antenna, to simulate the signal output of a rapidly rotating GPS antenna. The RF rotation simulator can generate rotation rates from 0-300 revolutions per second, and simulate any arbitrary antenna pattern. Use of a breadboard ASVN receiver and the RF rotation simulator enables the demonstration of ASVN rotation tracking in real time.
Figure 3 shows a servo loop capture test. At the start, the system operates with the tracking servo disabled, with an offset of 250 degrees per second between the rotation rate estimate and the simulated roll rate. Note the relative phases of the advanced, centered, and retarded channels. At the time indicated in the figure, the rotation-tracking servo was enabled. The servo quickly corrects the error in the rotation rate and angle estimates, thereby equalizing the advanced and retarded correlation magnitudes (|[S.sub.A]| [congruent to] |[S.sub.R]|). This maximizes the center channel correlation magnitude (|[S.sub.C]|) (see Figure 1).
Benchtop testing demonstrated rotation-angle tracking at spin rates from 50-300 revolutions per second (18,000-108,000 degrees per second). Loop capture and tracking dynamics were good, with typical roll-angle acquisition times under one second, and the ability to track instantaneous steps in rotation rate exceeding 1 revolution per second. Angle-tracking-error noise was low at less than 1 degree per square root Hz. This performance level is more than adequate for a typical artillery shell or missile application in which the angular acceleration profile is relatively predictable, and the roll-angle estimate only needs to be within a few degrees to properly coordinate steering commands.
The Ordnance and Ground Systems division of Alliant Techsystems (ATK) helped validate these benchtop test results. ATK developed a mock-up antenna ground plane representative of the shape of a typical artillery shell application and compatible with its spin fixture.
An off-the-shelf GPS patch antenna (designed for use on a flat ground plane) saved time and money during the test. The directionality and efficiency of the resulting antenna was sub-optimal, but adequate for the purposes of the spin demonstration.
The research team measured the phase and amplitude characteristics of the mock-up as a function of both roll and pitch angles, and programmed the ASVN bread-board system and the RF rotation simulator with antenna characteristics.
The RF rotation simulator served to initially characterize the breadboard system. Later, the team connected the ASVN breadboard system to the ATK spin fixture for live testing, performed for two days (July 30-31, 2003) at rotation rates from 100 to 250 revolutions per second. The testing team achieved rotation tracking on the first attempt, and no modifications to the breadboard were required for testing. Servo performance with mechanical spin fixture matched closely that measured with the benchtop simulator, helping validate integrity of the benchtop RF rotation simulator and its utility in system integration and performance evaluation of new antenna designs.
Figure 4 shows the rotation frequency measured for both the spin fixture and the ASVN breadboard. The receiver frequency estimate closely follows the actual spin rate.
