That time Galileo was stuck in the past
On Thursday 11 July 2019 a rather extraordinary - and hopefully incredibly rare - event occurred. The Galileo Global Navigation Satellite System (GNSS) went out of whack. What followed was a week of confusion and concern, with very little formal information coming out into the open from the powers that be. Galileo did come back online again, with a polite note from the European GNSS Agency (GSA) reminding everyone to look the other way because Galileo is not yet fully operational. So what do we know about what happened and how does Focal Point’s technology help if this ever happens again?
- The satellites stopped broadcasting their own orbital data correctly. The Galileo satellites transmit orbital parameters to allow receivers to calculate the satellite positions in space. These parameters are updated every few hours and have a “use by date” tagged to them. It was clear on Thursday 11th, as this use by date expired, that the Galileo satellites were not receiving fresh data sent to them from the ground, or had lost the ability to turn this uplinked data around and broadcast the new information.
- Communication with the satellites had probably been lost. During the service outage the satellites themselves were not reporting the “do not use me” message flag. This suggests that there was no communication made to the satellites from the ground during the outage, regardless of the inability to update the orbital and clock corrections for each satellite.
- The satellites were still being monitored properly from the ground. The online databases of satellite orbital and clock data could still be used in conjunction with the signals being broadcast to calculate sensible positions. So all of the processes of tracking and monitoring the satellite positions and signals were still functioning and Galileo was actually still usable in any assistance mode where the orbital and clock data came from a different source from the signal itself.
- GSA said it was a dicky ticker. The information that was provided to the public referred to issues with the Precise Timing Facility in Italy. This refers to the Earth-based timing system that the atomic clocks onboard the satellites are referenced against. However the fact that the online databases of information could still be used to produce valid position fixes suggest that Galileo System Time itself was still plodding along in the usual reassuringly-predictable fashion and that the satellites were still having their clock corrections referenced against it. So the problem was not the ground-based clocks, nor the system for comparing them to the satellite broadcasts, but must have been the link back to the satellites to update the information stored and rebroadcast from them.
- There should be backups. It shouldn't be possible for things like this to happen if backup systems exist to prevent them, and that is the case here, there is a facility in Germany that can be used instead of the facility in Italy if the Italian one needs a bit of time off (pun intended). So for some reason we do not yet understand the backup system in Germany was not able to take over control from the non-functioning facility in Italy.
- It came back up again! Suggesting at the very least that the GSA had not been locked out of their constellation completely by some ransomware attack looking for Galilean quantities of Bitcoin. But we did learn, worryingly, that it is possible for a system like Galileo to go down for a week when we assume that redundancies in GNSS should mean that multiple independent failures are needed, (or pesky space weather) to cause such an outage. It looks here however that there was probably a single point of failure within human control.
Where do we go from here? It is a sobering incident for the PNT community. Putting the spin aside, there is no way that GSA intended for Galileo to go offline for a week, nor did they have a rapid response way of handing control over to the backup facility. The biggest question of all right now is not whether Galileo has implemented the fix to prevent this ever happening again - the question is actually whether the GPS constellation is at risk of the same failure. A week-long loss of GPS would be catastrophic to the World economy and would certainly occupy a few more column inches than Galileo's mishap did.
When an entire constellation goes down the impact on a receiver’s performance could be dramatic, especially figures associated with availability, integrity and continuity of service. Focal Point’s S-GNSS technology is designed to boost the performance of each satellite being tracked by mitigating multipath interference and isolating the line-of-sight component from the sea of reflected signals when in an urban environment. This allows a receiver to provide maximum performance with a minimum number of satellites in use, a useful protection against a low availability of satellites. Admittedly, we designed this for the urban canyon problem, but catastrophic failure of an entire constellation also falls in the remit where Focal Point adds value to a receiver design.