GPSJam: Daily maps of possible GPS interference | Hacker News Reader

GPSJam: Daily maps of possible GPS interference

gpsjam.org

220 pointsscanny89 comments

Hello, HN. I made this site. Link to previous discussion: https://news.ycombinator.com/item?id=32245346

I'm curious about HN's take on a couple things:

1. Why haven't FlightRadar24, FlightAware, or any of the other flight trackers done this? Not enough people actually interested? I know there are companies that use GPSJAM to help brief their pilots.

2. Monetization! I can't really justify spending more time to work on projects like this one when they don't make money, and do cost (a little) money. But I absolutely do not want to turn these projects into another another job--I don't want contracts, obligations, deadlines. Do I try to get (very niche, but I bet they exist) advertisers to cover site costs and some of my time? Do I crowdfund for general development? Let me know your thoughts. And this isn't just for GPSJAM; I have other projects. For example...

A couple other aviation related projects I've done recently:

1. The Global Aircraft Event Viewer: https://aircraft.social/events/ A near real-time map of higher level aircraft behaviors around planet Earth: Circling, "near misses" (RAs), takeoffs, landings, emergency squawks, and other stuff. A very early experiment, but I think it's kind of neat (especially the RAs!).

2. Closest Points of Approach: https://skycircl.es/cpa/ An analysis tool for checking to see just how close two (or more) aircraft got. For example, if you saw the New York Times Story "How a Series of Air Traffic Control Lapses Nearly Killed 131 People"[1], here's a link that visualizes and animates the scenario that happened in Austin, where a Fedex jet almost landed on a Southwest passenger jet: http://skycircl.es/cpa/?kmlurl=https://gist.githubuserconten... My tool estimates they got within about 150 feet (assuming idealized point aircraft without volume!)–The NTSB said they got "within 200 feet".

1. https://www.nytimes.com/2023/10/11/business/air-traffic-cont...

Would you be interested in handing it off to a maintainer owning it in a 501c3 as a public good? If not, do you want to continue to run it as a non profit public good with donations? I am willing to donate, just curious what the costs look like, and am a huge fan of projects like this (sensor networks/fabrics).

How do you know that a RA has happened? You infer it from the flight paths?

Also, I'd just add ads and some recurrent donations and see how it goes - it looks to me that you're the perfect person to run projects like these. It would make no sense if money was the only problem that stops you.

Just a note from the sites FAQ: the data is actually showing places where airplanes report LOW NAVIGATION ACCURACY via their ADS-B transponders.

Since planes tend to default to INS navigation with use GPS, GLONASS, Galileo or other systems to supplement the drift, it may be subject to certain data skew (e.g. if there's an area on the world where most planes aren't equipped with GPS navigation system, have poor quality navigation tech or you might be seeing "jamming" where there might just be a common place on airline route where INS drifts).

It should be easy to correct for - only count aircraft that actually report good navigation accuracy sometimes.

It only counts aircraft as experiencing interference if they have previously reported high navigation accuracy on their flight, and then begin reporting low accuracy. It's not going to just show places where INS drifts. However, it may show areas, typically military, where aerobatics or hard maneuvering are taking place, which can mask an aircraft's GPS antennae from the satellites.

To pile on about what this actually shows, further the data is sourced from ADSB Exchange so there's only data in areas where there are sufficient participants, and skews coverage towards Europe and North America

https://map.adsbexchange.com/mlat-map/

Project owner might be able to add data from stationary GNSS reference stations to improve global coverage of positioning accuracy available.

https://geodesy.noaa.gov/CORS/

https://epncb.oma.be/

You're saying aircraft are reporting position based on their INS? That seems… weird. I've also been told the opposite, that ADS-B navigation accuracy is always reported based on GNSS, not INS. What's your source?

I hope this doesn't confuse the issue, because I don't think it's actually related to what you're talking about, but recently there have been reports of inertial nav systems being "corrupted", essentially, by spoofed GNSS. The hypothesis is that the INS/IRS is corrected using spoofed GNSS data after which the INS/IRS reports bad info: https://ops.group/blog/faa-warning-navigation-failures/

Err, yeah - the airliners still mostly use INS for their navigation (with corrections, either manual or automatic, depending on equipment and age), but ADS-B indeed needs to report its location from a GNSS "High-integrity position source". So scratch the INS part.

Pure INS should be pretty rare these days, no?

I believe it's now usually a combination of multiple sources (e.g. ground-based navaids like VORs and DMEs, GNSSs/GPS, and INS to reject most spoofed signals or in case everything else fails).

Not sure what do you mean with "pure" INS. Primary INS is still standard - as in, all the navigation is driven primarily from INS system outputs.

But, as it always was, it is regularly corrected with GNSS or navaid corrections. As far as I know most airliners still don't automatically enter corrections for INS and require manual input (basically retyping of GNSS data from a screen into a keypad) to put human into the loop which can verify if GNSS is being spoofed. The pilots can alternatively do the correction using VOR/DME navaids or even their eyes if they feel like it.

As far as I know, it's becoming much more common to have INS with ability to automatically correct drift from GNSS in smaller planes (and I'm sure that will make it to airliners sooner than later).

You're both right, the aircraft position as reported to aircrew, displayed on maps, and fed to the flight director (the autopilots outermost loop) is a blended INS/GPS solution. The opensource drone projects Ardupilot and PX4 have a lot of info and implementations of this actually.

However the autopilots inner loop and Flight Direction Indicator (artificial horizon) are going to be driven almost entirely by INS. This is because GPS positions can't update fast enough, are noisy, and don't provide velocity, acceleration, or orientation (outside of even slower position integration). Also those systems are not particularly involved in where the aircraft is, rather they care about where it is going.

I haven't looked into what the requirements of ADSB position information is, I don't suspect that they mandate GNSS but they may mandate a degree of accuracy. From an ATC perspective within crowded airspaces ADSB would be augmented by primary surveillance radar to prevent mishaps, spoofing, and errors.

For what it's worth my day job is aircraft avionics, navigation equipment is pretty much daily for me.

Ps. Hi HN, Reddit refugee and my first post here.

To be clear GPS does provides high-quality Doppler derived velocity. It's a fundamental output of all GNSS systems, separate from the positional data and certainly not provided by differencing positional data.

---

I suspect much or all of the data boradcast is GPS derived. Would be nice to have some hard data there.

For almost all general aviation aircraft in the USA the data source for ADS-B Out is GPS. All the GA STCs I've seen are GPS only position sources. Many modern GA aircraft will have AHRS but not a full INS, and so INS is not an option for them. Even on larger business jets that may have INS with GPS integration I am not clear any actually use it to drive ADS-B Out, the few installations I've looked at were all GPS.

Boeing was an early advocate for GNSS alone driving ADS-B Out. I think their argument was simplicity, everybody has the same data, and not dependent on erroneous INS or INS settings. I'm not sure what they ended up doing with all their actual installations. If the intent is to send more exotic data such as roll-angle they'll need some INS or at least AHRS integration, but there is no requirement to broadcast that data. The few early third party STCs I saw for ADS-B Out installations in Boeing aircraft only used GPS data sources.

To my understanding, most aircraft use a combination of INS and GPS for both redundancy and spoofing resilience.

This actually makes spoofing attacks very difficult, since you'd need to target each receiver individually and in a very precise way (e.g. a reasonably filtered system would reject a GPS signal indicating that the plane just by a couple of kilometers and changed direction by 90 degrees without any measurable acceleration).

That would explains why Ukraine went from mostly green around 20 February 2022, to ‘no data’ several days later.

It's very rough. There are GPS jammers in Moscow Kremlin, but they only affect the area maybe 1 km around red square. It has practical consequences of taxi apps freaking out when you pass this area. This map shows the entire Moscow, and a huge area around it, in red.

Jamming is not only near the Kremlin, but also in large parts of Хамовники, near places like Лужники, ЦСКА (because of the mall & stadium), Лубянка, Сити and so on.

Since this data is reported by airplanes I wonder if they experience more jamming because the jamming signals distribute easier upwards than through the jungle of buildings in the city.

I am guessing those places have jammers to prevent the attack by GPS guided munitions?

Problems with GPS around Kremlin existed way before the war and I experienced them myself. Didn't know about the other places because last time I've been to Moscow itself (not the airports) was several years ago.

It shows no jamming over Ukraine.

However, since the map actually shows the percentage of flights reporting low-navigation accuracy, what it really means is that there are no normal air traffic over Ukraine.

It actually shows "no data" over Ukraine.

Ukraine enforces no flight zone because of the war. Therefore there is no data.

Wait, does that mean there are no flights to Ukraine at all? Or is it mostly over the war zone in the east?

No flights at all. See it yourself https://www.flightradar24.com/

To answer the question of what has taken the place of air travel, domestic within as well as international mass transport to and from Ukraine has been predominantly train based during the most recent invasion:

https://apnews.com/article/russia-ukraine-war-riding-the-rai...

https://apnews.com/article/poland-railway-disruptions-arrest...

Some serious interference around areas of military tension: East Mediterranean, Black Sea South-West, Baltic Sea. Scary.

It is protection against drones

If you’re a business in Moscow or St. Petersburg and want accurate time, how are you supposed to sync your GPS clocks?

GLONASS supports time setting too.

Assuming this is to protect against drone attacks, wouldn’t (civilian) GLONASS also be jammed there?

Interesting. India's Punjab state has a massive problem with drugs- and weapons-carrying drones coming over the border from Pakistan and GPS jamming is one of the techniques to fight those drones.

What package is used to display the globe?

Mapbox GL JS https://docs.mapbox.com/mapbox-gl-js/guides/

Ouch, I hope GPSJam has a deal with them...

Mapbox's pricing for this would be insane (for a hobby site) if you got a lot of attention... $1 per 200 page loads once you breach the free tier -- though I'll admit the free tier looks generous for a small hobby site.

No deal. I get 50,000 map loads per month for free, and have only exceeded that once. That time, mapbox did give me a freebie because they liked the project.

MapTiler is _way_ cheaper and a bit more ethical.

Also: you can double dip. When the free tier of mapbox runs out, swap over to MapTiler/StadiaMaps or any other provider of vector tiles.

Is there a good free one I can use on a local server?

https://maplibre.org/ is a free fork of Mapbox v1.x before they brought in their licensing model.

It seems to zoom out as flat map, I was looking for something that zoom outs to a globe.

Check out CesiumJS, its used heavily in aerospace applications. Heres their demo page https://sandcastle.cesium.com/

What's happening in that one spot in Texas over Casamigos Ranch?

It's probably military testing and/or training. See my earlier comment about the Laughlin and Randolph military operating areas–Casamigos Ranch is under the Laughlin MOA.

Is there a way that a non-– aircraft actor could measure this?

That is to say: could I establish a fixed ground station that measured these events and somehow augment this data from airplanes?

If you take your own measurements you can do a lot more than what this site is doing, but with the disadvantage that you have to take your own measurements. For example, you could:

1. directly observe signal integrity with your own GPS receiver

2. compare the location your receiver reports with the actual known location

3. compare the location your receiver reports on GPS with the location a different GNSS network reports.

Yes, a lot of GPS receivers provide RAIM or jamming indication, if they detect signals they're not expecting. Gpsjam just isn't set up to accept such data.

I think Galmon.eu may collect it along with all the other receiver-status data, but it isn't exposed in the main dashboard. You'd have to figure out how to extract it from the data, and I don't think anyone's done that yet, but we're in #galileo on OFTC IRC.

The main problem with this approach is that it's hyper-local. Ground stations aren't moving so they don't cover a very broad area, and they're not evenly distributed or even necessarily in interesting locations. Quite the contrary; regimes that employ GPS jamming are likely also hostile to citizens reporting on same.

I wonder why there is a high p of gps jam in Murcia.

US bases are in Jaen and Sevilla

Murcia–San Javier is a Spanish Air Force base.

I feel like this would be easier to read and interpret if the land were darker colored. The yellow and the land colors blend together when some solo hexagons are present.

I agree that the visualization is not optimal. I wish I was a better visual designer!

You're alright! Hindsight is 20/20 I am a backend dev not a visual designer either. It was straightforward for me to notice on my monitor that I'm used to etc but that doesn't mean on your monitor it looked the same. Just thought I'd share some feedback!

Interesting that the entirety of Cyprus appears to show interference. That must be pretty annoying for anyone living on the island.

interference does not necessarily impact people on the ground – these readings are from thousands of feet in the air and are hence exposed to a lot more signals.

You probably need to be in the line of sight of the interfering device, so it has much less coverage on the ground level.

The invasion and subsequent occupation must've been pretty annoying as well !

Snarky jokes aside, in my experience GPS works okay in Cyprus, so the jamming must be localised somehow.

(major assumption) That's more than likely the position of the USS Gerald R. Ford (CVN-78).

Its support planes are in the area [1], and it was sighted on Sep 28 near Italy [2], and if I understand it correctly, has been tasked to this area.

1 - https://twitter.com/wipljw/status/1712780646530118125

2 - https://www.marinevesseltraffic.com/vessels/USS-Gerald-R.-Fo...

It might not be obvious, but the date on the gpsjam page is an input field: You can change it and see the map for any day. If you do that, you'll quickly see that Cyprus is constantly, always being jammed.

Thank you for giving that insight against my assumption!

I discovered GPSJam a few months ago, and have enjoyed some of your content on Twitter, and really enjoyed your work on the longer form article from a few years ago (the FBI one; "Secrets of the Sky") [0].

Best wishes for your endeavors!

[0] - for anyone curious: https://twitter.com/lemonodor/status/1250581745935593472 or https://docs.google.com/presentation/d/1sowJrQQfgxnLCErb-CvU...

I remember reading power plants in the USA, especially nuclear, have GPS scramblers.

Might be some interesting locations to cross check.

There are no legal GPS jammers in the USA. In fact, the one of two things that will get the FCC on you really bad is to mess with GPS or Cell Phone signals.

For Example, US Prisons want to jam Cell Phones but the FCC says no way.

There are exceptions for the federal government itself and for companies that make devices for the federal government for special (basically defense) purposes.

Obviously, stuff like this is made some how: https://www.rtx.com/raytheon/what-we-do/sea/ngj

Looks great! Is it possible to increase the resolution? Each hexagon cover too much area

The map uses the H3 tiling system, and the cells displayed are resolution 4. Sometimes for specific cases I look at higher resolutions. For example, I believe this map I did during the still-unexplained Dallas-Fort Worth major interference event of 2022 uses resolution 6: https://twitter.com/lemonodor/status/1582403839004508160

The caveats of that tweet still apply to higher resolutions: Specifically that you start to see the effects of individual aircraft, which can be harder to interpret.

What is going on in Spain in the triangle: Mazarron - Cartagena - Murcia ?

There are civilian flights across Australia yet no data. Does anyone know why?

Poor ADSB coverage there?

You need to have one nerdy geek to record the data per ~500 mile circle of the globe.

Thats why ocean coverage is bad too.

Yeah, gpsjam.org can only show data where 1. aircraft are flying and broadcasting ADS-B, and 2. there's someone there to receive the messages (and send them to ADS-B Exchange). You can see a map of where ADS-B Exchange feeders are located here: https://map.adsbexchange.com/mlat-map/

What's up in West London?

What's going on near london?

Does seem strange, there aren't any NOTAMs mentioning it that I can see.

There's Windsor Castle near the green area of the center (Windsor Grey Park). And some related military installations close by. Easy to see why it would be jammed...

No. There wouldnt be jamming of civilian navigation right next to one of the busiest airports in the world. Most likely cause is lots of gps warming up at lhr.

shouldn't every airport show that, including major ones like Atlanta?

Do they use ground mlat there? That is the only reason i can think of that you would get cold gps transmitting. Lht was first operational ground mlat site, but i cant find a current list.

This gpsjam site doesnt have much info on methodolgy. Seeing the data would probably show what is causing this issue.

London is the busiest "airport system" in the world and handles a lot of international traffic, so there might be something unique about it that you don't see somewhere like Atlanta.

That being said, it seems like London only has interference recorded today and every other day is green. I would imagine it's more likely to be a quirk of how the data is recorded rather than geniune jamming.

Not sure how you measure busy, but Atlanta handles way more passengers than London. It may be that London has far more flights due to the use of smaller aircraft however.

https://en.m.wikipedia.org/wiki/List_of_busiest_city_airport...

In terms of airport systems, Atlanta isn't even the top in the USA.

No. As another commenter mentioned, if that were the case you would most likely see it at other airports, and you don't. Additionally, the data that GPSJAM shows is filtered so that an aircraft isn't considered to have "bad" accuracy unless it has first reported "good" accuracy recently.

> Windsor Castle near the green area of the center (Windsor Grey Park). And some related military installations close by.

Both unlikely otherwise (a) you would see similar around other active royal sites (b) same around other military installations.

> What's going on near london?

As others have said, more likely dodgy methodology by gpsjam than anything sinister.

For example:

    - As well as being an active airfield, Heathrow also houses a small number of maintenance bases (e.g. BA have one, United are currently upgrading an ex-Virgin one etc.). I'm not familiar with aircraft maintenance, but I suppose there could easily be scope for RF signals related to test procedures to be leaky and be picked up by geeks.

    - There is also Farnborough.  Whilst Farnborough is predominantly a GA (General Aviation, i.e.  private jet) airfield, it is also on of the many homes of QinetiQ in the UK.  QinetiQ being a defence contractor.  So again, one could speculate about leaky test signals.

I'm not sure. It's very unusual to see even low levels of potential interference in that area. Looking at detailed stats using the history feature of GPSJAM PRO, I'd say it's likely there's a real effect that is having an impact on a significant percentage of aircraft in that area. It's true it can be tricky to figure out exactly what it is having an impact.

Data:

  Date        % w/ poss. interference    # Good    # Bad
  ---------   -----------------------    -------   -----
  2023-10-04  2.82%                      1172      33
  2023-10-05  0.00%                      1204      0
  2023-10-06  2.55%                      1253      32
  2023-10-07  0.42%                      1184      5
  2023-10-08  0.00%                      1261      0
  2023-10-09  4.03%                      1216      49
  2023-10-10  2.21%                      1178      26
  2023-10-11  0.00%                      1176      0
  2023-10-12  0.00%                      1203      0

"# Good" is the number of distinct aircraft that flew through that hex that never reported low accuracy with their GNSS. "# Bad" is the number of distinct aircraft that at some point during their flight reported good accuracy with their GNSS, but some time after that, including while flying through that hex, reported low accuracy. I would say that based on the number of aircraft showing up as Bad, and the fact that it's just that hex seeing this effect, it's likely to be showing a real effect.

My list of the most common things that make an aircraft report low GNSS accuracy:

1. Intentional GPS jamming.

2. Unintentional GPS interference.

3. Aerobatics that cause the aircraft's GPS antennae to be masked from satellites by the body of the aircraft.

4. On-board equipment issues.

I think the data suggests it's not #4. #3 is typically seen in military operating areas (MOAs) in the U.S.; I don't know if there might be fighters practicing in the skies west of London.

Edit to add: FWIW I've gotten one report from a person in the yellow hex that they weren't able to get a GPS fix. That's all I know.

US border is interesting.

People always think we're jamming GPS near the border, but all those yellow and red hexes are in the Laughlin and Randolph military operating areas (MOAs). It's much more likely to either be military testing/training involving systems that cause GPS interference, or jet fighter training where aircraft maneuver so hard their GPS antennas are masked from satellites by the body of the aircraft.

Map showing the possible interference overlaid on an FAA sectional chart showing the MOAs: https://i.imgur.com/vieGhgN.png

TIL

Ok, enough with the map nerd slight of hand. Put the Mercator projection on the top of the list of options, not the bottom. Mercator is the qwerty of map projections, the one users expect and are most likely looking for. It is also often the most useful as it is the projection designed for actual navigation.

I would hope that nobody uses this particular map for actual navigation!

And how many people looking at maps on the internet really perform (globe-scale) navigation where it would even matter?

Getting a sense for distances and/or proportions is arguably the most common use case, with “a sense for the precise bearing” a distant follower.

Reply on news.ycombinator.com