4
Technical Enhancements for Future Consideration

In the previous chapter, several GPS upgrades were proposed than could provide a stand-alone position accuracy approaching 5 meters (2 drms). Even though such an accuracy would satisfy many user requirements, as discussed in Chapter 2, better accuracy would still be required for many applications, such as Category II and Category III aviation landings, surveying and mapping, all-weather aircraft carrier landings, and some scientific applications.

Since satellite block changes are likely to occur at intervals of 5-10 years, there are a limited number of opportunities to take advantage of worthwhile technical advances and to refine the specifications based on new applications. Because of the anticipated worldwide dependance on the system, the committee believes that it would be shortsighted not to consider significant future improvements that would make GPS more generally useful and forestall the possible development of competing systems.

Below, several options for further GPS improvement are considered. Although the NRC committee determined that the supporting analyses for these options were not carried to the point where specific recommendations could be fully endorsed, the committee believes that the options presented have particular merit and should be seriously considered for future incorporation. Thus, these options are presented as suggestions for consideration rather than as recommendations. First, technical enhancements to improve the overall performance of the GPS for all users are presented; these are followed by enhancements that will benefit specific GPS user groups.

GPS Improvements To Improve Overall Performance

Use of a 24-Satellite Ensemble Clock

Currently, clock offset corrections are determined on the ground and then sent to the individual satellites once a day as they pass over a GPS monitoring station. The Block IIR satellites will have the capability to determine their clock offsets autonomously relative to a space-based ensemble clock and exchange clock information with other satellites via crosslinks every 15 minutes. As a result, satellites can obtain clock information more often than once per day, which should result in a reduced clock error.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 123
--> 4 Technical Enhancements for Future Consideration In the previous chapter, several GPS upgrades were proposed than could provide a stand-alone position accuracy approaching 5 meters (2 drms). Even though such an accuracy would satisfy many user requirements, as discussed in Chapter 2, better accuracy would still be required for many applications, such as Category II and Category III aviation landings, surveying and mapping, all-weather aircraft carrier landings, and some scientific applications. Since satellite block changes are likely to occur at intervals of 5-10 years, there are a limited number of opportunities to take advantage of worthwhile technical advances and to refine the specifications based on new applications. Because of the anticipated worldwide dependance on the system, the committee believes that it would be shortsighted not to consider significant future improvements that would make GPS more generally useful and forestall the possible development of competing systems. Below, several options for further GPS improvement are considered. Although the NRC committee determined that the supporting analyses for these options were not carried to the point where specific recommendations could be fully endorsed, the committee believes that the options presented have particular merit and should be seriously considered for future incorporation. Thus, these options are presented as suggestions for consideration rather than as recommendations. First, technical enhancements to improve the overall performance of the GPS for all users are presented; these are followed by enhancements that will benefit specific GPS user groups. GPS Improvements To Improve Overall Performance Use of a 24-Satellite Ensemble Clock Currently, clock offset corrections are determined on the ground and then sent to the individual satellites once a day as they pass over a GPS monitoring station. The Block IIR satellites will have the capability to determine their clock offsets autonomously relative to a space-based ensemble clock and exchange clock information with other satellites via crosslinks every 15 minutes. As a result, satellites can obtain clock information more often than once per day, which should result in a reduced clock error.

OCR for page 123
--> During autonomous ranging operation, each satellite will form an ensemble from the 14 satellites in view and will compare its offset relative to that ensemble. Further reduction of the clock error could be achieved if the clocks from all 24 satellites were used to create a single ensemble clock, as opposed to the current plan of letting each satellite form its own 14-satellite ensemble. For an ensemble of 14 clocks, the clock error is expected to be 1.1 meters (1σ) after a 4-hour period, as compared with an error of 0.9 meters (1σ) for a 24-satellite ensemble. This is discussed in greater detail in Appendix M. The major advantage, however, of using a single, 24-satellite clock ensemble is not improved accuracy. Rather, it would allow quartz oscillators to be used on some satellites instead of atomic clocks, which are heavier, more expensive, require higher power, and have lower reliability than quartz clocks.1 Since clock offset measurements are made frequently during autonomous ranging operation, the requirements on satellite oscillator stability are greatly reduced.2 Therefore, quartz clocks could replace atomic clocks on at least some of the GPS satellites.3 In addition, since atomic clocks require yearly maintenance, use of quartz clocks on some satellites also would reduce the ground control station workload.4 Finally, the formation of an all-satellite ensemble clock may permit a failed clock in any one satellite to be detected and replaced more quickly and reliably. In order to utilize an all-ensemble of all the 24 Block IIR GPS satellite clocks, the satellite software must be reprogrammed, and supporting ground software must be developed. In addition, further effort is needed to determine the optimal number and combination of quartz and atomic clocks. Reduced Satellite Clock Errors Through Use of Improved Clocks As discussed above, an ensemble reference clock can be used to reduce clock errors, relax requirements for clock stability, and eliminate the need for atomic clocks on some satellites. In order to improve the accuracy and the instantaneous frequency offset further, more accurate atomic clocks must be used on the satellites that will be carrying atomic 1   According to Martin Marietta Astro Space Division of Lockheed-Martin, atomic clocks have been used in the past on GPS spacecraft and have provided a mixed heritage of superb stability and long life in some cases but unexplained premature degradation and failure in others. Each Block IIR satellite will carry two rubidium clocks and one cesium clock. The total cost of all three clocks represents approximately 3 percent of the price of the GPS spacecraft. 2   With two-way time transfer measurements between satellites made every 15 minutes (900 seconds), the predictions need only to be good over this time period. Note that the clock error is the product of clock stability and prediction time. It is the reduction in prediction time from 1 day to 15 minutes that reduces the clock stability requirement by two orders of magnitude and, thus, enables the potential use of quartz oscillators. 3   Since quartz clocks and atomic clocks have different frequency accuracies, their offsets would be weighted when determining a single ensemble time from all 24 satellites, that is, more weight would be given to the atomic clocks in the ensemble. 4   Maintenance on the GPS clocks requires that each satellite is out of service 1 day per year.

OCR for page 123
--> clocks. A candidate future reference clock is the hydrogen maser. For terrestrial use, oscillators based on hydrogen masers have become the standard because they provide the best combination of low-phase noise, acceptable short- and long-term drift, reliability, and cost. 5 Hydrogen masers have been developed for space use, but none have been flown to date. It is possible that hydrogen masers could possibly be incorporated on Block IIF spacecraft, and the feasibility of doing so should be examined. If it appears viable, a research and development program should be initiated to develop a suitable space-qualified hydrogen maser oscillator suitable for GPS spacecraft. Satellite-Based Integrity Monitoring Perhaps the most innovative and promising method of signal integrity monitoring is through space-based monitoring, rather than ground-based monitoring. This capability known as Satellite Autonomous Integrity Monitoring or SAIM, would require the instrumentation of GPS satellites to monitor transmitted L-band signals from each other for accuracy and usability. If an anomalous signal is detected, neighboring satellites could inform the faulty satellite through the crosslinks. The faulty satellite could then autonomously begin broadcasting a code that could not be tracked by users' receivers. At the same time, the faulty satellite could inform the master control station (through the crosslinks) that there is a problem. With SAIM, the response time for commanding the faulty satellite to transmit a nonuseable code to the users after detection of a signal anomaly would be less than 1 second. Such a response time would meet many of the current integrity requirements, including those of the most stringent aviation applications. In order to fully implement SAIM, however, extensive satellite modifications are necessary.6 For example, a new crosslink design concept is required that is based on a CDMA (Code Division Multiple Access) protocol rather than the current Time Division Multiple Access (TDMA) protocol. This new crosslink would transmit the same navigation message observed by users to each neighboring satellite, which could then detect anomalies in the message. Since the required design modifications could be significant, fully operational SAIM would probably have to be incorporated in the Block IIF satellite design rather than the Block IIR satellites already under construction. There is, however, a less extensive modification that could be incorporated in the Block IIR satellites to provide significant interim improvements in integrity monitoring. This modification would consist of the installation of a radio frequency field probe in the antenna near-field regions of the Block IIR satellite, which would monitor the integrity of its own satellite's L-band transmissions. Since the Block IIR crosslink transponder data unit is currently designed to transmit data every 36 seconds, the integrity information derived from 5   Hydrogen masers are used for very long-baseline interferometry, which is used by Earth scientists to monitor tectonic deformations and Earth orientation. 6   Based on information submitted to the committee by Martin Marietta Astro Space Division of Lockheed-Martin, which was reviewed by the Block IIR payload supplier, ITT, 24 January 1995.

OCR for page 123
--> the radio frequency probe could be transmitted to all other satellites in the constellation every 36 seconds. This would not be timely enough to meet many stringent integrity (time-to-alarm) requirements, but would provide much better integrity than is currently available. Increased L2 Signal Strength One enhancement to the existing signal structure (C/A-code and Y-code on L1, and the Y-code on L2) that would improve performance for both civilian and military users is an increase in L2 signal strength. Currently, civilian receiver manufacturers attempt to correct for ionospheric errors through innovative codeless tracking techniques, with varying degrees of success. The chief limitation in the use of these somewhat expensive receivers is that, in order to function effectively, the carrier-to-noise ratio needed on the L2 signal must be considerably higher than that required by a military PPS receiver. In many environments codeless receivers work very well. However, both L2 pseudorange measurement precision and tracking margin for these receivers are considerably worse than for PPS receivers. In vehicle applications where there is signal attenuation due to foliage, the codeless receivers are more prone to signal loss. After loss of signal, the codeless receivers take a longer time to reach a given level of accuracy than a well-designed PPS receiver would. In order to increase the L2 signal strength by 6 dB, some spacecraft modifications must be made.7 If an additional signal is added to the GPS satellites, as recommended in the previous chapter, then civilians would have access to another frequency for ionospheric corrections, so enhancements such as increasing the signal strength of L2 would not be as vital. However, the military benefits obtained with an increased L2 signal strength would not be addressed by adding another frequency, so such an enhancement should be considered on this basis alone. As shown below, the performance of codeless receivers can be improved significantly if the transmitted power of the L1 signal is increased by 6 dB. Cross-Correlation Type Y-Codeless Receivers This receiver recovers the L2 observables by correlating the L2 and L1 Y-codes. A 6-dB increase in the GPS L2 signal causes a 6-dB signal-to-noise ratio increase in the reconstructed L2 carrier phase and pseudorange. This can be exploited in three ways: 7   According to Martin Marietta Astro Space Division of Lockheed-Martin, an increase of 6 dB to 12 dB would require several spacecraft modifications. None are major except for DC power and thermal control, and these changes only become important at end-of-life when specification-to-performance margins will be lower than normally required on U.S. Air Force programs. Other factors such as harnessing, re-balancing, and panel re-layouts need to be assessed in detail but should not be significant problems. If an L4 signal is also added to the Block IIR spacecraft, power sharing will be required, decreasing the amount by which the L2 signal could be strengthened.

OCR for page 123
--> (1)   The same antenna can be used. In this case the signals can be tracked to lower elevation angles. Given a typical gain fall-off from a survey-type antenna, the effective tracking limit would move from 15 degrees elevation to 5 degrees. Tracking to a 5-degree elevation angle is required for WAAS reference station sites and is beneficial for all differential GPS networks. Tracking to low elevation angles is also important when dual-frequency Y-codeless GPS is used on kinematic platforms such as aircraft, where bank angles can reduce antenna gain toward satellites at relatively high elevation angles. In addition, reducing the minimum tracking angle from 15 degrees to 5 degrees will increase the maximum tropospheric signature by about a factor of three. For high-accuracy GPS users who solve for tropospheric delays either to remove it as an error source from baseline measurements or to monitor tropospheric parameters such as water vapor content, the lower elevation tracks give about a threefold increase in accuracy. (2)   In applications where the limiting error is signal multipath originating from reflectors at low elevation, the system designer may decide to exploit improved signal-to-noise ratio by specifying an antenna with more rejection at low-elevation angles. (3)   Under some conditions, ionospheric variations cause a Y-codeless receiver's L2 tracking loop to slip cycles.8 Given an L2 signal with 6 dB more power, the receiver's L2 tracking loop bandwidth could be increased by a factor of two. L2 Squaring Y-Codeless Receivers This receiver recovers the L2 observables by multiplying the L2 Y-code by itself. A 6-dB increase in the GPS L2 signal causes a 12-dB signal-to-noise ratio increase in the reconstructed L2 carrier phase and pseudorange. These same benefits apply to the squaring receiver, with increased effects. Y-Code Tracking PPS Receivers For the military, a 6-dB increase in L2 signal strength would assist in direct Y-code acquisition and would improve the anti-jam margin, especially if L1 was jammed during a conflict. For example, if the power of the L2 signal is increased by 6 dB, then 6 dB in anti-jam capability could be provided to military users. The important parameter is the ratio between the power of the desired signal and the jammer power. Since the latter decreases 8   Personal communication between committee members and Bill Krabill, NASA, Wallops Island, March 1994.

OCR for page 123
--> as the square of the distance, a fourfold increase in signal power will allow the distance at which the jammer defeats GPS tracking to be halved. Military Enhancements Block IIF Signal Structure Military Enhancements If the NRC committee's recommendation to add an additional, unencrypted L4 signal on Block IIR satellites and increase the signal strength of L2 is adopted, the one remaining area in which further improvement might be considered for Block IIF satellites is that of enhanced resistance to RF (radio frequency) interference for military users. To achieve this capability, wider-band signals than currently provided by the present Y-code can be used or the desired signals can be supplied at greater intensity, possibly using spot-beam techniques, which would illuminate an area of conflict. Wide-Band Signals at High Frequency A significant increase (approximately 10 dB) in anti-jam capability could possibly be achieved on the Block IIF satellites by using another wide-band signal, occupying perhaps 100 MHz to 200 MHz.9 Such a broad signal would require that the carrier be at S-band frequency (approximately 2 GHz) or higher frequency. Although moving to a higher frequency would require receiver and spacecraft antennas to accommodate the signal, as well as other modifications, the move to a higher frequency would result in a reduced nulling antenna size and increase its performance. Such a high frequency would also provide increased immunity to the effects of ionospheric scintillation, which can degrade receiver performance when it is present.10 To demonstrate the effectiveness of a wide-band signal against a jammer (assumed to be co-located with a target), calculations have been performed for jammers operating at power levels of 100 watts and 10 kilowatts. (See Appendix L). At these two power levels, code and carrier tracking thresholds were estimated as a function of range from the jammer. For many applications, the key parameter is not the minimum range for loss of signal lock, but the minimum range for acceptable miss distance (range error) at the target. Therefore, the minimum range-to-jammer for a 1-meter range error was also determined. 9   Additional information on wide-band signals is given in Appendix L. 10   Ionospheric scintillation is a phenomenon in which the Earth's ionosphere introduces rapid phase and amplitude fluctuations in the received signals.

OCR for page 123
--> Figures 4-1 and 4-2 are the pseudorange errors as a function of distance for various receiver alternatives described in Appendix L and the two jammer power levels.11 The difference between the narrower-band Y-code and wide-band options is rather dramatic, even on the log-log plots. The most capable system operates below the 1-meter level to within about 45 meters of the 100-watt source. At 1,000 meters, the code tracking error is below the centimeter level. As shown in Table 4-1, carrier phase tracking and code loop aiding are available within several hundred meters of the jammer. The miniaturized nulling antenna with aiding is good down to about 175 meters. Both wide-band options, which are combined with inertial aiding, are substantially more capable than the best performing existing Y-code system. Figure 4-1 Wide-band GPS with a 100-watt jammer. 11   Data generated by J. W. Sennott, Bradley University, Peoria, Illinois.

OCR for page 123
--> Figure 4-2 Wide-band GPS with a 10-kilowatt jammer. Tables 4-1 and 4-2 summarize the results of this exercise. The most significant finding, perhaps, is that with the wide-band signal using unaided tracking and a simple antenna, a vehicle can approach a 100-watt jammer to within 6 kilometers before a 1-meter range error has accumulated. With aided tracking, this range is reduced to about 3 kilometers. For many airborne weapons systems, this is sufficiently close to permit a successful mission when using inertial navigation for the balance of the flight, that is, assuming the worst case scenario in which the jammer and target are co-located. Considering that the size and cost of current nulling antennas may prohibit their use on certain weapon systems, this is a significant finding and supports the notion that consideration should be given to the eventual inclusion of a new, very wide-band waveform.

OCR for page 123
--> Table 4-1 GPS Wide-Band Signal Augmentation Performance with a 100-Watt Jammer System Option Code Status Carrier Telemetry Status   Jammer distance at loss of lock (meters) Jammer distance for 1-meter range error (meters) Jammer distance at loss of lock (meters) Range error at loss of lock (meters) 1. Y-code unaided standard antenna 18,000 90,000 90,000 1.0 2. Y-code aided standard antenna 10,000 35,000 21,000 —— 3. Y-code aided nulling antenna 550 1,000 1,400 1.9 4. Wide-band unaided standard antenna 6,000 6,000 35,000 0.1 5. Wide-band aided standard antenna 3,100 3,100 6,500 0.27 6. Wide-band aided miniature antenna 175 175 450 0.19 7. Wide-band aided null/beamforming antenna 45 45 215 0.19

OCR for page 123
--> Table 4-2 GPS Wide-Band Signal Augmentation Performance with a 10-Kilowatt Jammer System Scenario Code Status Carrier Telemetry Status   Jammer distance at loss of lock (meters) Jammer distance for 1-meter range error (meters) Jammer distance at loss of lock (meters) Range error at loss of lock (meters) 1. Y-code unaided standard antenna —— —— —— —— 2. Y-code aided standard antenna —— —— —— —— 3. Y-code aided nulling antenna —— 20,000 —— —— 4. Wide-band unaided standard antenna —— 60,000 —— —— 5. Wide-band aided standard antenna —— 31,000 —— —— 6. Wide-band aided miniature antenna —— 1,800 —— —— 7. Wide-band aided null/beamforming antenna —— 450 —— —— Spot Beams The advantages of introducing a new, 200-MHz wide-band signal at a higher carrier frequency for coping with a jamming environment were discussed above. While this offers the best technical solution, the difficulty of finding a suitable frequency band and the need to develop a new suite of military receivers to acquire the signal must be considered. An alternative solution to a wide-band signal for improved anti-jam margin would be the use of spot beams. By employing a steerable spot beam on the satellite to illuminate an area of

OCR for page 123
--> conflict, the desired signal power at the receiver could be increased. For example, if a 3-meter, steerable-reflector L-band antenna (or phased-array antenna) could be added to the spacecraft, then a gain of approximately 20 dB would be obtained, which would increase range-to-jammer penetration by a factor of 10.12 While adding a 3-meter steerable antenna to the GPS satellites is a very significant change with attendant complexity, weight, and cost penalties, this is clearly a preferable approach to simply boosting the overall L2 transmitter power. In summary, in addition to increasing the L2 transmitted power, military anti-jam capabilities can be further improved by using a new, very wide-band signal (approximately 200 MHz), a spot beam, or some combination of both. Enhancements For High-Precision Users GPS Transmit Antenna Calibration High-accuracy users of GPS rely on differential carrier phase measurements to obtain millimeter- to centimeter-level results. High accuracies are obtained because for the differential measurements, most satellite-based errors are common mode errors and cancel in the differencing process. One error, however, that does not cancel is the error due to variations of the effective location (phase center) of the transmitting GPS antenna. These variations are a function of the angle to the user, primarily the angle off the GPS antenna array boresight. Satellites that require precise orbit determination, such as Topex/Poseidon, are vulnerable to this error because the satellites view the GPS antenna from large angles off boresight. The maximum boresight angle to receivers on the ground is about 13 degrees, while the angle to a satellite in an orbit at 1,300 kilometers altitude is about 17 degrees. Variations of a few centimeters in the GPS transmit antenna phase center would induce variations of about 10 centimeters in the altitude of the Topex orbit. Even for ground-based measurements, these effects may contribute a small (~10-9 x baseline length) error. Phase variations are expected to be much greater at larger boresight angles. For such applications, knowledge of the transmit antenna phase variations is needed to reliably obtain centimeter, or subcentimeter, accuracy. By measuring the transmit antenna phase center, the error currently limiting accuracy of very-high precision users can easily be eliminated. Since elaborate antenna measurements are already being made prior to launch, it should be relatively simple to make the measurements required to determine the actual phase center.13 12   Since the nominal GPS antenna has a gain of + 11 dBiC L2 and + 13 dBiC L1, at 14.3 degrees off axis, the benefit of the postulated spot beam is about 20 dB. 13   B. R. Schulper, B. L. Allshouse, and T. A. Clark, ''Signal Characteristics of GPS User Antennas," Navigation: The Journal of the Institute of Navigation 41, no. 3 (1994).

OCR for page 123
--> Knowledge of Spacecraft Characteristics Another limitation in the accuracy obtained by precision GPS users is caused by errors in the dynamic model used to predict the behavior of a GPS satellite as it expands and contracts due to the space environment. The current model could be improved if better information was available on the thermal emissivity of the exterior of the spacecraft surfaces, including its solar panels. In addition, telemetry information on temperatures of spacecraft surfaces is needed. For long baseline applications, the limiting error source is usually the GPS orbit error, even when high-accuracy post-processed orbits are utilized. To reach accuracies of 10-9 times baseline length, the orbits must be known to about 10-9 times the distance to the satellites (approximately 20,000 km), or about 0.02 meters. In other terms, an unmodeled acceleration of 10-12 g would accumulate to 0.02 meters after a 12-hour orbit. Unequal radiation of heat from sides of the GPS satellite causes accelerations much larger than 10-12 g. With a minimal amount of effort, the thermal properties of the materials on the exterior surfaces of the Block IIR satellite could be determined. To accomplish this, instruments to measure temperature could be added to the GPS satellites prior to launch, and the data received from these instruments could be transmitted to the ground. This would allow accelerations of the spacecraft surface, which result from uneven heating in space, to be calculated. These accelerations could then be included in models to improve accuracy. In the absence of deterministic models developed through actual measurements, these radiation pressure parameters can only be estimated. Improved L1 Signal Reception at Angles Below the Earth's Horizon In order to support the Block IIR crosslink capability, the specifications related to the UHF antenna on the Block IIR satellites are different from the Block IIA specifications. As a result of these UHF antenna changes, the L-band antenna on the Block IIR satellites will be less symmetrical and will have a narrower pattern. For angles beyond the limb of the Earth as viewed from GPS satellites, this change will probably result in a reduction in the L1 power currently observed with Block IIA spacecraft by approximately 3 dB.14 For spacecraft applications of GPS, this reduction in received L1 power over that currently observed with Block IIA satellites and the narrower antenna pattern could decrease the ability of low-Earth orbit and geosynchronous satellites to receive GPS signals and make GPS-based positioning more difficult from orbit. By increasing the L1 signal power or improving the symmetry of the L-band antenna, spacecraft applications using GPS could be greatly enhanced. 14   Currently, there is no official specification by the Air Force for end-of-life-power requirements beyond the Earth's horizon.