Xona Pulsar: Obvious and Not-so-Obvious Issues of Next-Gen GPS
The first Pulsar satellite had successfully lifted off just several hours before I started to write this text.
The satellite is a demo of a new Xona Pulsar navigation system. The goals of this system are ambitious: to deliver resilient and accurate navigation globally. But a proposed solution is pretty simple: to build GPS retranslators on low Earth orbits (LEO).
Xona is going to launch about 300 satellites. Each sat will transmit its navigation signals to the Earth, just like GPS. But the LEO is much-much closer to the Earth’s surface in comparison with GNSS’s orbits: several hundred km vs twenty thousand km. So, the Pulsar’s signals can have x100 power on the users’ side!
Each Pulsar satellite will have a GNSS receiver for the orbit determination and satellite’s clock synchronization purposes. As the result, the satellites needn’t have any full-size onboard atomic clocks and the system needn’t have wide ground control networks.
Pulsar satellites are not reachable by terrestrial and aviation jammers, so the jammers cannot violate the system’s operation. On the other hand, the satellites are going to provide powerful and encrypted signals, which are hard to suppress too. As the result, Xona Pulsar claims great PNT resilience.
There is some accuracy improvement too. It can be achieved by a PPP technique with shorter convergence and ambiguity resolution time due to fast working constellation changes. But it’s discussible :)
It is claimed, that Starlink, OneWeb, Spire, HawkEye 360, and other New Space companies have prepared technologies, and the Xona’s satellite can be small, lightweight, cheap, and affordable. It’s about $1 million per satellite for manufacturing and $1 million per satellite for launch.
What’s a fascinating idea! But will it be so easy? I can see several technical issues.
Frequency allocation for Radio Navigation Satellite Services
The Pulsar signal plan has not been published, so there are some questions about the Pulsar signal’s characteristics.
Radio frequencies are very limited. Countries are participating in the International Telecommunication Union (ITU) to share the resource. The ITU publishes the ITU Radio Regulations’ (ITU RR) document that describes the utilisation of radio frequencies from 9 kHz to 275 GHz.
The document determines a frequency allocation for Radio Navigation Satellite Services (RNSS) like GPS, GLONASS, IRNSS, Transit … and Xona Pulsar, I suppose. It is satellite, it is navigation, isn't it?
There are several bands allocated for RNSS in the ITU RR for the space-earth direction: several parts of the L-band, and two 20-MHz chunks in the S- and C-bands.
The L-band is popular and it is overcrowded already. All the modern global navigation satellite systems (GNSS) are there.
Theoretically, it is still possible to add a new Pulsar navigation signal to the L-band. But the signal must not decrease any existing service’s performance. So, the possible new Pulsar signal should use free frequencies or be weak enough. If you develop the signal, you should minimize power spectrum densities (PSD) intersections of the new Pulsar signal and old GNSS ones. If you cross them, the new signal PSD should be low enough in those areas. And it is a problem for Xona Pulsar, because, yes, its signal is 100x more powered in comparison with GNSS ones.
How can we solve the problem? We can try to spread the PSD among L-bands. We have about 200 MHz of bandwidth in 1164–1300 and 1559–1610 chunks. So, if the new Pulsar signal has a similar bandwidth, then its PSD level would be about x10 in comparison with signals like GPS L5 or Beidou B3I. But the level is still too high. It will cause significant intersystem interferences.
Another option is the utilization of the S- or C-band chunks. But it will become a serious barrier for civilian commercial applications. In this case, user devices must be equipped with new antennas and new frontends. It will postpone a lot of civilian use cases and revenue.
I guess Xona Pulsar will use a pair of civilian and military signals. I suppose they will choose L-band for the new civilian signal and the S- or C- band for the military one.
The power of the civilian Pulsar signal can be decreased to meet the ITU RR requirements. The signal would have a lot of sub-carriers. You will be able to receive it by your phone with a minimal firmware adjustment.
I guess the military Pulsar signal will be maximized by power. And resilience advantages will be multiplied by miniature controlled radiation patterns antennas on the users side. The ionospheric delay is small on the high frequencies, so it is not necessary to have two carriers.
We can look at the first demo-sat photo and find some clues:
You can see two antenna arrays. The biggest one contains 12 antenna elements those sizes are suitable for the low L-band. The second one is 3–4 times smaller. It could be an S/C band TX-antenna array or a telemetry antenna.
How to receive GNSS signals on the satellite?
In the first stages the Xona Pulsar system relays on GPS and other GNSSs for synchronization and satellites’ trajectory determination. But a satellite is not a regular GNSS user, because it is a beacon too. It is constructed to transmit powerful signals in the same, probably, frequency band.
The conducted power to the Pulsar transmitter antenna is about 20–60 Wt. At the same time, a GPS signal has a level of 0.0000000000000001 Wt. The difference is 17 orders of magnitude! In my experience, you can expect about 20–40 dB isolation between TX and RX antennas, even if the antennas are located on the different sides of the satellite. It is not enough. The Pulsar signal will suppress any GPS signal reception, if you don’t undertake special efforts on the system level.
What can you do? You can try using special signal processing techniques to compensate Pulsar’s signals on the sat’s GPS receiver side. In the case you need a GPS frontend with incredible linearity, it is almost impossible.
The second option is to use another frequency band, but we have already discussed the frequency limitations. You can limit the sat’s GPS receiver by L1 only signals, but they are rather narrow-band.
In the third option, you can use any time multiplexing and frequency hopping for your L-band signal. For example, you can periodically switch off the transmission and use these pauses for GPS signal reception by the Pulsar satellite. The period can be several hundred nanoseconds or microseconds:
Maybe they will choose 1-ms pauses to exclude signals of other Pulsar satellites.
I can imagine the L-band signal with several sub-carriers. In this case, you can switch off the sub-carriers one by one:
These impulse signals have disadvantages. The average signal power will drop. And you will have an issue keeping a constant envelope for the group signal to maximize High Power Amplifiers (HPA) efficiency.
How to keep a circular polarization?
A satellite’s attitude relative to a user is different in cases of GPS and Pulsar. We always see GPS satellite in front, fullface, but we will often see Pulsar in profile.
A Pulsar’s beam should be much, much wider. It’s necessary for ground users and for inter-satellite links.
The wider beam the lower power density, but it is not a big problem. Actual problems are:
- Low satellites elevation,
- Satellite signal polarization degradation.
The low elevations can cause serious multipath and tropospheric errors in PVT observations.
Big nadir angles can destruct Pulsar signal polarization. It can degenerate from right hand circular to linear. And it’s a really bad piece of news. Circular polarization allows GPS to additionally suppress multipath signals up to 20–30 dB. Xona Pulsar can lose this serious advantage or can use non-planar TX antenna arrays to fix the problem. There is another one solution: to increase the number of Pulsar satellites and limit an elevation mask on the user side.
How to deliver keys?
There are two reliable approaches to defend a user receiver from spoofing attacks: space-time signal processing and cryptography. Xona proposes to use the second one.
I guess Pulsar’s signals will be covered by non-periodical sequences like P(Y) or M-code GPS signals. In this case, it is needed to deliver keys to manufacturers of devices and users. And there is a serious risk. You cannot control the cellphone and automotive electronics manufacturers like military ones.
Anyway, it’s much harder to implement any spoofing attack in this scenario.
Conclusion
Global Navigation Satellite Systems were rather boring past few years. No action, and no drama. And I’m happy to see new ideas like Xona Pulsar. The future of the system is not clear yet, and it is going to be interesting at least.
