737 GPS Interference
ฝัง
- เผยแพร่เมื่อ 5 ก.ย. 2024
- This covers the topical subject of GPS Interference, such as jamming and spoofing which is prevalent in areas of conflict or tension such as Ukraine, Syria, Israel and the NATO border with Russia. Although GPS interference is aimed at downgrading military capability it also affects commercial air transport. In this video I describe how the 737 navigates, how GPS works, what effects can be seen during interference and what measures can be taken to mitigate these effects..
Contents:
2:02 FMC Position
10:55 Radio Position
17:56 IRS Position
31:58 GPS Background & Theory
49:29 GPS Equipment
1:00:03 GPS Interference
1:26:52 Hybrid GPS-Inertial
I'm not even a pilot but I love watching these.
Same
Thank you. I admire your dedication to watch such in-depth videos when you don't fly - kudos!
Same
Likewise, I am not a pilot, but I pretend to be a Ryanair pilot flying out of Liverpool and Manchester on my Microsoft flight sim....using PMDG 737-800. I also play the 737-400, but pretend to be a contract pilot working through a crappy agency flying cargo in the middle of the night out of Dublin. This -400 simulation aircraft has low availability and poor reliability, with lots of "in flight" issues.
These videos bring me closer to the machine, as all airliners I think are beautiful machines. My flying career ended with a migraine attack in my mid 20s. I have two pilot friends (A320 and 787/777 rated) who fill in the blanks on procedures etc. The latter is an instructor and examiner on the 78/77.
@@SigmundAnschutz-wi2fj Sorry to hear that your flying career was cut short but so pleased that my videos give you more insight into the aircraft.
His voice and calm approach to the facts are always easy to listen to.
Thank you and thanks for watching
I'm a pilot based in Russia, and we encounter this issue daily. Thank you very much for such a detailed explanation. You are doing amazing work!
My pleasure, glad to help your understanding.
New Captain Brady Video? Sign me up. Grabbing popcorn 🍿
Hope it didn't disappoint!
@@ChrisBrady737 you’re videos are the best out there. Thank you so much for all the work you put in
Very interesting topic that pertains to current events. Military GPS systems are currently being developed to be more resistant to interference. Looking forward to the next video!
Thank you, I'm not sure what my next topic will be yet.
@@ChrisBrady737 There are probably dozens of possibilities. Have learned so much about the 737 the last few months as I've watched your videos.
It is GREAT to have you back with another video Chris. Ironically I was just reading about the FR1269 incident at STN in the news today and just like magic a notification popped up for this!
Sorry this one took so long. There was a lot to research and a lot of flying to be done in my day job!
Yes. The INU’s are well designed to save their internally measured “Biases” after each flight, but sometimes the gyros will be “Running away” excessively, more than the internally recorded biases can compensate for. So I like that you have a criteria for reporting a bad INU. Note! PLEASE DO NOT suddenly shut down or cause an INU to have immediate loss of power. A “Controlled” shutdown is desired, and designed for. A discrete command is sent to tell it “Shutdown”, and allow it approximately 30 seconds max (I don’t know your exact model) to save the last known “Biases” to NVM. The design intent is that after N flights, the accuracy of the INU will INCREASE over time because it uses the last X number of saved gyro “Biases” to compensate for during your ground alignment.
29:30 “coasting” is loss of all valid states. The “Prime Navigation” data will provide inertial true heading, position, velocity, and all linear and angular angles and accelerations in the “High Rate Navigation” feed.
Good info, thanks
That is not our definition of coasting
That PULL UP alert over Iraq is frankly scary.
Great detail - I thought I was fairly familiar with basic GPS Interference. Now I feel I know nothing but have learnt a lot. Thanks Chris
Agreed, but it would be even more scary if you were anywhere near MSA! Greta to hear that you found the video useful.
1:36:00 EXCELLENT knowledge of the KF. Bravo! But understand that the “covariance matrix” from an input state tells the KF “How good I think I am”. I have the honor of meeting Mark!
Thank you, glad I got one part right!! :) I would love to hear about your meeting with Mark, the guy is a genius.
@@ChrisBrady737 Honestly, the peak was the Kalman Filter course taught by Peter Maybeck. Was like getting a batting lesson from Cal Ripken.
I would have loved to have sat in on that course.
@@desertengineer1covariance is not an indication of "goodness", it is a weighted measure of variation between a prediction and the current measurement, where the process noise forms the absolute lower bound.
The aircraft can be on the wrong side of the world, and if the prediction and measurements are consistent, the eigenvalues of the covariance will be small.
@@DumbledoreMcCracken So yes. At the office, I do stand corrected as we should be more specific as we do generalize here. State vectors, Covariance, and “Noise” matrices tend to incorrectly get mixed when we are digging into code. I deserve that correction!
Again, EDUCATION BASED COMMENT ONLY…. 1:13:20 is NOT spoofing. It is something well known called “multi-path”. You are getting error from reflections off of buildings that reduce GPS accuracy. In these situations, you should do what is called a “Ground Alignment” using your manually entered known good position for the parking spot you are on, not the GPS. You should notice as you taxi away from buildings, GPS updates resume and the deviations should disappear.
I can tell that you are very knowledgeable on the subject but in this case I respectfully disagree. This is only seen in areas of GPS Interference such as Larnaca, it would therefore be too much of a coincidence for it not to be GPS Interference.
Multipath can cause errors of tens of metres, maybe a hundred metres in extreme cases, but not multiple miles
@@ramseyfaragher6810 Early instances caused the IP to be 1.5 miles off during ground align. Once they taxiid free of the buildings, the sudden change of GPS position caused the KF to reject updates. We had to develop a “Non-GPS” aided alignment procedure for them, having to manually select “NAV” on the receivers when they made it to open areas. The latest generations of new receivers don’t do this anymore, but it was a pain for a while.
Fly out of PFO and the GPS interference here is insane. Thanks for the presentation
No problem, glad to help
This video is great. I have a HackRF One which is an SDR device. I would NEVER do anything illegal but I find this type of stuff fascinating. Another interesting tidbit, in the desktop sim world, RAIM is very rare to find to be simulated. I have a few cockpit edition avionics suites that has RAIM simulation. Chris, this is awesome work, thank you!!!!!!!!!
What is RAIM?
Oh nevermind I am now on that part!
You are very welcome, thanks for watching
Excellent work, sir! I wish the CBTs and classroom lectures, in the companies I've flown for, could have been so detailef and well delivered.
While flying in China, regularly traversing the east coast and South China Sea, I would observe lots of GNSS dropouts. Usually, one would become unavailable, soon followed by the other. After a minute or two, they would become available. Not a problem enroute, but effen' disruptive if we were doing a procedure needing a tight nav precision.
Thank you for your kind words.
Chris, again an amazing and useful video. I have to cut it into bits because I am headed out to fly on the 737 in a GPS spoofed area of the south Mediterranean. Very relevant. Very detailed. Very informative. IRS are GOOD I have no problems to turn off the GPS updating if I am the slightest in doubt of their reliability. We flew without them installed on the Classics for Gods sake. The old 747s used to cross the Atlantic on IRS only…… amazing technology, would love a video into the tech details inside of the ADIRU BOX itself.
Thank you, I am of the same view as you on the efficacy of the IRS, especially a hybrid GPS-IRS. I may well do a video purely on the IRS.
I have never heard about the Schuler cycle until this video. Flat earthers are definitely going to be annoyed with this lol.
True!!
Interesting navigation interference event. In an aircraft with a very similar FMS, going into one of the areas concerned, experienced an actual position shift (ND Map Shift; UNABLE RNP with ANP >15nm; EGPWS and ATC/ADS-B issues, naturally) approximately 10 minutes after GPS Updating was selected OFF during the Descent/Approach briefing (VOR, DME and LOC Updating all ON). As we were on the STAR, there was (unfortunately) no time to gather more information; requested vectors until ANP had reduced below RNP. LNAV reselected, the approach and landing were uneventful. It’s still unclear to me why/how this could happen with GPS Updating OFF prior the interference event unless there is radio spoofing occurring also, which “sends the investigation in a whole new direction” when we discuss mitigation strategies.
More than thanks I want to point out the hours of study and dedication is really amazing ! 🙌
Thanks Damien, this particular video took months of work
@@ChrisBrady737 I know, I took me +3 hs just to study the video. !! once again thank you and congratulations. hugs!🤗
Fantastic stuff. No one else is doing this. Thanks.
You are welcome, glad you found it useful
You ARE AMAZING!!!
Thank you, glad you liked it.
Another great video. Thanks Chris
Glad you enjoyed it
Very good presentation BTW. Yes, the POS REF, shows the final KF derived solution (FMS POS) and what we call the “Pure Inertial” solution for each IRSU. In your example at 20:54, it is telling you IRS 1 is decidedly “Prime”, the “top dog”, for lack of a better term, and you can pretty much be at ease IRS R is almost totally ignored. If something happens to IRS L, you’ll get IRS 2 as a backup, but now you know the confidence of its’ inertial solution. But with good GPS updates, the error should close down to nearly zero based on periodic updates sent from the KF.
Thank you. The terminology varies between "Pure" and "Raw". As our manuals describe it IRS 1 feeds into FMC L, which may or may not be the primary FMC on the flight.
@@ChrisBrady737 I forgot that. It’s been a while, but somewhere in my memory you select CMD in each FMC?
Well done Chris another amazing video. Marco
Many thanks Marco
Shuler cycle is automatically compensated for internally.
1:18:02 is GPS denied with a bad pressure rate input that affects your Z-axis INU “Reference Altitude”. Keep reporting these through the airline engineers.
Superb!Thankyou so much Chris.Brgds Paul.C.
Thanks Paul, I miss you as a photo source!!
Thank you so much 🎉
You are very welcome
FYI As I understand GPS sats are unlikely to orbit decay and fall back to earth anytime soon as they are in MEO (medium earth orbit) and it would take quite a while to do so >100 years. Old GPS sats may be placed in a "graveyard orbit" when they fail / run low on fuel.
Ok, perfect example is 1:12:14. The GPS is NOT being “spoofed”, the GPS “covariance” (noise/error estimation) is probably so high, the KF algorithm is using only a “little” amount in the state updates. So you suffer INS drift error, as well as ADC, and VORTAC atmospheric and possibly slant range error. The GPS’s are probably good, (otherwise the receivers based on the 2018 mandate would go “invalid”), but your Kalman Filter just doesn’t trust it until the reported covariance goes “down”. Thresholds are unknown because they vary based on STATISTIC estimation.
This occurred deep in an area of known GPS Interference, it surely must have been either jamming or spoofing, no?
@@ChrisBrady737 I find it hard to digest “Spoofing” of GPS time. The method of determining the distance from each satellite is by decoding a Pseudo Random Number (PRN), which is simply a known sequence of binary transitions that does not repeat itself, ever, for 48 hours. Once your receiver determines the PRN “time”, it resolves the GPS time from that satellite (SV). It also will receive the most recent almanac on the 60 Bits per second “Navigation Message”. Given a known time (you know approximately where you are, exactly where the satellite is, and measure the PRN transitions) and accurate almanac, the difference between actual GPS time and the PRN “time” will give you the distance to that “known” satellite (6.18 microseconds per mile, APPROXIMATELY - the atmosphere and ionosphere effects can be “estimated” but not wholly). This is why the military L2, when decoded, can use the frequency difference and get a MUCH better solution. In the end, nanoseconds matter, and beyond a few feet of error, your EPE and EVE will become very high and newer “smart” receivers (mandated back in 2018 to fight denial) will just kick it off the list of satellites chosen for your solution. 3 minimum for a 2-D solution, 4 minimum for 3-D (altitude). So knowing the timing has to be accurate down to nanoseconds (even the ‘group delay’ in the coax cable is accounted for), and receiver will know if it’s not, I have no idea how the GPS time gets “spoofed”. Now, if your system time/date is somehow compromised, the GPS won’t know where the satellites are. However, even this “cold start” condition has a fallback. The receiver will “scan” sequentially for every SV# until it finds them, recovers a time, then when it re-acquires 3, is getting barometric altitude, it rapidly recovers. But if recovering from a totally dorked condition, it will take anywhere from 6 to 12 minutes to re-download the almanac from the nav message. So I don’t know what’s happening with your time. Maybe when all of L1 is denied, the FMS is “defaulting” to something strange. But for a while, your INS is still in “Free Inertial” mode, and your system clock SHOULD be doing the same - estimating the GMT based on the last good known “I-11” time mark pulse (TMP).
The KF knows nothing about error, in the sense called "bias".
It is because the receiver is simply trusting the TOW data in the spoofed radio broadcast without making any checks at all. So you simply get given whatever new time the spoofer has set the TOW data to be in the broadcast. It would appear that the receivers in these avionics have not implemented even the most basic of checks for malicious signals
@@desertengineer1the civilian prn repeats every millisecond by the way. I think you are thinking about the P(Y) military code which is one week long. The civilian C/A PRN is one millisecond long
Excellent as usual. Thanks for including the systems information at the start of the video. How can I know if the hybrid GPS is installed in an aircraft or not? If the IRS align without the need to enter a position, does it then mean it is installed?
Yes, self aligning is the way you identify hybrid GPS
There is now a FOTB 737 24-01 for 737 Max with Hybrid GPS-Inertial enabled..!
Excellent news
My friend is a 787 pilot who flies Europe/Far East and spoofing is a regular fact of life. GPWS warnings at FL41. The clouds of war are gathering.
These technical presentations are fascinating, speaking as a non-professional. GNSS interference must drive you absolutely bonkers, from the outside it appears such a nuisance!
They are an unwelcome facet of flight in certain problem areas of the world.
Hi chris, random question about 737NG vs max. In the max variant you have the equipment smoke light so in case of fire or smoke its very clear where the source is and u just run the NNC for that, in theNG series theres no such thing so what do you do just run the smoke fire n fumes checklist ? Theres no way of knowing if source is avionics bay in the NG how do u deal with this and more importantly how are aware of the fact that u have a smoke source in ur avionics bay in NG series ? Or is there no way to know cheers. Keep up the amazing videos
I don't think that description of the Schuler cycle is quite accurate, the oscillation isn't happening intentionally but is a side-effect of a feedback loop in the IRS, and generally it's happening entirely in the math inside the IRS rather than anything physical. The short explanation is that while keeping track of orientation is mainly the job of the gyros, it is also adjusted based on the computed velocity in order to keep the vertical axis pointed down (aligned with gravity) rather than fixed in space (which is what the gyros would do naturally). If everything was drift-free and perfectly aligned then this would stay aligned without oscillation, however in the (in practice inevitable) presence of any velocity error, the plane will think it is slowly tilting, which causes it to misinterpret a small portion of the 1g vertical acceleration (due to gravity) as being horizontal, in the opposite direction of the velocity error, which causes the velocity error to first be zeroed and then reversed until the plane's imaginary tilt has been cancelled, and then the whole cycle repeats, resulting in the undampened oscillation with 84.4 minute period.
I thought that was what I was saying. My apologies for any confusion.
Chris, you talked about the four different GNSS systems which are even working on slightly different frequencies but do you have any information about which of those GNSS are used in different GNSS receiver generations/versions and whether they are all equally affected by jamming and spoofing? Particularly spoofing might be quite challenging by the spoofer as he would have to provide all the different GNSS signals at once.
I am not sure but I suspect that any constellation can be programmed into any MMR. the newer MMRs can receive multiple constellations. Which are affected depends upon who and how is doing the interference.
There have been reports of IRS being "contaminated" by spoofed GPS inputs. Is that a concern on the hybrid systems on 737s?
That is all covered towards the end of the video.
Excelente! 👏🏼👏🏼👏🏼
Thank you
You mentioned that the antenna is L1/L2 capable. Are either the IMMR or GLU using the L2 signal in their position calculations? The L2 signal is encrypted for military use only. Some creative engineers have figured out how to recover the carrier frequency from the encrypted L2 signal, which can then be used to work out ionospheric dispersion, but for all the extra complexity, and the need for a completely separate RF front end on account of the large frequency difference, the gain in accuracy isn’t any better than can be achieved by use of SBAS, which for GPS operates in the same L1 band.
The manuals suggest that they use both
1:20:34 GPS to the transponder is invalid. Mode S ADS-B DF-17 message, intended to replace legacy ATCBRS, becomes invalid, legacy ATCBRS (Old modes 1, 2, 3/A, C) is enabled, and RAPCON systems see dual replies, one from ATCBRS and then a randomly delayed (Squitter enabled) degraded mode S. Two replies are received at a varying distance, sounding a proximity warning in RAPCON. Annoys controllers. Nobody uses raw video anymore. You’re simply double-replying to synthetic radar.
typo: ATCRBS (not ATCBRS)
@@MatthijsvanDuin ATCBRS = Air Traffic Control Beacon Reference System. The legacy, individual transmitted interrogated pulse-pairs for mode 1, 2, 3/A and C. This was replaced by the Mode S, DF-17 word, which cover both ADS-B and ATC mode S for legacy ATCBRS.
@@desertengineer1 Must be rather obscure terminology given that Google yields literally zero results for "Air Traffic Control Beacon Reference System", and the results for ATCBRS all appear to be typos for ATCRBS (air traffic control radar beacon system) or unrelated nonsense
@@MatthijsvanDuin I am human sometimes. I just focused on designing circuits.
@@MatthijsvanDuin Yup. Typo. It's essentially the legacy modes 1,2 , 3/A, and C. Mode S with DF-17 or ADS-B is supposed to replace it. Terminal synthetic radars are already interrogating both.
Regarding EGPWS false alerts. Some times they will occcour even after you left the RFI area. I.E. you take off from LLBG or LCLK and. Get a false alert descending through 8000 feet in EDDF or EgLL.
This is because you had GPS spoofing. Unlike jamming the spoofing doesn’t go anywhere.
Yes, because some MMRs continue to be affected until their NVM can be cleared
Dear captain. Regarding TOGA position update. It will not work if the calculated PP is more than 2 nm away and GPS PP is on. It is so deigned by Boeing.
We had that issue few days ago, the GPS update was OFF, DME update was ON we had map shift of approx 80nm and when we pressed TOGA it didn’t update the position.
Soviet Front? Chris that term would no longer be appropriate, although I understand why you used it, it has to be Russian Federation now. Excellent presentation as always, thanks.
You are correct, I was showing my age!
The hybrid is great in theory. Where is the chart taken from?
From the paper by Mark Ahlbrecht referenced in the presentation
@@ChrisBrady737thanks
UUWW ❤
1:16:07 is not spoofing.
Contact me offline for a predicted cause of the situation in 1:13:00.
I would love to contact you and pick your brains for more insight on this subject but I cant see any contact details. My details here: www.b737.org.uk/contact.htm
Nothing is easily fixed.
Relatively