3
Performance Improvements to the Existing GPS Configuration

Introduction

As pointed out in the previous section, civil users of the GPS have accommodated themselves to the currently available SPS (Standard Positioning Service) in attempting to meet their individual performance requirements, and a number of innovative uses of GPS have been demonstrated with the existing system. An even more capable system would likely result in a larger number of applications. Improved accuracy, integrity, availability, and reliability of the signal could provide improved results at significantly lower cost. For example, if the stand-alone GPS could provide an accuracy approaching 5 meters (2 drms), the need for many of the existing or planned differential systems could be avoided.

In accordance with the committee's statement of task, this chapter will recommend a sequence of enhancements to the GPS that will serve to improve the accuracy of the system for civilian, commercial, and military users. After a discussion of the current performance achievable from the basic GPS, the subsequent sections address specific accuracy improvements focused on enhancing civilian, commercial, and military use of the system. Many of the suggested improvements also will have benefits other than better accuracy, such as increased integrity, improved availability, and enhanced resistance to RF (radio frequency) interference. These improved characteristics are discussed where appropriate. The final section of this chapter presents an overall strategy for implementing the recommended improvements. As noted throughout the text, some of the improvements are meant to be applied to the current GPS satellite constellation and others to the Block IIR and Block IIF constellations. When available, the approximate cost of each improvement also is given.



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 67
--> 3 Performance Improvements to the Existing GPS Configuration Introduction As pointed out in the previous section, civil users of the GPS have accommodated themselves to the currently available SPS (Standard Positioning Service) in attempting to meet their individual performance requirements, and a number of innovative uses of GPS have been demonstrated with the existing system. An even more capable system would likely result in a larger number of applications. Improved accuracy, integrity, availability, and reliability of the signal could provide improved results at significantly lower cost. For example, if the stand-alone GPS could provide an accuracy approaching 5 meters (2 drms), the need for many of the existing or planned differential systems could be avoided. In accordance with the committee's statement of task, this chapter will recommend a sequence of enhancements to the GPS that will serve to improve the accuracy of the system for civilian, commercial, and military users. After a discussion of the current performance achievable from the basic GPS, the subsequent sections address specific accuracy improvements focused on enhancing civilian, commercial, and military use of the system. Many of the suggested improvements also will have benefits other than better accuracy, such as increased integrity, improved availability, and enhanced resistance to RF (radio frequency) interference. These improved characteristics are discussed where appropriate. The final section of this chapter presents an overall strategy for implementing the recommended improvements. As noted throughout the text, some of the improvements are meant to be applied to the current GPS satellite constellation and others to the Block IIR and Block IIF constellations. When available, the approximate cost of each improvement also is given.

OCR for page 67
--> Current GPS Performance Accuracy As can be seen from Table 3-1, the contributors to civilian SPS signal accuracy errors are SA (Selective Availability), the atmospheric error, the clock and ephemeris errors, the receiver noise error, and the multipath error. For the military PPS (Precise Positioning Service) signal, the largest error contributors are the clock and ephemeris errors, the receiver noise, and multipath errors, since the PPS signal is not degraded by SA. The ionospheric error for the PPS signal is small relative to that for the SPS signal since the military has access to both the L1 and L2 frequencies and can correct for the ionospheric error. Table 3-1 Observed GPS Positioning Errors with Typical SPS and PPS Receiversa Error Source Typical Range Error Magnitude (meters, 1σ)   SPS with II/IIA satellites PPS with II/IIA satellites Selective Availabilityb 24.0 0.0 Atmospheric Error     Ionosphericc 7.0 0.01 Troposphericd 0.7 0.7 Clock and Ephemeris Errore 3.6 3.6 Receiver Noisef 1.5 0.6 Multipathg 1.2 1.8 Total User Equivalent Range Error (UERE) 25.3 4.1 Typical Horizontal DOP (HDOP)h 2.0 2.0 Total Stand-Alone Horizontal Accuracy, 2 drmsi 101.2 16.4 a. It is assumed here that a ''typical" SPS and PPS receiver has a four-satellite position solution. b. J. F. Zumberge and W. I. Bertiger, "Ephemeris and Clock Navigation Message Accuracy in the Global Positioning System," Vol. I, Chap. 16. Edited by B. W. Parkinson, J. J. Spilker, P. Axelrad, and P. Enge (To be published by AIAA, in press 1995). This error is manifested as increased clock and ephemeris errors when SA is on. c. For the SPS signal, the ionospheric content is quite variable, with large diurnal variations, and large variations over the 11-year solar cycle. Depending on the Total Electron Content (TEC), a delay at L, ranging from less than 1 meter to 70 meters can result. A typical SPS receiver has an algorithm that can remove about 50 percent of the ionospheric error, leading to an error ranging from less than 1 meter to 35 meters. For the above table, an error of 7 meters was used, which is typical for a daytime mid-latitude ionospheric error near the maximum of the 11-year solar cycle, after correction by the standard algorithm. Because the ionospheric error is not independent between satellites, it should not strictly be considered a range error to be multiplied by HDOP (Horizontal Dilution of Precision). When the ionospheric content is uniform above the receiver, such as during the pre-sunrise morning, it contributes little to horizontal error, but maps into errors in the vertical position and receiver clock. When there are significant gradients in the ionospheric content, however, such as exist at local dawn and dusk, errors are induced into the horizontal position. Therefore, the use of 7 meters for a range error, which is multiplied by HDOP, is a somewhat conservative choice. For the PPS signal the ionospheric error is removed by a linear combination of the L1 and L2 observables. This correction leaves residual ionospheric error of 1 centimeter or less.

OCR for page 67
--> d. For a typical SPS or PPS receiver, software models correct for all but around 0.7 meters (la) of the tropospheric error. The tropospheric error is even more highly correlated than the ionospheric error, due to its uniform distribution. The errors introduced by the troposphere normally map into the vertical position and receiver clock errors. As for the ionospheric error, the multiplication of this error by HDOP to obtain the horizontal error is a conservative calculation. e. This value is based on observed data as noted in "Ephemeris and Clock Navigation Message Accuracy in the Global Positioning System." (See note a above). The combined clock and ephemeris error does not contain SA epsilon error in the broadcast ephemeris nor the SA dither error in the broadcast time. f. For a SPS receiver, the receiver noise for independent 1-second measurements can actually range from around 0.25 to 2.0 meters, depending on its design. For a PPS receiver, the single-frequency pseudorange noise error is less because the ten times faster Y-code chip rate overcomes the 3 dB to 6 dB signal-to-noise ratio penalty relative to the C/A code. In forming the linear combination required to removed the ionospheric error, Y-code corrected = 2.55(Y)L1-1.55(Y)L2, the noise error of the Y-code is effectively multiplied by the root sum square of 2.55 and 155, which is approximately 3. (A single-frequency PPS receiver like the Plugger would have a receiver range error smaller by a factor of three, but at the cost of retaining a 7-meter error due to the ionosphere). The PPS receiver noise error can range from 0.1 to 0.8 meters (1s), for independent 1-second measurements. g. For a SPS receiver, multipath can typically range from 0.4 to 5 meters (1s), depending on the antenna, antenna surroundings, and receiver design. For a PPS receiver, the single-frequency multipath error is somewhat less, typically by a factor of 0.5, because of the faster chip rate. In forming the linear combination required to remove the ionospheric error, Y-code corrected = 2.55(Y)L1-1.55(Y)L2, the Y-code multipath error is effectively multiplied by the root sum square of 2.55 and 1.55, which is approximately 3. This explains why the PPS multipath error exceeds the SPS multipath error. (A single-frequency PPS receiver like the Plugger would have a multipath error smaller by a factor of three, but at the cost of retaining a 7-meter error due to the ionosphere). The PPS multipath error can range from 0.3 to 2 meters. h. HDOP can vary depending on the geometry of the satellites. For a typical SPS or PPS receiver, the geometric strength of a four-satellite solution is limited, so a conservative HDOP of 2.0 was used. i. These values are based on observation and differ from the accuracy values specified by the DOD (Department of Defense), shown in Figure C-7, Appendix C. Specific technical modifications to GPS to reduce the errors discussed above and improve the accuracy for both the military and civilian communities are discussed in detail below. As explained in the table notes above, the exact numbers in the tables can vary. If all of the recommendations are implemented, the committee believes that the stand-alone horizontal GPS accuracy will approach 5 meters (2 drms). Greater stand-alone accuracy could take the place of differential GPS systems for some users who require accuracies of a few meters (2 drms). For example, greater standalone GPS accuracy would allow many vehicle positioning and navigation requirements to be met without the use of DGPS. To use a military example, precision weapons, such as missiles and smart bombs that have been equipped with GPS, presently require expensive

OCR for page 67
--> terminal guidance packages or access to differential corrections to meet accuracy requirements of a few meters (CEP). In order to take advantage of GPS accuracy, accurate knowledge of the target location is essential. Various target-determination techniques are being developed, but until they are available, munitions delivery, even with GPS, will still require highly accurate terminal guidance systems. Using an enhanced GPS with greater accuracy for guidance would provide two levels of benefit. First, the requirements on an inertial navigation device can be relaxed because more accurate GPS determination of position and velocity will be possible. Second, under some conditions where jamming is not likely, GPS could be used to provide a very economical and accurate stand-alone munitions guidance system. Integrity and Availability In general, improving the system ranging accuracy also improves integrity and availability. As noted in Appendix C, availability is the percentage of time that a user's positioning errors lie within a specified accuracy. If the ranging errors are decreased, then positioning errors will remain inside the accuracy bounds for higher DOP (Dilution of Precision) values. As a consequence, the amount of time that the system is available increases, especially in the presence of satellite outages. Improvements in availability for the SPS as ranging accuracy improves will be shown throughout this chapter by comparing availability values for Chicago, Illinois.1 Integrity checking algorithms also benefit from improved ranging accuracy. Most integrity algorithms, such as RAIM (Receiver Autonomous Integrity Monitoring), are based on consistency checks among redundant sets of measurements.2 Poor consistency indicates the possibility of a position solution error exceeding the protection limit. However, when the range errors are large, the consistency checks are not reliable except under very favorable satellites geometries. With improved range measurement accuracy, consistency can be reliably measured even under poor user-satellite geometries. Selective Availability And Anti-Spoofing The GPS was designed to provide our military forces with an advantage when engaged with other military forces, while still providing a reasonable positioning service to 1   Chicago, Illinois, was randomly chosen by the MITRE Corporation, which determined the availability values presented in this chapter using a GPS availability model developed for the FAA (Federal Aviation Administration). The analytical model accounts for individual satellite short-term and long-term failures and restorations for the 24-satellite constellation and assumes a conservative serial restoration strategy (that is, only one satellite can be replaced at a time). The GPS receiver was assumed to have an elevation mask angle of 5 degrees. 2   RAIM, which utilizes receiver software algorithms to detect unreliable satellites or position solutions, is defined in Appendix C.

OCR for page 67
--> the civil community. Two features were incorporated into the GPS to provide this advantage. The first, SA, degrades the GPS signal so that the unencrypted signal on L1 will provide a stand-alone horizontal accuracy of 100 meters (2 drms).3 The second, A-S (Anti-Spoofing), provides encryption of the P-code on L1, and L2 in order to deny the signal to the adversary and to increase resistance to spoofing. Selective Availability Currently, the full accuracy of GPS is denied to stand-alone non-PPS users of GPS for both navigation and time transfer through the implementation of SA. SA comprises two functions: (1) fluctuation of the GPS satellite clock frequency, known as dither, and (2) transmission of incorrect ephemeris parameters in the navigation message, termed epsilon. SA affects all GPS observables, which include the C/A-code and P-code pseudorange measurements and the L1 and L2 carrier phase measurements. SA is discussed in greater detail in Appendix C. The DOD has stated that the degradation produced by SA will be limited to a value that maintains the 100-meter (2 drms) specified stand-alone horizontal accuracy of the SPS. Furthermore, at a recent meeting of the DOD/DOT (Department of Transportation) Signal Specification Issues Technical and Policy Groups, additional specifications were discussed and agreed upon for limits on the individual satellite range rate and acceleration errors, shown in Table 3-2.4 Table 3-2 SA Errors from DOD/DOT Signal Specification Issues Technical Group Type of Error Specification Range Rate Bound Not to exceed 2 m/s Range Acceleration Bound Not to exceed 19 mm/s2 Range Acceleration 8 mm/s2 (2a) Under special circumstances, the level of SA errors can be set to zero or increased to a larger value, but only by the National Command Authority. For example, SA was set at a very low level during the Persian Gulf War and during the initial occupation of Haiti 3   The Under Secretary of Defense for Research and Engineering officially established the 100-meter (2 drms) accuracy level for the SPS on June 28, 1983. This policy is reiterated in each biannual publication of the Federal Radionavigation Plan. 4   Report of the DOD/DOT Signal Specification Issues Technical Group to the Policy Group, Washington, D.C., 13 December 1994.

OCR for page 67
--> (it was then set back to 100 meters, 2 drms), because of the lack of PPS equipment fielded by the U.S. military at those times. PPS receivers are able to completely remove the effects of both SA dither and epsilon from their observations through the use of a security module. SPS receivers can eliminate the effects of SA through the use of local or wide-area DGPS broadcasts of differential corrections. DGPS reference stations typically broadcast observed range and range-rate errors. The level of SA-induced range acceleration determines the rate at which the corrections must be updated to keep the user error within acceptable bounds. Satellite position errors produced by the epsilon technique will decorrelate as the separation between the reference station and the user increases. Wide-area DGPS will provide orbit corrections for each satellite to compensate for this effect. Military Utility of SA The DOD has stated that SA is an important security feature because it prevents a potential enemy from directly obtaining positioning and navigation accuracy of approximately 12.5 meters CEP (30 meters, 2 drms) from the C/A-code. Since the military has access to a specified accuracy of 8 meters CEP (21 meters, 2 drms), they believe U.S. forces have a distinct strategic and tactical advantage. With SA at its current level, a potential enemy has access only to the 42-meter CEP (100 meters, 2 drms) accuracy available from the SPS. The DOD believes that obtaining accuracies better than 42 meters CEP requires a substantial amount of effort. DOD representatives have expressed their belief that our adversaries are much more likely to exploit the GPS C/A-code, rather than DGPS, because its use requires less effort and technical sophistication than is required to use DGPS. In addition, some DOD representatives contend that local-area DGPS broadcasts do not diminish the military advantage of SA because they could be rendered inoperative, if warranted, through detection and destruction or by jamming. It is the opinion of the NRC committee however, that meter-level accuracies are readily obtainable, even in the presence of SA set at its current level or even at higher levels. As shown in Figure 3-1, several DGPS systems, operated by both commercial and government entities, routinely provide position accuracies approaching 1 meter (2 drms) in the United States and in most of the populated areas of the world. Further information on commercially available systems is provided in Appendix C. Even within the U.S. government, civilian agencies such as the Federal Aviation Administration, the Coast Guard, and the Army Corp of Engineers are planning to operate systems that will, in combination, cover the entire United States and beyond, as shown in Figure 3-2. Furthermore, if the full GLONASS constellation is completed in 1995 as currently planned, this system also will provide properly equipped users with an additional source of highly accurate positioning data, as shown in Figure 3-3.5 5   Unlike GPS, GLONASS does not deny accuracy to some users through the use of SA or a similar technique.

OCR for page 67
--> Figure 3-1 DGPS coverage provided by commercially available systems, including Skyfix and Sercel. (Courtesy o National Air Intelligence Center)

OCR for page 67
--> Figure 3-2 DGPS coverage provided by the planned FAA WAAS (Wide-Area Augmentation System). Source: Innovative  Solutions International, Inc., presentation at the National Technical Meeting of the Institute of Navigation Meeting, California, January 1995.

OCR for page 67
--> Figure 3-3 Position estimates from GPS and GLONASS obtained from measurement snapshots taken 1 minute apart over an entire  day. Position from (a) GPS with SA off, (b) GPS with SA on, (c) GLONASS, and (d) GPS plus GLONASS.  (Courtesy of MIT Lincoln Laboratory)

OCR for page 67
--> Even if potential adversaries are not taking advantage of DGPS at this time, the NRC committee believes that it would be prudent for the DOD to recognize the potential capability that currently exists. In addition, the establishment of a low-cost, militarily controlled local-area DGPS network for use by an adversary in a theater of conflict is an even more likely possibility. Local-area differential systems are easy to build or buy and are inexpensive. Furthermore, the NRC committee believes that the detection and elimination of these military local-area DGPS stations, either in wartime or peacetime, would be difficult. Local-area DGPS reference stations are small and can be installed in less than an hour. Signals from such systems are difficult to detect because they can be broadcast at low power and at spread-spectrum frequencies or in rapid on/off cycles, with very short transmission times. Therefore, they are not easy to detect electronically or visually. The NRC committee expects that any enemy of the United States sophisticated enough to operate GPS-guided weapons will be sophisticated enough to acquire and install local-area differential system or take advantage of an existing commercial system. These systems can have the capability to provide velocity and position corrections to cruise and ballistic missiles with accuracies that are equal to or superior to those available from an undegraded C/A-code. Even if the level of SA is increased, DGPS methods could still be used to provide an enemy with accurate signals. Further, as previously mentioned, if the full GLONASS constellation is completed in 1995 as currently planned, this system also will provide properly equipped users with an additional source of highly accurate positioning data. The unencrypted C/A-code, which is degraded with SA, still provides our adversaries with an accuracy of 100 meters, 2 drms (42 meter CEP), which would still be more than adequate to deliver chemical, biological, or even explosive weapons, if creating terror in a city is the enemy's objective (see Figure 3-4). Further, any enemy encountered is not likely to share the U.S. military's interest in limiting collateral damage. With SA set at zero, the stand-alone accuracy improves to 30 meters, 2 drms (approximately 13 meters CEP) or better, depending on the solar cycle and user equipment capabilities. While this improvement enhances the ability of an adversary to successfully attack high-value point targets, significant damage also can be inflicted with accuracies of 100 meters, 2 drms. Therefore, in either case (30-meter or 100-meter accuracy, 2 drms) the NRC committee believes that the risk is sufficiently high to justify denial of the L1 signal by jamming. The jamming strategy has the additional benefit of denying an adversary all radionavigation capability including the even more accurate DGPS threat.

OCR for page 67
--> Figure 3-4 Horizontal scatter plot of 42 meters CEP (100 meters, 2 drms) with SA at its current level and horizontal scatter  plot of approximately 10 meters CEP (24 meters, 2 drms) without SA (Figure Courtesy of Mr. Jules McNeff, Office of the Assistant Secretary of Defense, C3I)

OCR for page 67
--> experienced and the movement of the vehicle carrying the receiver. If the L1 signal is jammed, the current receivers cannot acquire the C/A-code and as a result are denied access to the encrypted Y-code as well as dual-frequency ionospheric corrections. One receiver improvement that would enhance military access to the encrypted Y-code in a jamming environment would be the ability to acquire the Y-code rapidly without first acquiring the C/A-code. A method for improving L2 ionospheric corrections in an L1 jamming environment is addressed later in this chapter. In order to obtain direct Y-code access, the signal acquisition processing capability of current PPS receivers must be improved through the use of massively parallel correlators, built using application-specific integrated circuit (ASIC) technology.51 The technology is now available that would allow the incorporation of at least 1,000 parallel correlators per receiver at a reasonable cost. This would allow direct Y-code acquisition within 2 seconds in a non-jamming environment, without prior acquisition of the C/A-code. This would allow faster receiver time-to-first-fix after power-down and would enhance signal availability after a blackout interval.52 The ability to directly acquire the Y-code on L2 would ensure that the selective denial of the L1 signal and the C/A-code through spoofing and jamming would eliminate or seriously degrade an enemy's use of GPS without impacting U.S. capabilities. According to experts in the field of military receiver technology, the technology for direct Y-code acquisition is in hand and in fact, the current military ''Plugger" receivers do try to directly reacquire the Y-code after signal loss.53 A military receiver with the capability to initially acquire the Y-code directly could be developed in 9-15 months depending on: (1) the amount of input received from the military regarding specifications; (2) the level of trade-off accepted between jamming-to-signal ratio versus the amount of time for direct Y-code acquisition; and (3) the ASIC development.54 The development of receivers that can rapidly lock onto the Y-coded signals in the absence of the C/A-code should be completed. The deployment of direct Y-code receivers should be given high priority by the DOD. 51   Massively parallel correlators using ASIC technology, permit the receiver to compare, at very fast speeds, the internally generated pseudorandom noise codes to the received codes, which contain data about the satellite's position and time the code was transmitted. 52   See Appendix K for calculations showing a direct Y-code acquisition time of 2 seconds with current ASIC technology. 53   Personal conversation with Mr. Tyler Trickey, Rockwell-Collins, February 1995. 54   Information provided to the NRC committee by Mr. Charles Trimble of Trimble Navigation Ltd., 31 March 1995.

OCR for page 67
--> Antenna Subsystem Improvements The jamming analysis shown in Appendix L assessed three candidate techniques for improved jamming resistance: aiding of tracking loops with inertial sensors, increased processing gain via wider signal bandwidth, and nulling/directive antenna systems. Nulling antennas were found to provide the single biggest improvement in jamming performance on the order of 25-35 dB. Both military and civilian users have deployed multi-element antenna structures for several years. In the late 1970s, work on the multi-element military AE-1 began. This antenna system was designed to effectively null a single jammer. In both this unit and its derivative design, the AE-1A, element phase shifting and combining is carried out in the radio frequency analog domain. These units have been deployed on a number of aircraft, but have not yet been widely utilized on other military weapons systems, primarily due to cost and size considerations. Recently, more effective techniques have been developed wherein element phase shifting and amplitude weighting is done after spread-spectrum signal correlation, eliminating the RF phase shift components in favor of lower-cost correlator ASIC's and signal processing at baseband. With processing gain applied before nulling and beam forming algorithms, much improved jamming-to-signal margins are available.55 These and future developments aimed at reducing the size and cost of antenna structures should be actively pursued. Nulling antennas and antenna electronics should be employed whenever feasible and cost effective. Research and development focused on reducing the size and cost of this hardware should actively be supported. Inertial Aiding Improvements Aiding can be defined as the usage of any non-GPS-derived user dynamics and clock information in GPS receiver signal-tracking and navigation functions. The availability of such data can have a profound impact on GPS signal acquisition time, code and carrier tracking thresholds, interference and jamming resistance, anti-spoofing capability, and receiver integrity. Aiding works by providing auxiliary observations, which sense a vehicle's motion parameters. Inertial aiding is especially effective because of its invulnerability to electromagnetic interference and because its error characteristics are complementary to those of radionavigation systems, that is, inertial noise errors are low frequency and GPS signal tracking errors are high frequency. Since the earliest days of GPS, the military has exploited synergism, at first with loosely-integrated inertial navigation systems (INS)/GPS built around existing aircraft INS mechanization, and more recently with "embedded" INS in which the inertial sensor and 55   These concepts have been privately developed and patented by the Magnavox Electronic Systems Company (MESC), Patent 4734701. MESC has been continuing to enhance these concepts since inception.

OCR for page 67
--> GPS components reside in the same box. Embedded architectures combine GPS tracking-loop estimates with INS accelerometer and gyroscope outputs to correct INS biases. This provides fast GPS loop-aiding commands for a 10 dB-15 dB increase in tracking threshold and jamming margins and also supports rapid pull-in after signal blockage. This is referred to as a "tightly-coupled" INS/GPS structure. In less sophisticated aiding systems, often referred to as "loosely-coupled" structures, inertial positions and GPS positions or pseudorange data are merged in a cascaded filter structure, missing the benefits of improved GPS signal tracking margins. However, these loosely-coupled INS/GPS structures do extend the length of time that inertial operations can provide useful accuracy and a GPS integrity check, and also speed GPS signal acquisition. Historically, inertial aiding had been too expensive for many tactical military applications. It was not until the 1980s that less-costly strapdown ring-laser gyroscope technology became common aboard military aircraft. However, in the last 5 years there have been other encouraging developments that could lead to wider implementation of aided GPS in tactical military applications. Fiber-optic gyroscopes and solid-state accelerometer configurations have come into use, and more recently, batch-fabricated quartz rate sensors and quartz and silicon accelerometers have been developed. These technologies should have a major impact on the cost of aided receiver systems. The development of low-cost, solid-state, tightly-coupled integrated inertial navigation system/GPS receivers to improve immunity to jamming and spoofing should be accelerated. Signal Processing Improvements The estimation of path delay and Doppler for all satellites in view is the most fundamental task of any navigation receiver.56 Conventionally, delay and Doppler parameter estimates are extracted in delay-lock and phase-lock tracking loops consisting of dedicated loop software and correlator ASIC channels for each satellite. The resulting pseudorange and carrier phase quantities are then fed to the navigation filter routines, wherein these estimates are combined to produce updated position and velocity. In this traditional setup, predating the availability of fast and inexpensive digital signal processing and reduced instruction set computing (RISC) capable of hundreds of millions of double-precision floating point instructions per second, raw correlator data from a given satellite are processed without reference to state and error from other tracking loops, or from the receiver as a whole. Fast computing permits statistically optimal validation and weighting of correlator data from all satellites early in the processing chain, based upon the full receiver state model. By taking advantage of inter-satellite path correlations, and by rapidly adapting filter gains to encountered signal amplitude and noise fluctuations, tracking thresholds can be improved on the order of 10 dB, and tracking can be made resistant to spoofing, multipath 56   Doppler refers to the relative shift of frequency due to satellite-to-user motion.

OCR for page 67
--> capture, and cycle slip. This signal processing approach also can be combined with the inertial aiding techniques above, whereby correlator data, as well as accelerometer and gyroscope data are combined in an optimal fashion. The development and operational use of GPS receivers with improved integration of signal processing and navigation functions for enhanced performance in jamming and spoofing should be accelerated. Improved L2 Ionospheric Correction In a hostile environment where the L1 signal is jammed, PPS receivers will have access only to L2 signals, thus eliminating their ability to perform dual-frequency ionospheric corrections. In such situations, military users must rely on an ionospheric model such as that contained in the broadcast navigation message or on a single-frequency correction process such as DRVID (Difference Range Versus Integrated Doppler). 57 As the GPS signal travels through the ionosphere, the modulating codes (Y and C/A) are delayed by an amount proportional to the inverse square of the frequency. To first order, the carrier phase is advanced by the same amount, producing an effect termed "code-carrier divergence." DRVID is a technique that exploits this effect to compute the change in ionospheric delay over time. In order to determine the total ionospheric delay, an initial delay value must be known. This would work well in a scenario in which a receiver is initialized in a clear environment, that is, outside the region of L1 jamming, and then tracks signals into the jammed region using DRVID to make ionospheric corrections. The primary disadvantage of DRVID is that it relies on continuous carrier tracking, which is not likely to be possible in a high-jamming and possibly high-blockage environment. Currently, single-frequency receivers employ the 8-parameter Klobuchar model that is contained in the broadcast navigation message. This model is considered to be effective in eliminating approximately 50 percent of the total ionospheric delay with a day-to-day variability of 20 percent to 30 percent. It is suggested that enhancements to this model could improve the performance to the 70 percent to 80 percent level.58 Furthermore, based on the current performance of local-area DGPS, the NRC committee believes that local-area estimates of ionospheric conditions made just outside the jamming region could provide even greater improvements. 57   P.F. MacDoran, "A First-Principles Derivation of the Difference Range Versus Integrated Doppler (DRVID) Charged-Particle Calibration Method," JPL Space Programs Summary 37-62 II, 31 March 1970. 58   JA. Klobuchar, "Potential Improvements to the GPS Ionospheric Algorithm." Presentation at the GPS/PAWG Meeting, 14 July 1993, Peterson Air Force Base, Colorado.

OCR for page 67
--> Military receivers should be developed that compensate for ionospheric errors when L1 is jammed, by improved software modeling and use of local-area ionospheric corrections. Possible Interim Operational Procedures The NRC committee believes that the most significant shortcoming of a GPS denial strategy is the current inability to operate in high levels of enemy jamming, while at the same time denying GPS to an adversary. The implementation of technical enhancements to military user equipment, such as direct Y-code acquisition capability, improved nulling antennas, better inertial aiding capabilities, enhanced signal processors, and improved L2 ionospheric corrections would assist in the optimal solution to this problem. Although the NRC committee believes that these technical capabilities are now available, unfortunately, such capabilities are not currently fielded by the military. GPS receivers are especially vulnerable during their signal acquisition phase. This weakness is magnified by the inability of most military GPS receivers to acquire the Y-code during periods when the C/A-code is being jammed. Future receivers capable of direct Y-code acquisition will go a long way toward solving this problem. In any event, tactics must be developed and put in place to facilitate acquisition during jamming. Some possible disciplines that can be implemented in the near-term are presented below. (1)   Develop military procedures to remove jammers and DGPS stations. As with existing plans to destroy radars in a hostile area, plans and procedures should be developed to remove jammers and DGPS stations. (2)   Acquire the Y-code outside the jamming area. Prior to entering the jamming area, the C/A-code can be used to acquire the Y-code. Once the Y-code is obtained, and while still within the active jamming area, PPS receivers should be operated continuously or be re-powered every few hours in order to maintain accurate time. Accurate time will aid in faster, direct reacquisition of the Y-code. This technique can be extended to aircraft-based GPS-guided munitions using low-powered C/A-code retransmissions aboard, or by hardwiring of time-transfer circuits. (3)   Review training exercises, procedures, and policy manuals. The current training procedures and policy manuals should be examined to make sure U.S. troops are properly instructed to operate in both hostile jamming and denial jamming environments. For example, ground forces can make use of natural terrain and man-made obstructions to obtain some shielding from ground-based jammers.

OCR for page 67
--> (4)   Schedule denial jamming/spoofing. Tactically, the U.S. military can interrupt denial jamming/spoofing for short time periods, typically 2 to 3 minutes hourly, to assist those friendly forces in need of C/A-code to reacquire the Y-code. These scheduled times would be short and random to prevent hostile troops from taking advantage of interrupted jamming. Dependence upon this technique will diminish as improved training procedures and time discipline techniques are disseminated into the force structure. (5)   Develop and utilize C/A-code selective denial techniques that minimize impact upon friendly L1-only military receivers, such as the Plugger. The L4 selective denial analysis of Appendix J suggests a four-part approach to selective denial of C/A-code on the L1 band: apply shaped denial jamming combined with spoofing; use a switchable retrofit bandstop RF front-end filter; and, improve clock discipline, through operator training. Improvement Implementation Strategy Because of the relatively long life time of GPS satellites (5 to 10 years) and the length of time required to replace the total constellation of 24 satellites, opportunities for introducing enhancements and technology improvements to the system are limited. Figure 3-6 shows the current plan for satellite replacements. According to the GPS Joint Program Office, current plans for the Block IIF contract include 6 short-term, and 45 long-term "sustainment" satellites. As currently planned, the Block IIF satellites will be designed to essentially the same specifications as the Block IIR satellites. The current program and schedule make it possible for another country to put up a technically superior system that uses currently available technology before the United States can do so. Under current planning, the earliest opportunity for an infusion of new technology in the GPS space segment would be after Block IIF, probably sometime after the year 2020. As noted throughout this chapter, the NRC committee believes that there are significant improvements that could be made to the system that would not only enhance its performance for civilian and military use, but also make it more acceptable and competitive internationally. One method to incorporate technology in an efficient and timely manner is through a preplanned product improvement (P3I) process. With this type of approach, planned changes and improvements could intentionally be designed into the production of the satellites at specific time intervals. Assuming that the first improvements suggested in this report are incorporated in the later half of the Block IIR satellites, additional funding might be required to incorporate changes for the already completed Block IIR satellites. However, the NRC committee believes that the timely improvement in system performance is adequate justification for the additional cost. Recommended improvements to the space segment and the operational control segment are summarized in Tables 3-13 and 3-14.

OCR for page 67
--> In addition to the specific recommendations given in this report, the NRC committee also discussed several enhancements that it believes have particular merit and should be seriously considered for future incorporation. These items are discussed in Chapter 4. Although a few enhancements could be included on the Block IIR spacecraft, especially if a P3I program were implemented, most of the enhancements would have to be incorporated in the Block IIF spacecraft design. Figure 3-6 Current Plan for Satellite Replacement. (Courtesy of the GPS Joint Program Office)

OCR for page 67
--> Table 3-13 Space Segment Enhancements Proposed Enhancement Satellite Block Implementation Date Benefit of Enhancement Turn SA to zero. Block II/IIA, IIR, IIF Immediately Approximately 30- meter (2 drms) stand-alone accuracy for civil users. Add a new L-band signal (would be usable before the Block IIR constellation is complete). Block IIR, IIF As soon as possible Approximately 12- meter (2 drms) stand-alone accuracy for civil users. Enhanced integrity monitoring. Use inter-satellite crosslinks to relay satellite health information and commands. Block IIR, IIF As soon as possible Improve overall system reliability and availability Use inter-satellite crosslinks to relay ground-based integrity monitoring information and commands. Block IIR, IIF As soon as possible Improve GPS signal integrity for all users.

OCR for page 67
--> Table 3-14 Operational Control Segment Enhancements Proposed Enhancement Implementation Date: Benefit of Enhancement Provide more frequent correction updates. As soon as possible Improve stand-alone GPS accuracy for PPS and SPS users (if SA is off and 48-hour embargo is lifted) by reducing combined clock and ephemeris errors by half. Add more monitoring stations. As soon as possible Improve stand-alone GPS accuracy Improve overall system reliability by allowing prompt detection of satellite anomalies. Allow for uninterrupted tracking of all satellites. Improve Kalman Filter and dynamic models. Added to 1995 Operational Control Station Request for Proposal Improve accuracy by reducing combined clock and ephemeris errors with non-partitioned Kalman Filter (15%) and with improved dynamic model (5%). Establish procurement coordination of improved monitor station receivers, computers, and software contracts. As soon as possible and in conjunction with the 1995 contract award Improve accuracy. Allow for integrity monitoring of C/A-code. Reexamine planned Block IIR operation and compare to the accuracy advantages gained by incorporating inter-satellite ranging data in the ground-based Kalman Filter and uploading data at some optimal time. Immediately Possibly improve accuracy over planned Block IIR operation.

OCR for page 67
--> Proposed Enhancement Implementation Date: Benefit of Enhancement Use Block IIR satellite communication crosslinks to the extent possible with the existing crosslink data rate to support on-board satellite health monitoring. Block IIR satellites Improved reliability and availability Permits more rapid response time by the ground station. Use Block IIR inter-satellite communication crosslinks to relay integrity information determined through ground-based monitoring. Block IIR satellites Permits more rapid response time for integrity monitoring. Permanent backup master control station. Immediately Reduce risk and improve reliability of overall system. Provide simulator to test software and train personnel. Immediately Reduce risk and improve reliability of overall system, improve efficiency of operations. Update the operational control segment software using modern software engineering methods in order to permit easy and cost-effective updating of the system and to enhance system integrity. This should be specified in the 1995 OCS upgrade request for proposal. Should be specified in 1995 Operational Control Station Request for Proposal. Easier to make modifications to software. Reduces cost and complexity

OCR for page 67
This page in the original is blank.