By – Gene H. McCall* Los Alamos, New Mexico
Abstract: This paper is a rough, preliminary proposal for a positioning, navigation, and timing(PNT) system using the experience gained from the fielding of large constellations, such as Starlink and Oneweb, and developing 5g communication technology. The result will be a system of 100 low earth orbit satellites, under civil control, separated from the military, which will give a position accuracy of one centimeter, or better, and time accuracy of one nanosecond, or less, at a cost far lower than that of the current GPS program. The system will be practically immune to jamming and spoofing.
I. INTRODUCTION
It is not the purpose of this paper to define, precisely, a Positioning, Navigation, and Timing(PNT) system. Rather, it is intended as an outline that proposes a path to the future of PNT for the United States, in particular, and, even, the remainder of the world. It also defines a business case such that total government funding is not necessary.
The current system that provides the standard of PNT in the United States(US), and much of the world external to the US, is the Global Positioning System(GPS). Although the GPS was the first Global Navigation Satellite System(GNSS), other nations have since either built their own systems or have built augmentations to the GPS that improve services in a limited, or regional, area. Reference [1] gives a good summary of those systems.
Several studies have emphasized the importance of PNT capabilities to a modern nation. Two that are of interest are one done by the National Institute of Standards and Technology[2] which treats the economic benefits of GPS to the United States, and another done by London Economics which estimates the cost of the loss of PNT capabilities[3] to the UK over a five day period. For the U. S., the result was 30bn USD to 45bn USD depending on the time of year, and in the UK, it was 5.2 bn GBP. Perhaps, a good summary is that PNT capabilities can be valued at 1.5bn USD per day in a modern nation.
The value of PNT capabilities is well-known. Many have asserted the need for backup capabilities that could assume the role of current PNT systems if the existing ones were suddenly unavailable. This requirement translates into typical accuracies, as reported by the FAA[4] for 95% of the time as a horizontal accuracy of 1.891 m horizontal, 3.872 m vertical, and time accuracy of 10 ns, with a mean of 2.5 ± 4.1ns.
Many claim that accuracy at this level is adequate at present, and it will remain adequate well into the future. Nothing could be further from the truth. Many important functions, such as farming, could benefit from higher accuracy. One of the most important uses of the GPS is to provide instrument landing capabilities at U. S. airports. The GPS alone is incapable of providing adequate accuracy, and an augmentation system, WAAS, is necessary to provide the accuracy required to enable safe landings in bad weather. Commercial augmentation systems are also available. The Galileo system[1] has a special access service that provides an accuracy of about 40 cm for certain users. Still, the acquisition time required to access the service can be as long as 400 seconds.
In general, the existing PNT systems are inadequate for many important uses.
Recall that the GPS was pronounced fully operational in April 1995, 26 years ago. However, the first satellite was launched in 1978, 43 years ago. The design work was done even earlier. Therefore, in terms of the age of modern technology standards, the GPS has a basic design that is more than 40 years old. It is, thus, an antique in terms of its technology age. Its components are, certainly, modern, and the program engineers have worked very diligently to make fairly modest performance improvements. Even then, many of these improvements benefit the military and not civilian services.
When one adds the extreme sensitivity to interference, natural or artificial, and spoofing, perhaps the analogy of dementia for navigation systems, one realizes that the existing system has almost reached the limits of possible improvements and is inadequate for the nation’s future.
An important example is a self-driving automobile. Positioning accuracy of nearly two meters is inadequate, even for freeway driving, and a weak interference signal could cause a massive pileup, with considerable loss of life. This paper will present some possibilities for significant improvements in accuracy and jam resistance.
A fee-for-service approach will also be described.
II. Antique satellite design
When the first GPS satellites were designed and built in the 1970s, satellite launches were infrequent and frequent failures. It was believed that the most effective use of scarce resources was to build satellites that used the entire payload mass and space of the launch vehicle. Satellites were, mostly, a one-of-a-kind device that should enable as many functions as possible, independent of their relation to one another. Because of the expense, satellite launches were a highly political event. The development of the philosophy that space could be used more effectively if multiple satellites, each with the minimum number of necessary independent functions, were combined in a cooperative constellation was still years away. Although the GPS may have been the first example of modern philosophy, the possibilities appear to have passed by the creators of the GPS without notice. As a result, a heavy, power-hungry system for nuclear explosion detection(nudets) was added to each satellite. The nudet designers appear to have noticed the possibilities of a cooperative constellation, and a system of satellite communication crosslinks at UHF frequency was added to support the nudet function. However, because of noise from terrestrial UHF sources, the crosslinks never operated properly. A separate downlink frequency, L3 at 1381.05 MHz, and telemetry system were also dedicated to the nudets. Even today, nudet designers claim that their support for the GPS system made it happen.
Many in all the services indeed opposed the creation of the system, and opposition also developed in Congress. The GAO wrote three negative reports about the system. In the end, though, the creation of the GPS as a joint military program and delicate bureaucratic maneuvering by Malcolm Currie(DDRE) and Col. Brad Parkinson(USAF) carried the day, and the GPS became
Figure 1: GPS-III assembly area at Lockheed-Martin
a reality. The essay by Lt. Col. M. E. Skeen in the Joint Forces Quarterly(JFQ) is an interesting description of the political history of the system[5]. It is interesting to compare the scales of antique designs, such as GPS, to modern designs, such as Starlink. The GPS III assembly area is shown in fig 1, and the Starlink assembly area is shown in fig. 2
Perhaps, one can imagine why a Starlink satellite costs $1M in orbit, and a GPS III satellite costs $400M.
III. collaborative constellations
The original idea of satellite design and deployment recognized that the use of space that could enable visual access to places on the earth that were, otherwise, denied. Each satellite was a unique vehicle, generally different in many aspects from all others. It has now been recognized, however, that constellations of identical, relatively simple satellites can provide important worldwide capabilities which cover the entire earth simultaneously. Although the GPS system has similarities to a modern constellation, the first system designed intentionally to operate in this way was the Iridium system with 66 identical, collaborating satellites forming a simultaneous, worldwide communication system[6].
More recent, highly publicized examples of collaborative constellations are two, which are intended to provide simultaneous, worldwide internet services. Oneweb, with a large part of its ownership residing in the British government, now plans a constellation of 6372 satellites with a mass of 150 kg at an altitude of 1200 km, fewer than the original 47,844. The current number in orbit is 648.
The Starlink constellation, owned by SpaceX, currently has 1635 active satellites, each with 260 kg at an altitude of 1100 km. Solar power is said to be approximately 6 kW, The Starlink satellites are said to cost $1M each, and the company is producing four per day. They are launched in groups of 60 by a Falcon 9 launch vehicle. Each satellite is equipped with an ion engine to enable deorbiting at the end of life. Six kilowatts is a rough estimate of available solar power based on photographs.
It is apparent that satellites having fairly complex functions can be deployed in low earth orbit(LEO) at nominal cost and high effectiveness.
IV. GPS Backup: ill-defined and inadequate
Defense appropriation bills for several years have proclaimed the necessity for a GPS backup system. Few details were given, and the concept remains poorly defined. Sensing increased funding, various government departments, including Transportation, Homeland Security, and the FAA, have undertaken the task of demonstrating continued PNT capabilities at a level that some would consider adequate. Little mention is made of the fact that GPS is already inadequate for future applications in the first place. Only the Army appears to be taking the limitations of GPS, even at the GPS III level, seriously. At the Global Force Next event sponsored by the Army Association[7], it was said by an Army official that:
Access to situational awareness data positioning on a lethal multidomain battlefield is critical for the warfighter, whether on the ground, in the air, or manning a long-range precise weapon system. Yet, traditional global positioning systems (GPS) are often vulnerable in a hyper-active environment where near-peer threats challenge our military across each of the domains. Military leaders must explore alternative ways of delivering and accessing assured position, navigation, timing (APNT) data to ensure it is always available within even the most contested environments.
Thus, it is recognized, at least by the army, that what is needed is a continuous, protected, and accurate source of PNT. This concept holds, especially for the civil community. It is encouraging that the Army philosophy is reflected in recent Congressional language, although it is identified as being Army and not as a general principle.
It is becoming apparent, therefore, that the term GPS backup has become a vacuous concept that has little meaning.
V. Vulnerability to commercial interests
The community of GPS users, operators, and suppliers was energized by requesting the Federal Communications Commission(FCC) from Ligado networks. Ligado asked for permission to use L-band frequencies adjacent to GPS frequencies for their 5g communications network. It was feared by some in the GPS community that spillover from the Ligado system could interfere with GPS signals. Some evidence was offered to support that claim.
However, after consulting with technical advisors, the FCC voted, unanimously, to grant permission to Ligado[8]. Although there were many complaints, Including some from Congress, The FCC issued a press release confirming the authorization[9].
The decision should surprise no one. The FCC commissioners are lawyers, each with a strong interest in supporting and advancing future communication capabilities, and the FCC mission includes the encouragement of new communication technologies. Although the GPS is very valuable to the nation, as described in section I, it could be argued that the communication industry is more valuable. Also, it was not shown that the Ligado system would definitely interfere with the GPS always, only that there was the possibility of, perhaps intermittent, interference. Apparently, the FCC commissioners did not feel that the technical arguments were strong enough to justify rescinding the authorization.
The World Economic Forum has estimated that intelligent internet connectivity enabled by 5G technology will produce an economic output of 23.3 trillion dollars and create 22.3 million jobs by 2035[10]. Given that the Commissioners are instructed to be biased toward aiding future communications, the value of the GPS does not appear to be an overriding consideration.
Thus, it has been shown that threats to the continuous, proper operation of the GPS are not just technical ones from those who believe there is an advantage to them from jamming, or spoofing, the signal. There is, perhaps, a threat, at least as significant and government-approved, from domestic commercial and political sources.
Given the threats, worldwide and domestic, that can interfere with accurate and continuous PNT capabilities in the United States, and the technical limitations described above, it is clearly time, or past time, to move PNT functions out of L-band and to field a PNT system that performs better than the GPS. One that will also have no problems operating under conditions of nearby jamming, or spoofing, can easily resist adjacent band interference, and has a clear path to future improvements. The basic features of a possible system will be described below.
VI. Performance for the future
Basic performance for the next PNT system should be:
• Position accuracy of one centimeter, or better.
• Timing accuracy of one nanosecond, or better, relative to UTC.
• Capable of proper operation in all foreseeable jamming and spoofing environments
• Capable of taking advantage of future developments in signal generation and propagation
VII. Accuracy Considerations
It will be assumed that the new system will continue to use pseudorandom noise sequences with correlation in the receiver to improve signal-to-noise ratios. It may be appropriate, though, to consider new possibilities in this area. The limiting effect of ionosphere delay on accuracy is well-known. Because the delay is dispersive, and therefore, dependent on signal frequency, the general approach is to measure the delay at two, or more, frequencies, and, then, to use highly accurate models to determine the delay at all frequencies to an accuracy of a few centimeters, or, with more frequencies, even, millimeters.
i. Ionosphere Delay
It is well-known that the free electrons in the earth’s ionosphere change the dielectric constant of the atmosphere[11]. Interestingly, the change is such that the phase of a propagating wave is advanced, and a propagating pulse is delayed by the same amount. This dielectric constant is dispersive with frequency, and the effect can be modeled very accurately by making delay measurements at two or more frequencies[12].
ii. Troposphere Delay
Less well-known is the delay in the earth’s troposphere. The delay depends on two components, the normal delay in the dry air produced by the earth’s atmospheric dielectric constant and the delay produced by water vapor in the air. If both components are known accurately, several models will produce very accurate results with only a centimeter or less error. Unfortunately, the components are not dispersive at L-band frequencies, and, thus, they cannot be measured by sampling wave transmission at, even, several frequencies. Models, which guess at the components can produce an error as large as one meter or slightly more. The components become dispersive at frequencies near the water vapor absorption line at 183 GHz and above[13], and this fact will be used to advantage below.
VIII. frequency selection
Given that 5G technology will enable higher speed, PNT systems should take advantage of 5G technology advances to solve jamming, spoofing, and accuracy. Efficient traveling wave tube(TWT) transmitters with data rates higher than 20 gigabits per second(Gbps) at a carrier frequency of 300 GHz have been designed[14], as have small portable receivers with compact on-chip antennas[16]. Data transmission rates above 22gbps have been obtained.
Although there can be weather issues at these frequencies, the processing gain can be made high enough to overcome those easily.
IX. Signal structure
The direct sequence spread spectrum ranging signal has served well, and it probably should be used in the new system. Perhaps, though, other options should be considered. The primary considerations are that there be no need for additional systems to provide enhanced accuracy or protection. At a bandwidth of 20 GHz, 2 × 1010 Hz, a navigation message at the rate of 1 kHz can be modulated onto the ranging signal to give a processing gain of 70 dB, enough to prevent jamming at the 300 GHz signal each satellite sequence. A nav message of a few kilobits, enough to convey satellite orbit information and clock corrections, can be downloaded in 3 or so seconds.
X. Satellite Configuration
A complete design of the satellites and their constellation will not be given here, but general design characteristics will be presented. . It is easy to show that the fraction of the earth’s surface, f, covered by a satellite at an altitude, h, which uniformly illuminates all the area in view, is:
f = h/(R + h)/2.
Where R is the mean radius of the earth, 6378 km, the GPS satellites at an altitude of 2 0,200 km illuminate a fraction f = 0.38. If one arbitrarily chooses a factor of 10 dB over GPS, for the LEO satellite signal, the LEO altitude is 524 km. That is, perhaps, a bit low. A satellite at 500 km can be expected to have a mean lifetime of 10 years. That is probably, sufficient, but that value indicates a significant amount of drag and orbit perturbation. It would also imply one or more reentries per year from the constellation.
Choosing an altitude of 700 km should decrease the drag by a factor of ten. Thus, an altitude of 700 km will be chosen as the constellation altitude. Other considerations may change this value during the design phase of the program.
Lang and Adams have calculated the number of satellites required to provide full earth coverage of a signal for the required number of satellites in view ranging from one to four[17] using both streets of coverage and a Walker approaches. Either design method gives the number of satellites required for the fourfold coverage required of a navigation constellation to be approximately one hundred.
A significant change to antique satellite characteristics should be the use of chip-scale atomic clocks(CSAC) as the timing reference. The change will make a significant difference in both the mass and the cost of the satellites. At present, the large atomic clocks used in the GPS have higher accuracy and lower drift rates than the CSAC, but the CSAC performance is improving steadily, and the time drift for the orbital period of a 700 km satellite is, even now, only about 2 ns. With adequate monitoring, calibration, modeling, and steering, it will be possible to reduce the error below 1 ns[18]
XI. Satellite Monitoring and Control
It is essential to monitor the status and operating parameters in the constellation and communicate these parameters to the user of the PNT service. If satellites are observed at precisely known locations, the position and time errors can be determined continuously by transmitting the measured values to a central control station. Transmit delays can be measured precisely, and the control station can use an accurate timekeeping method to determine errors in satellite time and orbits.
The number of monitoring stations required for the continuous monitoring of all satellites in the constellation can be estimated by referring to fig. 3.
The fraction of the entire sky in view is determined by the angle θ. The minimum viewing angle above the horizon, α, is usually taken to be 15 degrees. In fig. 3, R is the radius of the earth,
Figure 3: Geometry of extent of viewing from a monitoring station
6378 km and h is the satellite altitude, 700 km. The angle, θ can be determined by solving for β using the law of sines, and, then observing that θ + α + π/2 + β = π. Then, θ = 14.5 degrees. The fraction of the entire sky is, then, (1/2)(1 − cos(θ) = 0.0159. The number of observation points required is 1/0.0159 = 62.7. Thus, approximately 63 observation stations placed at optimum locations will be required. The number is rather large, but all that is required is a secure location, a multi-channel receiver, a processor, and a communication channel to the control center.
The corrections must then be transmitted to the satellites to be incorporated in the navigation message. Existing PNT systems, such as GPS, transmit this information from a ground station that can communicate with all satellites only a few times per day. Recently, however, the inter-Satellite Relay System(IDRS) has been demonstrated[19].The system downlinks data from geosynchronous satellites to LEO satellites. The geo satellite is in continuous view of the ground station, and it is crosslinked to other geo satellites, such that the LEO satellites can be sent data at all times. The possibility of using this system to transmit navigation information to the PN constellation should be investigated.
XII. receiver
A handheld 300 GHz handheld receiver with a bandwidth of 27 GHz, adequate for detecting the ranging signals, has been designed[20] for 5g applications. If the chip rate of the ranging signal is set to 10 Gbps with a one-second-long sequence, the receiver can use an autocorrelation time of one second with a correlation gain of 100 dB to defeat noise and jamming.
Tightly coupling the rf channels to a set of mem accelerometers will give kHz position response with, effectively, a calibration every second. The bias correction of these devices is, now, good enough to provide a one-centimeter accuracy over the one-second integration time. Position can be calculated with a kHz bandwidth. This system will provide, essentially, one-centimeter accurate positioning performance with an update each second.
If the ranging signal is modulated with a navigation message, which gives the satellite orbit parameters, clock offset, and health, at a rate of, say, one kHz, a processing gain of 70 dB is obtained. At the 300 GHz frequency, that should be enough to make the signal unjammable.
XIII. Cost Estimate
The satellites will be, internally, simpler than the Starlink satellites, which are said to cost one million dollars, each, on-orbit. Most of the PNT satellite cost will be absorbed in launch costs. If one assumes one million dollars per satellite, the constellation of 100 satellites will cost 100 million dollars. The geosynchronous link, IDRS, will, perhaps, cost another 100 million, and control and monitoring stations another one-hundred. Add software costs and communication links, and another one-hundred will result. Thus, it should be possible to field the entire system for approximately 400 million dollars. That is, approximately, the single-year cost of the GPS III program.
XIV. A possible Business
The encrypting of the ranging signals at the level used for satellite TV will eliminate spoofing from all but the most sophisticated, technologically advanced tamperers. A major nation could possibly defeat the system, but other methods can be used against such attacks[21].
An encryption system, such as VideoGuardTM, used by satellite television companies to protect programming, would give adequate protection and would enable a low-cost subscription service for the PNT services.
XV. Summary
Existing PN systems, such as the GPS, have severe limitations, which make them unacceptable for use in important applications which will appear in the 21st century. For example, their accuracy is inadequate for self-driving automobiles, and even some existing needs, such as farming, could benefit from higher accuracy. The systems are also highly vulnerable to interference, such as jamming and spoofing. Most of the systems also do not have a clear path to future improvements in accuracy and security. In the case of some, such as GPS, satellite design, and deployment, are controlled by the military, and civil requirements are given low priority. Of course, some threats, such as solar storms, or anti-satellite weapons, can never be entirely eliminated, but their danger can be reduced significantly by making them less effective in destroying system performance or making the use of weapons highly expensive for the attacker.
Some of the needs can be summarized:
- Move out of L-band to reduce interference, and legal, problems.
- Use a single-purpose low earth orbit satellite constellation for higher accuracy, and lowercost.
- Design for one centimeter, one nanosecond performance
- build in jamming and spoofing protection.
- Use 5g technology as it develops.
- Use modern radiation resistant design and construction.
————–
* Gene H. McCall* *Laboratory Fellow(retired) Los Alamos National Laboratory and Associate Fellow, Royal Institute of Navigation