Low orbit navigation system: pitfalls

An hour before I sat down to write these lines, the first demonstration satellite of the Pulsar low-orbit navigation system separated from the booster block.

I outlined the idea of ​​the system in a previous article. Xona plans to put about 300 small satellites into low orbit. Each satellite will be equipped with a global navigation system (GNSS) signal receiver, and the satellites will use their signals to determine their trajectories and synchronize. Due to their distance from the earth’s surface and the orientation of the antenna, the satellites will remain out of reach of terrestrial jammers.

Each satellite in the system will generate its own signal, which, due to its proximity to the surface of the planet, will be 100 times more powerful than GNSS signals. Together with cryptographic signal protection, this promises navigation resistance to interference and spoofing attacks. And due to the simplicity of satellites and their output, the simplicity of the control complex, the cost of the entire system can be placed in several hundred million dollars.

Sounds promising, but what are the pitfalls? The structure of the signals and their carrier frequencies are not yet published by the company, and here I have many questions.

RNSS frequency allocation

The frequency resource is finite and limited. In order to somehow divide it and streamline the use of frequencies, countries participate in international telecommunication union (ITU, English ITU). Every three to four years, the ITU organizes a world radiocommunication conference (eng. WRC), at which it reviews the fundamental document – the Radio Regulations (ITU RR). The Radio Regulations govern the use of the radio frequency spectrum and satellite orbits, including frequency allocation for certain services. The regulation is mandatory for execution by all countries that have signed the ITU convention, and these are almost all countries of the world. At the state level, decrees are issued that duplicate the regulations. For example, in Russia this is a government decree “About Table Approval allocation of radio frequency bands … “.

There is also a distribution in this regulation for satellite radio navigation services (RNSS), which includes both GNSS and Xona Pulsar. In the space-to-ground direction, several wide sections are distributed in the lower L band and one narrow section, 20 MHz each, in the S and C bands:

Frequency allocation for radio navigation satellite services
Frequency allocation for radio navigation satellite services

The lower L band is overpopulated with existing systems, these are GPS, GLONASS, Galileo, Beidou, IRNSS, QZSS, SBAS:

Adding new signals to the L band is possible, but new signals should not significantly affect the operation of old systems. To do this, on the consumer side, the spectra of new signals should not cross the old ones, or their power spectral density (PSD) should be comparable with the PSD of old signals. To ensure sufficient accuracy and noise immunity, Pulsar must use wideband signals: units, and even better tens of megahertz. This means that it will not be possible not to intersect with old signals in the L range. But it is also difficult to ensure a low PSD level, because we are promised a signal 100 times more powerful than GNSS signals!

You can try to spread the signal over the entire frequency domain, say from 1164 to 1300 MHz. But even in this case, its PSD level will be 10-13 dB higher than the PSD level of such signals as GPS L5, Galileo E5, Beidou B3I and others. And this will cause a significant level of intersystem interference.

In addition, Xona announced a two-component signal. It is not clear from the release whether two carrier frequencies or two components on one carrier are meant. If the first option, then where will we smear? If the second, then the level of intersystem interference will increase by another 3 dB.

Placing signals in the S and C bands has a number of negative consequences. Firstly, this will require a separate antenna and separate paths, which means that the implementation of the system in mobile phones, etc. will become more complicated. devices. Secondly, in proportion to the square of the frequency, the power of the received signal will drop, all other things being equal. Thirdly, the influence of the troposphere and precipitation, which are relatively small in the L-band, will increase.

Receiving GNSS signals against the background of your own transmitter

To synchronize the satellites and determine the trajectory of their movement, a GNSS receiver is installed on the satellites. I can assume that these will be GPS or GPS + Galileo receivers. Although the Pulsar satellite moves in space, it is not much closer to the GPS satellite than any terrestrial consumer. The signal level at the output of its receiving antenna is still tiny, on the order of hundreds of zeptowatts. But from the same satellite, probably in the same frequency range, it is necessary to emit a signal with a power of tens of watts!

In my experience in developing similar systems, spreading the antenna to different sides of the satellite can provide an isolation of 20, maximum 40 dB. This will not help if the power difference between the transmitter and receiver signals is 17 orders of magnitude; without additional measures, the reception of navigation signals will be tightly suppressed.

What additional measures can be taken? It would be possible to spread the signals into different frequency ranges. But see item 1, this solution has big drawbacks.

Pulse signal
Pulse signal

You can make a pulse signal. For example, alternate turning on and off the transmitter with a tempo from several hundred ns to units of ms. To maintain an average signal level, this will force us to increase the instantaneous power by another 3 dB, exacerbating the already acute problem of inter-system interference a little.

Variant of a signal on several subcarriers with disconnection of separate ones for receiving GNSS signals
Variant of a signal on several subcarriers with disconnection of separate ones for receiving GNSS signals

You can make a signal with multiple subcarriers and turn them off one by one. So peculiar PRFC. But the effect and problems will be similar. And a new one will appear – the preservation of a constant envelope, without which the efficiency of amplifiers on the satellite will drop from fifty percent to very modest values. And these are problems of power supply, heat dissipation, and dimensions.

Signal polarization

The angular dimensions of the Earth as viewed from a GPS satellite are 30 degrees. The GPS satellite radiates a cone in this direction, the power is focused into a narrow beam. The satellite from Xona is located above the surface itself, it will need a much wider radiation pattern. This means that the signal level will drop.

GPS satellite transmit antenna pattern
GPS satellite transmit antenna pattern

But the drop in signal level is not the main problem. When using a flat antenna array and significant deviations from nadir, the signal polarization will begin to degenerate from circular to linear. And polarization is one of the most effective ways to deal with multipath. When a signal with circular polarization is reflected, it often changes the direction of polarization. This unwanted signal with a different direction of polarization is perfectly suppressed by the receiving antenna. The loss of this quality will negatively affect the accuracy of positioning.

Key distribution

To protect against spoofing attacks and relayed interference, the Pulsar signal is promised to be encrypted. I can assume that the protection will be built on the principle of authorized access signals of the GPS system. In this approach, the high rate signal is modulated with a pseudo-random sequence that is a hash of the current time. To fully receive a signal, you need a key to reflect the time in that very sequence.

There is no need to transfer the full key to each consumer. It can be broken down into many black-red pairs. From each pair, black is sewn into the receiver during production, and red is already transmitted via communication channels. So each receiver will have an individual key. Interception of a private key without a receiver to which it is intended is useless.

But this scheme can work when the receivers are military and are strictly registered. What about civilian consumers? Account for every iPhone sold and every Tesla?

This is probably one of the reasons for the declared two-component signal. One can only be given to the military, the second to civilians. But then most receivers, civilians, will continue to operate on a spoofed signal. Although this will complicate the spoofer itself.


Behind the simplicity and elegance of the idea of ​​a low-orbit navigation system lies many technical problems in its implementation. It will be interesting to see how the developers of the system get around them.

Similar Posts

Leave a Reply