Dallas Air Traffic Rerouted As FAA Probes Faulty GPS Signals (bloomberg.com) 54
An anonymous reader quotes a report from Bloomberg: Flights into the Dallas area are being forced to take older, cumbersome routes and a runway at Dallas-Fort Worth International Airport was temporarily closed after aviation authorities said GPS signals there aren't reliable. The Federal Aviation Administration said in an emailed statement Tuesday it's investigating the possible jamming of the global-positioning system that aircraft increasingly use to guide them on more efficient routes and to runways. So far, the agency has found "no evidence of intentional interference," it said. American Airlines, the primary carrier at DFW, said the GPS issue is not affecting its operations. Southwest Airlines, which flies from nearby Love Field, said it also isn't experiencing any disruptions. The FAA reopened the closed runway earlier on Tuesday.
The GPS problem -- despite the lack of impact -- highlights the risk of widespread reliance on the weak GPS radio signals from space used for everything from timing stock trades to guiding jetliners. The FAA occasionally warns pilots in advance of military testing that may degrade the GPS signals and pilots sometimes report short-lived problems, but the interference in Dallas is atypical, said Dan Streufert, founder of the flight-tracking website ADSBexchange.com. "In the US, it's very unusual to see this without a prior notice," Streufert said in an interview. ADSBExchange.com monitors aircraft data streams that indicate the accuracy of the GPS signals they are receiving and the website began seeing problems around Dallas on Monday, he said. The military has told the FAA it isn't conducting any operations that would interfere with GPS in that area, said a person familiar with the situation who wasn't authorized to speak publicly about it. The primary way FAA's air-traffic system tracks planes is based on GPS, but older radars and radio-direction beacons have remained in place as backups.
The GPS problem -- despite the lack of impact -- highlights the risk of widespread reliance on the weak GPS radio signals from space used for everything from timing stock trades to guiding jetliners. The FAA occasionally warns pilots in advance of military testing that may degrade the GPS signals and pilots sometimes report short-lived problems, but the interference in Dallas is atypical, said Dan Streufert, founder of the flight-tracking website ADSBexchange.com. "In the US, it's very unusual to see this without a prior notice," Streufert said in an interview. ADSBExchange.com monitors aircraft data streams that indicate the accuracy of the GPS signals they are receiving and the website began seeing problems around Dallas on Monday, he said. The military has told the FAA it isn't conducting any operations that would interfere with GPS in that area, said a person familiar with the situation who wasn't authorized to speak publicly about it. The primary way FAA's air-traffic system tracks planes is based on GPS, but older radars and radio-direction beacons have remained in place as backups.
GPS should be encrypted (Score:2)
Re: (Score:3)
Military GPS is encrypted.
Re: (Score:3)
There is a relatively easy fix for this. Place Differential GPS [wikipedia.org] stations around all airports that use GPS landing aids.
A Differential GPS station receives GPS signals at its location, compares the reading to its stored, very precise LAT/LON location, and immediately transmits any correction for that latest set of GPS signals. It reduces the standard distance uncertainty from about 15 meters down to about 1-3 centimeters. So even if bad actors were transmitting erroneous GPS signals, the DGPS correction w
Re: (Score:2)
There is a relatively easy fix for this.
You are 100% correct... but that costs money and would require upgrades for every airport and every plane. Rollout time would likely be a decade if they started now. However, without identifying the cause of the GPS issue here then you may simply be trading one problem for another problem.
Re:GPS should be encrypted (Score:4, Insightful)
You dont want this for aircraft because they require absolute position. The master GPS fails in RTK system or becomes inaccurate, the whole system suffers. This is useful for driving a tractor around a field because tractors are stable and the error can be removed, but planes are faster.
Position is best determined from a multitude of different sensors that use different principles and pulls them all together like the Extended Kalman Filter used in Ardupilot. If one sensor fails it's detected by its deviation from the other sensors.
Planes could use a combo of gps, dead reckoning (with accelerometers/gyros/mags), optical flow, radio beacon, sonar, lidar/led, radar, barometer, even AI looking at cameras like Teslas use... anything and everything that is consistent and accurate. Then you have fallback upon fallbacks.
Re: (Score:2)
Isn't RTK GPS accurate to within 1-3cm between the gps units (relative), but the absolute position is still dependent on the error of the primary gps in the system?
To answer my own question, yes I am correct.
From wiki:
https://en.wikipedia.org/wiki/... [wikipedia.org]
"This allows the units to calculate their relative position to within millimeters, although their absolute position is accurate only to the same accuracy as the computed position of the base station. The typical nominal accuracy for these systems is 1 centimetr
Manipulated input data is the problem. (Score:2)
As long as the US military has the ability to manipulate the GPS location input data,
it doesn't matter whether you encrypt the data with quantum computing proof encryption methods.
Re: (Score:3)
My info here is a bit dated, so no idea if this is still the case, but GPS satellites used to broadcast two messages (C/A and P/Y) on two frequencies (L1 and L2 - L1 had both, L2 just contained P/Y).
C/A is the coarse message, and is unencrypted - this was the message which was tampered with to produce Selective Availability.
P/Y is the precise message, and is encrypted - this was the message which military receives could use to avoid the Selective Availability error.
All civilian usage is via the C/A message
Re:GPS should be encrypted (Score:5, Informative)
That is still all true, although the US government has published enough about the current Y-code structure that receivers can correlate the L1 and L2 signals to get improved positioning performance. The USG plans to break that ability in the future [gps.gov] -- originally scheduled for 2020, now pushed out almost a decade.
Newer GPS satellites -- since Block IIR-M -- now broadcast CNAV (civil navigation) messages on the L2C (L2 civil) signal, which is mixed in with the P/Y and M-code signals on L2.
And an even newer subset of satellites -- since Block IIF -- also broadcast CNAV on the L5 signal, which does not have a military signal.
Even newer GPS satellites -- GPS III and newer -- also broadcast CNAV on another signal using the L1 frequency, L1C. L1C improves accuracy compared to L1 C/A and uses a modulation approach that is shared with other constellations (Galileo, Beidou and QZSS).
Re: (Score:2)
Is this level of control still an issue in a world where multiple alternate services exist? The idea of the military's control over GPS was strategic in case of conflict, but when anyone can simply use GLONASS or Galileo does it matter anymore? My phone already receives all three, my smartwatch 2 of them.
Re: (Score:2)
What level of control? L2C, L5, L1C are all for civil use and each has its own target use case, slightly different from each other.
P/Y- and M-code are the GPS military signals. Beidou and GLONASS also have military-only signals. Galileo has its PRS, which is limited to authorized government users (but supposedly for civil authorities).
The military/restricted signals are for anti-spoofing purposes: in case of war, the military trusts those signals. Civil signals may be degraded, disabled or corrupted in
Re: (Score:2)
I wonder if they will go through with their plans to degrade GPS by breaking the ability to correlate the two signals.
Aside from anything else, Galileo and BeiDou both offer higher accuracy than GPS anyway. Even if GPS gets degraded or blocked in certain parts of the world, unless those two cooperate it might be limited impact.
Re: (Score:2)
Using any of the GPS civil signals on a second frequency is technically much better than the P/Y correlation approach (called codeless or semi-codeless operation, depending on implementation details) because the target correlation pattern is known perfectly rather than being measured by a receiver. That's why the codeless processing sunset is contingent on having enough satellites broadcasting one of those. L5 was chosen because it's in a frequency range where aeronautical use gets priority (the L2 signal
Re: (Score:2)
Encryption does not do that, unless you have a secret that is shared between the sender and receiver. (Even then, you need a checksum or similar inside the encryption, to make it likely that tampering will invalidate the message.) When you want the general public to receive the message, that stops being secret.
There is a lot of investigation about how to authenticate GNSS messages. GPS leans towards Chimera [insidegnss.com], Galileo leans towards TESLA [esa.int], etc. These do not use traditional digital signatures because the bi
Re: (Score:2)
What you actually do is cryptographically sign a hash of the data. If the signature doesn't check or the data doesn't match the hash, it's been corrupted or deliberately altered. Everyone knows the public key. The secret key is kept secure.
Re: (Score:2)
Please read up about how crypto signatures do not depend on a shared secret (in fact, the more public the public key is, the better). Then check your own face for egg.
Re: (Score:2)
The legacy channel has a low bitrate. The new ones run at a bit over a megabit per second. I'm pretty sure they can shoehorn a signature in there somewhere, at least every N samples. The legacy channels can't be changed anyway since they are for legacy devices.
GLONASS and Galileo alternatives (Score:2)
Re: (Score:1)
So the back-up hen-house guard would be a wolf?
Re: (Score:2)
So the back-up hen-house guard would be a wolf?
Given the way things are going, more like a sheep in wolf's clothing, but mangy and tattered.
Re: (Score:3)
Galileo is not yet operational and US-certified avionics would almost never be certified to use GLONASS. (Galileo has quite good performance even without being operational, but that is enough to make it unusable for safety-of-life applications. Also, the standards to allow avionics to use Galileo are effectively not developed enough for hardware to be certified against them. EUROCAE ED-259 is formally published, but is regarded as not "production ready".)
Also, when approach operations need SBAS/WAAS augm
Re: (Score:2)
My older cell phone can see some 20 GNSS satellites from at least four constellations (GPS, Galileo, Glonass, and Beidou) most of time and refine a pretty accurate fix based on all those observations. The FAA would be foolish to ultimately deny certification to multi-band, multi-constellation receivers, provided the receivers could properly recognize a bad satellite and discount it. No doubt that is the hold up.
A fix from a modern GNSS receiver with multiple constellations can achieve an uncorrected accur
Re: (Score:3, Informative)
I am pretty damn libertarian, and am generally very quick to believe the government is delaying things, but the delay here is not due to the FAA. The FAA will definitely certify dual-frequency, multi-constellation (DFMC) receivers as soon as they are ready and sufficiently safe.
The standards for DFMC equipment are not yet published. The Annex 10 SARPs [icao.int] are in the final stages of being published. RTCA is still developing the DFMC version of DO-229. EUROCAE ED-259 is not expected to be "usable" until Rev A
Related to a 5G rollout? (Score:3, Interesting)
It wasn't that long ago that the US Military was furious at the 5G rollout because they claimed it could interfere with GPS:
https://arstechnica.com/tech-p... [arstechnica.com]
Someone turned on ... (Score:2)
INS should be ubiquitous (Score:2)
Miniature INS units should be incorporated in all GPS nav systems as backup.
Re: (Score:3)
That's a whole lot easier to say than to actually do.
Either way, aviation has several ways to navigate without GPS, and they've been very reluctant to fully remove the older navigational aids because, well, GPS is great, but every once in a while it fails, and they've already got backups.
That said, inertial guidance systems tend to be expensive, larger than GPS units and can only give sufficiently accurate information for a limited time, so they're a good backup to GPS in some applications, but not all, at
Re: (Score:2)
Some cars apparently do have them. They're just "accurate" enough to fill gaps like driving under a bridge, with a lot of software guessing.
Re: (Score:1)
Re: (Score:2)
Based on performance I suspect that most cars are using GPS only for this functionality, no compass at all. This is smart because you have to recalibrate mag sensors. It's better to just have the compass be wrong until you have at least two 2d fix points than to risk having it be wrong semi-randomly.
Re: (Score:2)
Most do yes. Wikipedia says there are some systems that use vibrating gyros and the odometer. I've never seen one, but it doesn't sound like a feature that would be in a car I would buy.
Sure it's GPS? (Score:2)
Re: (Score:3)
Airplanes don't use RTK. When using GNSS signals, they use RAIM [wikipedia.org] (in FD or FDE mode) or SBAS. At some point they will also be able to use ARAIM.
Re: (Score:2)
Disaster Tolerance (Score:2)
Am I barking up the wrong tree here? This seems eminently doable, but I'm just spitballing.
Re: (Score:2)
They already have and use ground-based guidance. It's all they had, prior to the 80s, for anyone trying to land IFR.
It's been too many years since I was in flight school, so I don't remember what exactly the beacons are called, but a quick Wikipedia glance mentions "ground-based systems such as DME/VORs and NDBs".
Re: (Score:2)
The US war bureau is called the Department of Defense (DoD), not MoD. They will occasionally monkey with the GPS broadcasts, but normally announce it in advance. Lots of other people should have observed if the GPS birds were acting oddly -- and they did not, even across town at Love Field.
Also, Paul Wellstone [wikipedia.org] died in a plane that was being manually piloted, not on autopilot. Both the pilot and co-pilot were known to have poor flying skills and lapses in judgment (like not wearing corrective lenses that
Differential GPS (Score:3)
Isn't this the sort of thing that Differential GPS was designed to resolve? A network of fixed stations with known locations, broadcasting their own signal to augment the satellite signals to improve resilience and accuracy?
Re: (Score:2)
Well, there is such a thing [wikipedia.org]. You would need an international standard for how the DGPS messages get to the aircraft, and training on when to shut DGPS off. Also, it's only going to get you from 20 meter accuracy to 10 meter accuracy [dfo-mpo.gc.ca] (that is really a refreshingly direct document, I should read more maritime stuff.)
Re: (Score:2)
Kind of. DGPS is not certified for use by airplanes.
The equivalent system for aviation is a Ground-Based Augmentation System [faa.gov] (GBAS; formerly Local Area Augmentation System, LAAS). Each GBAS installation is expensive, it only works in a small radius around a particular airport, and there is roughly one vendor who has a certified GBAS product, so it is not very widely used. Something like three US airports (EWR, IAH and SFO) have it.
Aviation -- and other users -- can also use a Satellite Based Augmentation [faa.gov]
Re: (Score:2)
For all except a Cat III landing the last bit of the landing must be done by a pilot flying visually. Only in a Cat III approach and landing can you be in zero visibility and have the autopilot
Re: (Score:2)
A man with a watch knows what time it is. A man with two watches is never sure.
In other words, how do you know if the GPS data or the DGPS data is the one to trust?
Newer GNSS chipsets use multiple satellite networks (GPS, Galileo, BeiDou, GLONASS) and can detect if one of them is giving readings that don't tally with the others.
But! (Score:1)
It's not like the runways are moving around (Score:2)
I'm surprised that the Flight Management Systems (FMS) aren't cross checking with other data. GPS is one portion of modern FMS but so are Inertial Reference (IRS) and other older technologies. No modern airliner just relies on one system to indicate where the aircraft is and where its destination is, those are input into the FMS usually before flight and can be changed if rerouting is required.
I can understand the concern if GPS is lost but FMS systems can handle that and as for runways, there's only 7 runw
Lufthansa was right (Score:2)
I'm not saying it's aliens, (Score:2)