Space PNT targets large LEO telecom constellations

Botteron.

SpacePNT SA — established in 2020 in Neuchâtel, Switzerland — provides advanced PNT technologies and solutions for satellites. I discussed the company and its products with its co-founder and CEO, Cyril Botteron.

What is your company’s niche within PNT?

We have developed our own FPGA-based hardware/software/firmware spaceborne GNSS receiver technology specially targeting the fast-growing New Space satellite market and the demanding applications requiring real-time and on-board dm-level positioning and ns timing accuracy, or highest signal reception sensitivity for GEO or Moon missions.

What is your background and that of the other people in the company?

I have been working in the PNT domain since 1999, when I started my Ph.D. in wireless localization at the University of Calgary, in Canada. Then, after finishing my Ph.D. in early 2003, I joined the Institute of Microtechnology in Neuchâtel, Switzerland and transferred a few years later to the Swiss Federal Institute of Technology Lausanne (EPFL). There, for more than 15 years, I led PNT and GNSS receiver R&D activities, several in connection with the Galileo project, which was just starting back then. In parallel, I also worked as part-time Galileo GNSS receiver expert for the European Commission for more than 10 years.

Today, SpacePNT is still growing and counts 11 people including the equivalent of 7-8 full-time engineers with many years of experience in their respective domains who have entirely developed the company’s hardware and software technology. Some of them followed me from EPFL at SpacePNT, while others were previously working for Syderal Swiss, a company that has developed electronics and space equipment for more than 50 missions without any failure, but that, unfortunately, stopped its operations in 2022.

One particularity about our core engineering team is that we have been able to bring together very talented and complementary people, allowing us to perform all the electrical and software design, analysis, development, verification, and qualification engineering tasks of our FPGA-based spaceborne GNSS receiver products internally. I think this is quite remarkable given our still relatively small size and the tremendous complexity in developing satellite GNSS receivers.

What are the origins of your company’s current product offerings?

It all started in EPFL more than 10 years ago, after we had developed some advanced FPGA-based GNSS receiver acquisition algorithms as part of an EU Galileo project aiming to acquire the GPS signals in difficult indoor environments, without assistance and with a very short time-to-first-fix. At that time, we realized that such algorithms could also be used to enable autonomous space navigation toward the moon thanks to the terrestrial GNSS signals. Indeed, when you are at moon altitude, or about 400,000 kilometers from here, it is very difficult to acquire the GNSS signals because they are so attenuated and there is no assistance network up there to help with the GNSS signal acquisition process.

So, we started to build a first proof-of-concept prototype implemented on a powerful FPGA commercial development board to see whether it was possible to acquire the GPS L1C/A signals at moon altitude. After a successful demonstration and because from the moon you cannot see so many satellites from a single GNSS constellation, we added to the prototype the capability to also receive the Galileo E1 signals in order to compute a position fix. Then, in order to improve the least-squares solution we were obtaining — which was very coarse, with an accuracy of several kilometers — we decided to add a second frequency in order to take advantage of the modernized GPS L5 and Galileo E5a signals providing better pseudo-range observables.

After that, the accuracy of the receiver prototype was still limited to a few kilometers at moon altitude due to the poor system geometry. Indeed, from a moon user perspective, all the GNSS satellites are constrained in the same direction towards Earth, leading to a huge dilution of geometry on the order of 1000. This means that even if the pseudo-range observables are estimated with a 1-meter accuracy, the position accuracy will still be on the order of 1000 m because of the poor system geometry. So, we made two additional important improvements. The first one was the addition in the receiver of a model of orbital forces to model all forces acting on the satellite and filter the solution. The second one was to aid the acquisition algorithms from the navigation solution to acquire more rapidly new GNSS signals.

At the end of 2017, we finally achieved a successful hardware-in-the-loop demonstration in our laboratory with this proof-of-concept prototype fed by a real radio frequency signal generated using a GPS+Galileo full constellation Spirent simulator, demonstrating an accuracy of just a few hundreds of meters at moon altitude. It is at that time that part of my EPFL team and I decided to leverage the knowledge we had accumulated toward the development of a commercial spaceborne GNSS receiver product.

Interestingly, the first product we started to develop was not a moon receiver but one targeting LEO satellites and LEO constellations called NaviLEO, because there was more demand for solutions covering LEO orbit satellites than for moon mission, especially in 2019. Today, moon PNT technology is also becoming very important.

How did you become involved with the European Space Agency’s moon mission?

After we started the development of our first NaviLEO spaceborne GNSS receiver product, we won an open competitive call from the European Space Agency (ESA) to develop a moon GNSS receiver prototype that we named NaviMoon. This development built upon the NaviLEO spaceborne receiver development that integrated high-performance, radiation-tolerant COTS EEE components and a radiation-tolerant HW/SW/FW architecture, including latch-up protections and ECC, but this time with a better clock and improved super-high-sensitivity algorithms.

What are the special challenges of making a lunar GNSS receiver?

There are several of them as the super-high sensitivity algorithms and the navigation algorithms are quite complex. One special challenge we had to overcome was related to the hardware. Indeed, for the proof-of-concept, we realized at EPFL, we used a commercial development board integrating a very large FPGA, which allowed us to rapidly develop the algorithms without being limited with the FPGA computational resources. However, when you need to make a space product, then you need to select radiation-tolerant components and also want to minimize power consumption, so the choice of a suitable radiation-tolerant FPGA is very limited. Therefore, a main challenge during the lunar receiver prototype development was to develop super-high sensitivity GNSS algorithms that could fit within the limited computational resources of the NaviLEO hardware. In addition, we also needed to find a better radiation-tolerant low-phase-noise clock allowing very long coherent integrations of the received signals to extract them from the environmental thermal noise.

What happened next?

After the first ESA contract to develop this NaviMoon engineering model, we won a follow-up competitive ESA call to build a flight model that ESA will send around the moon circa 2025 to demonstrate for the first time the use of terrestrial GNSS signals for autonomous navigation in a cislunar orbit. For the manufacturing and testing of the hardware, we partnered in this project with European Engineering & Consultancy (EECL) in the UK. Surrey Satellite Technology (SSTL), also in the UK, is the satellite prime in charge of the ESA/SSTL Lunar Pathfinder satellite that will host our NaviMoon receiver. It will also host a laser retroreflector array that will make it possible to verify the real-time positioning accuracy provided by the receiver in cislunar orbit. We already delivered the flight model to SSTL in June of last year and are very much looking forward to this in-cislunar orbit demonstration. It will be the culmination of a very long development that started 10 years ago at EPFL and has only been possible thanks to the hard work and dedication of all the people who worked on it, including the support from ESA.

What are your key innovations?

Besides the fully in-flight reprogrammable radiation-tolerant hardware we developed and the super-high sensitivity algorithms and orbital forces model integrated in our NaviMoon navigation filter, another key innovation we developed at SpacePNT is our own precise orbit determination (POD) algorithm that can process the clock and ephemeris corrections transmitted in real-time by Galileo satellites (the High Accuracy Service) or by GEO satellites (the Fugro SpaceStar service) and that we are integrating into our NaviLEO-POD product. Thanks to these real-time corrections received from the same GNSS antenna as used to receive the GNSS signals, our NaviLEO POD receiver technology can deliver to the other payloads onboard the satellites, totally autonomously and in real-time, a position and a time with sub-decimeter and ns-level accuracy, which is outstanding if we think of the velocity of a LEO satellite, which travels at a speed more than several tens of thousands of km/h.

One of your press releases refers to NaviLEO as a “spaceborne GNSS receiver product platform.” What does that mean?

When we started the development of our first GNSS product, NaviLEO, we already had in mind the development in the near future of additional receiver products to cover additional markets, e.g., with dual-antenna to provide optimal visibility from LEO to GEO, or with a better on-board clock to enable autonomous moon navigation. This is why we developed the original NaviLEO hardware as a flexible “product platform” or “technological base” that our other spaceborne GNSS receiver products could inherit and build upon.

This is also why I said in that press release that the successful in-orbit demonstration of our NaviLEO receiver product platform is a significant achievement towards future missions. What I meant is that the NaviMoon flight model we already delivered to ESA last year, as well as other NaviLEO flight models we are delivering to other customers this year, are also based on the same fully in-flight reprogrammable technological platform as NaviLEO. Therefore, having this receiver platform already successfully demonstrated in a LEO environment is a great achievement towards the future missions, including the coming ESA/SSTL Lunar Pathfinder demonstration in cislunar orbit. Moreover, this in-orbit validation has also allowed us to de-risk our second-generation product platform, because our second generation reuses the same key radiation-tolerant electronics components, repacked to enable a more cost-effective and larger-scale manufacturing. 

Are you mostly targeting telecoms?

The large LEO telecom constellations are one of our main targets for our second-generation product. Indeed, given the large quantities involved and the market pressure to make the satellites cheaper, it is necessary to develop a technology well-optimized for cost reduction and serial manufacturing. That is something we clearly had in mind when we defined the requirements of our second-generation hardware product that we will qualify in the coming months. We are also targeting additional markets, both with our current first-generation and coming second-generation products, for instance, the Earth observation market or LEOPNT market for which decimeter and nanosecond accuracy can make a huge difference to the quality and performance of the services these satellites can deliver to the end-users, or the GEO and Moon markets for which our super-high-sensitivity receiver technology is perfectly suited.

Use of a NaviLEO-POD receiver onboard each satellite of a LEO-PNT constellation, allowing the autonomous generation of PNT signals within each satellite. (Image: SpacePNT)

What are the key technical challenges using GNSS satellites “from the other side,” so to speak?

What makes it extremely difficult to use GNSS satellites at altitudes above them is the fact that GNSS satellites always have their main lobe antenna gain directed toward Earth and do not transmit any signal power toward outer space. So, when you are above a GNSS constellation, you cannot receive any signal power from the satellites directly beneath you, and in fact the only signals you can receive are coming from the spillover around Earth of the satellites that are on the other side of Earth, or that come from the secondary side lobes of the GNSS transmit antennas.

Since the antenna gain of the secondary side lobe is reduced by about 14dB as compared to the main lobe directed toward Earth, this is yet another reason why super-high-sensitivity algorithms are needed for moon and GEO missions, to allow the use of these lower power signals transmitted by the GNSS secondary antenna side lobes.

More specifically, what are the challenges of moon navigation?

Moon navigation, whether in transit to the moon, in a lunar orbit, or on the lunar surface represents several challenges. These include the definition of a reference time and geometric reference frame to be used on the moon, and the definition of standards for communications and positioning to be followed by the different moon users and moon service providers in order to achieve interoperability, amongst others.

Do ESA and NASA plan to place navigation satellites in orbit around the moon? If that’s the case, are you bridging the gap until the new system is deployed?

Yes, exactly. NASA and ESA are collaborating to define the LunaNet Interoperability Specification. It is a common framework of mutually agreed standards to be applied by users and service providers in a cooperative network and support missions on and around the moon. In this framework, PNT services are envisioned to be provided in two ways to lunar users, through dedicated communication links and through a GNSS-like lunar navigation system.

As there is no such infrastructure available yet, however, our NaviMoon GNSS receiver solution that ESA will demonstrate as part of the ESA/SSTL Lunar Pathfinder mission circa 2025 is a first step toward the effort to develop lunar PNT capabilities. It will also illustrate how GNSS can play a meaningful role in lunar PNT, analogously to the way that LEO-PNT complements GNSS for Earth users.

How can your system contribute to complementary PNT?

In a nutshell, a complementary PNT constellation can provide PNT services similar to a GNSS system, making each LEO-PNT satellite transmit PNT signals that contain both its real-time ephemeris and time of transmission. To do that, what every LEO satellite of a LEO-PNT constellation needs is a means to compute its own precise ephemeris and to precisely synchronize its time-frequency with the others. This is exactly what our NaviLEO POD solution can do. Thanks to its on-board real-time POD algorithm and the real-time GNSS clock and ephemeris corrections it can receive from GEO satellites, such as Fugro SpaceStar service, or from MEO satellites, such as the Galileo high accuracy service, it can disseminate within the satellite the real-time precise orbit determination of the satellite needed to compute and transmit its own precise ephemeris. It can also compute and transmit the ns-level timing frequency synchronization with GPS or Galileo system time. The LEO-PNT satellite can then use these data to generate the PNT timing signals sent toward the PNT terrestrial users. The beauty of this concept is that there is no need for inter-satellite links, additional ground station infrastructure or atomic clocks on the LEO-PNT satellites. The GNSS receiver equipment can do it all by itself, delivering a time fully synchronized with a GNSS time scale that is maintained by the atomic clocks onboard the GNSS monitor stations and GNSS satellites. Our system is also resilient to short GNSS outages, thanks to NaviLEO POD’s advanced algorithms that optimally combine the multi-constellation multi-frequency GNSS measurements with a precise model of orbital forces allowing the propagation of the navigation even in the absence of GNSS measurements. Finally, what I think is remarkable is that thanks to the MEO or GEO real-time correction signals used to correct the ephemeris and clock errors present in the real-time signals transmitted by the GNSS satellites, a LEO-PNT satellite equipped with our solution can potentially transmit ephemeris and clock signals towards the terrestrial users that contain fewer errors than the real-time ones transmitted by today’s GNSS satellites.

Plus, of course, the received signal on Earth from LEO satellites is much, much stronger than that from GNSS satellites, which has many advantages, right?

Absolutely. The additional power means that signals transmitted from LEO satellites are much more difficult to jam or spoof, thanks to the higher received signal power. In addition, because the LEO satellites travel much faster than the GNSS satellites above a terrestrial user, the signals are much more dynamic. So, even if one wanted to make a very complex spoofing attack with UAVs, everything is so dynamic and moving so fast, that it would be very difficult to implement. This may make the receiver more complex but also brings advantages. For instance, if one user application does not need a high accuracy fix, it is possible to use the Doppler effect to locate a receiver with just one or two LEO satellites.

Are you working on any other related projects?

We are also working on an enhanced orbit propagation tool called SimORBIT and commercialized by Spirent. It enables realistic testing of emerging LEO satellite constellations with the generation of output files in SP3-c format, as well as in the proprietary Spirent MOT and MOTI formats. We are also constantly improving our receiver technology and widening our product offerings.