Starlink signals can be reverse-engineered to work like GPS | Hacker News Reader

Starlink signals can be reverse-engineered to work like GPS

technologyreview.com

354 pointsSanjayMehta173 comments

I think there is a bit of an issue here that isn't addressed in the article. They reference using the SpaceX published ephemerides which are created from the GPS receivers onboard the SpaceX satellites:

"SpaceX satellites regularly downlink accurate orbital information from onboard GPS. We use this orbital information, combined with planned maneuvers, to accurately predict future ephemerides, which are uploaded to Space-Track.org three times per day" [0]

So using the positioning information of the SpaceX satellites is already dependent on GPS. Saying that it could be used as a backup to GPS is a bit non-sensical to me. Sure you could go back to using ranging measurements for each of the satellites to get TLEs from NORAD or LeoLabs or one of the other commercial space tracking companies, but it would likely be less accurate and not updated as frequently as the SpaceX satellites don't have a precision clock onboard for timing and propagation like the GPS satellites do.

[0] https://www.spacex.com/updates/index.html

They're different failure modes. It's unlikely the GPS constellation will go dark anytime soon. It's much more likely you're in an area without good view of the (few) GPS satellites required for a fix; or that there is GPS jamming, etc.

Ah, that is true I didn't consider it from that perspective. It may also serve as a decent anti-spoofing measure if you compare the two results for position that you are getting directly from GPS vs from the Starlink satellites.

Presumably GPS can be jammed locally on the ground and confuse ground terminals but not at the range required to block the SpaceX satellites reception in low earth orbit.

Are the phased array antennas utilized by the starlink satellites and user terminals more resilient to jamming?

no, but they're a different frequency so if someone was only jamming GPS, this would get through.

They are massively more resistant to jamming. The antennas are very good at rejecting signals that are coming from the wrong direction, that is, not from above.

This is the correct answer, phased array antennas have much more directivity so they are much harder to jam with a ground-based jammer. Although I suppose in some cases you might want to put the jammer on a drone anyway, to cover a larger area. In that scenario you would probably not see a significant difference for phased array antennas vs "normal" antennas.

Caveat btw: for both starlink and GPS, the satellite you are talking to will not always be "up". For GPS in particular, it is possible that some of the satellites are only barely above the horizon. So an antenna that only looks "up" is generally not what you want anyway, which makes ground-based jammer more effective again.

Phased arrays are good at rejecting incidental interference, which is why they were used to address radar jamming. Of course that also meant that billions were poured into developing technology to counter that resilient property. Anyone capable of blocking GPS on a non-trivial scale would be easily capable of blocking Starlink as well.

It depends strongly on how the phased array is being steered. If the components prior to phasing and summation are saturated by the interference then phased array beamforming will not reject the interference.

GPS jamming occurs local to the receiver, not the transmitter. So, you'd have to worry about the dish on your roof / vehicle, not the ones on the satellite.

I'm that sense, a phased array can help by filtering out signals that are not in line of sight with your satellite, but only one satellite fix does not provide a great estimate of the receivers position. Esp when signal strength is used to estimate bearing.

well, anyone having to deal with you (and I) do. I wouldn't put it past a APT to be able to do better and take action in space if their hand were forced.

Interesting! Is it theoretically possible to launch a "GPS jamming" satellite constellation? And if so, could you do it with far less satellites than the quantity you are attempting to jam?

It's much cheaper to just shoot them down.

... with friggin' lasers.

If you have a phased array, you can point it at a series of sats in rapid succession, or even all at once with more electronics.

That is a good point about pinging many sats. However all at once doesn't make sense to me in the context of a jamming scenario

Phased array antennas are mostly for getting programmable beam forming, and don’t have much to do with jamming, which works by destructively interfering with incoming waves. Starlink operating on a different frequency means that GPS jammers won’t be able to jam them out of the box, but presumably if they’re jammers built in the last 20-30 years they’d have onboard SDRs which can dynamically hop frequencies. It’s a neat trick to get GPS from Starlink, but won’t stop nation state attackers from jamming access to GPS.

Jamming isn't destructive interference, which would require knowledge of the exact signal being sent as well as the exact location of the transmitter and receiver. Jamming is overwhelming the receiver with a stronger signal in the same frequency band so the weaker one cannot be received. Think someone screaming over someone else whispering.

Phased array beamforming absolutely helps make jamming more difficult. Jamming is all about reducing the signal to noise ratio in the channel until it is unusable. Directional antennas (of which phased arrays are electronically steerable versions) have more gain in the direction of the desired signal and less towards unwanted signals located in another direction.

There are different kinds of jamming. It's my understanding that interference jamming is fairly uncommon nowadays. If it's not perfect, it mainly gets you a reduction in SNR and GPS is already way down below the noise floor, so receivers are inherently designed to mitigate the effects even if they aren't explicitly designed for anti-jamming. Even when they work, spectral filters exist. Broadband and spoofing are what Russia typically use. The defenses against those benefit from lots of processing power and large, beamforming arrays.

It's an arms race though.

The gps antennas on the starlink satellites will be pointed “up” i.e. away from the earth’s surface. Jamming from the ground to a receiver in space pointed away from earth would be… difficult. Especially when the satellites orbit about every 90 minutes, so you’d really have to do a global very very loud gps attack… it’d be simpler trying to hit several gps satellites with missiles than to accomplish that.

Aren't the Russian drones using GPS? Why wouldn't Ukraine jam GPS?

I'd assume the Russian drones would default to GLONASS

why wouldn't that be just as easy to jam?

It's basically the same frequencies too

It's harder to jam GPS for "just one side" so doing so would harm both side's operations.

Couldn't the US re-enable Selective Availability?

Theoretically yes, but with massive disruption in the civilian space.

DGPS [0] is a workaround for SA. If you can have a ground station with a known location somewhere in the same area - even 100 miles away should be OK - you can broadcast the correction.

DGPS was already "good enough" in 2000 when SA was turned off, so I'd expect it could achieve very close to the same precision as regular GPS by now.

[0] https://en.m.wikipedia.org/wiki/Differential_GPS

Although I wouldn't be surprised if one side is more likely to have dual-band GNSS receivers...

The random Chinese consumer drones, sure. You could also just block wifi to knock out most of them too.

Current block military drones being imported from Iran are not vulnerable to GPS manipulation because they have backup inertial navigation systems.

After Iran steered an American drone using false GPS signals, the US also implemented inertial and celestial navigation systems. https://en.wikipedia.org/wiki/Iran%E2%80%93U.S._RQ-170_incid...

In that article, they claim that the American drone was already using inertial navigation.

> American aeronautical engineers dispute this, pointing out that as is the case with the MQ-1 Predator, the MQ-9 Reaper, and the Tomahawk, "GPS is not the primary navigation sensor for the RQ-170... The vehicle gets its flight path orders from an inertial navigation system".

Your last line seems wrong. American aircraft have had inertial navigation since forever. GPS is a more recent addition. GPS is easily jammed and that's always been known to be the case. Nothing would be solely dependent on GPS.

This is false. All US military navigation systems, including the referenced drone, are primarily inertial and always have been. These systems can optionally accept fine-tuning GPS corrections, but only if those corrections are within the tight (classified) error margins of the inertial navigation system. It is a misconception that the US has ever used GPS for primary guidance in weapon systems; even cheap throwaway weapons like JDAM are primarily inertial. The kind of influence possible via manipulation of GPS is measured in meters. State-of-the-art INS technology is now good enough that the US military is considering phasing out GPS corrections for some newer systems.

GPS was designed for the purpose of accurately measuring the Earth during peacetime to build a precise model the world that could be used for inertial navigation systems in wartime. It was never intended to be used as a critical navigation system since it could be trivially destroyed by the Soviet Union when it was designed. Civilian systems tend to not concern themselves with this vulnerability and therefore happily use it for navigation.

The drones have inertial guidance and gps is used for corrections only. That means when the drone is guided to blow up the railway junction, maybe you throw it off a bit so it hits the apartment next door. Not necessarily a good plan. There was also the case of a drone being damaged by ground fire, and veering off course to hit an office building next to an energy junction that it was being aimed at. And the problem of all those people shooting small arms into the sky causes damage when the bullets fall. For these reasons, Ukraine has asked people to stop shooting at the drones -- the time to shoot down drones is when they are flying over fields towards the city, not when they are already in the populated areas.

You have the same problem with putting AA missiles in populated cities -- sometimes they miss and slam into an apartment/office building -- that has happened 3 times so far at least as publicly reported (most likely it happened more often). You are supposed put the AA missiles in defensive rings around your assets, not in your downtown area. Ukraine's problem is they don't have enough to deploy as effective rings -- e.g. they don't have a real AA system, and so they put the missiles directly in populated areas, in which case it's better to turn them off when the cheap drones fly by and save them for enemy jets. Those missiles are not cheap.

Really the drones are exploiting the fact that AA missiles are so expensive and there is a gap between the guy with a rifle standing on the hood of his police car and a $2 million Amraam. Even manpads cost far more than the drone. The production costs of one of these is about $1000 (although they are sold for 10-20x that retail) - using parts sourced from Alibaba. A cheap moped motor, a mirocontroller, it's really simple stuff. Like V1 [corrected] rocket tech. Israel is working on the "iron beam" laser defense, but right now, there doesn't exist a working defense that doesn't involve firing million dollar missiles at $1000 drones. It's something all the major powers should be investing in, and I'm worried the U.S. defense complex just isn't able to do "cheap" any more. Could they even design a system that could take out such a drone cost effectively?

You meant V1 rockets? The V2 missiles were a bit more complex.

Yes, V1. It's closer to a flying bomb. I'll correct.

At least one party in the conflict is generating enough interference on the ground to obstruct both GPS and GLONASS navigation in orbits passing over the region. Its loud enough to detect as soon as the region is visible (thousands of miles away). Source: we are operating a constellation of sophisticated GNSS receivers in LEO.

Hawkeye360 (a different operator of sophisticated RF instruments in LEO) detected GPS interference and geolocated it to Russian forces shortly before the invasion, but at this point it could easily be both parties.

Satellites don't need extremely frequent GPS fixes though, or do they? I think they can afford waiting half a revolution to correct their predicted location.

It's not clear to me if GPS is necessary here, or just a convenience?

Could you reasonably update the location/timing information on starlink satellites every time they passed over (for example) north america from fixed ground stations? Or do you need to update the satellites clocks more frequently than that or something?

Honestly, and this isn’t something I’d ever thought of until just now, it’s probably reasonable to assume that most Starlink ground terminals (the ones that users install at their homes and businesses) sit in a fixed location. It would be complex to do, but I suspect you could probably get quite good orbital positioning in reverse by using all of the fixed ground terminals as a giant reference set. If the terminals themselves have GPS receivers built-in, the whole system could pretty readily bootstrap itself (each GT sits and continuously averages its own GPS position and clock offset, then transmits that up to the overhead SVs as they pass by). Even without GNSS receivers in the GTs, though, you could probably still rely on the Starlink-owned GTs to know their own positions very accurately and augment the whole thing using the “static user GT” assumption. GTs that did move would just be treated as noise in the whole thing.

It’s been a while since I’ve done this kind of math, but it sounds like a really fun problem!

GPS is also coupled with WAAS which relies on a network of ground stations at known coordinates to determine GPS error. These errors are caused by bulges in the ionosphere that throw off timing. Error is calculated, broadcast up to satellites and then rebroadcast down to GPS receivers that are WAAS enabled. This is known as a satellite based augmentation system. The other satellite nav systems that compete with GPS are building similar augmentation systems (like GAGAN, EGNOS, MSAS and SDCM). These systems are required for precision approaches like LPV for aircraft on instrument flight plans. Regular GPS precision due to ionosphere changes is too low. They've enabled the FAA to lower the minimums for GPS approaches down to levels that are close to ILS, removing the need for expensive ILS installations at many regional airports. So replacing GPS isn't just about creating a system to locate a point in space. You'd need to replace critical systems like WAAS that enable precision approaches during instrument conditions, for example.

It’s mindblowing how many wrinkles there are to iron out in a complex system. Who knew that the ionosphere bulges!

SpaceX for one. When solar storms occur they have to release newly launched satellites at a higher altitude to avoid being slowed down by a puffed up atmosphere. Just recently they lost a few starlink sats to this effect.

Rather more than a few. Managed to save a few, instead. Expensive lesson.

FWIW, that particular problem isn't likely going to be an issue for Starlink. Starlink satellites orbit at 550km vs. 20000km for GPS. The transmission is much (much) closer to a straight line approximation and ionosphere refraction is negligible.

In fact, the line of sight transmission to and from a Starlink satellite is likely to be significantly better than that between an airliner and its local WAAS station, which is confounded by ground reflections. (Edit for clarity: WAAS is just digital data, this isn't an actual error in the signal. The point is more "you have to spend more engineering effort on the transmission environment of the correction signal for GPS than you do on the actual signal for Starlink")

Almost all of the ionosphere's total thickness is below 550km, so the lower altitude doesn't matter. Starlink sees less impact from the ionosphere because the carrier frequencies are higher, and dispersion is proportional to 1/f^2.

I don't think that's true? Plasma density goes up with altitude out to quite some distance, though it's nonlinear. And in any case it's a refractive thing. Having a somewhat lower index doesn't make much difference when one lens is 40x thicker than another.

I would imagine many more satellites to choose from.

(I'm uncertain, lower orbit might give you more satellites, but I don't know how many would be in view - many more or fewer)

> Regular GPS precision due to ionosphere changes is too low.

... for single-frequency receivers. Military receivers have been dual-frequency for a long time. GPS L2C's rollout is pretty slow (and L5 is even farther behind), but Galileo and BeiDou are both fully operational with dual-frequency civil signals right now. All dual-frequency receivers can naturally cancel out the effect of the ionosphere. That's part of why some smartphone baseband chips are coming out with L5 capability: GPS L5, Galileo E5A, and BeiDou B2A all operate in the same spectrum.

Apple Watch Ultra has dual frequency:

"The Apple Watch series has always offered GPS support (in addition to other global satellite navigation systems), but the Apple Watch Ultra is the first to come with dual-band GPS (L1 + L5). Most smartwatches on the market, including the Galaxy Watch 5, only support single-frequency GPS and can only receive satellite signals on the L1 frequency. Dual-frequency support allows the Apple Watch Ultra to lock onto L1 and L5 bands simultaneously. This greatly improves navigational positional accuracy and reduces multipath errors in urban areas and other challenging environments."

https://www.xda-developers.com/apple-watch-ultra-gps-support...

This apparently just came out, was thinking they had this out for some time.

And then there's RTK (I like to talk about https://centipede.fr ) that measures the offset between GPS signals and litteral ground truth and proadcasts it, enabling centimeter-level precision.

The more systems, the merrier, even if they aren't all accurate, you can use them to increase precision (kalman anyone?). Someone mentioned multi band. You can also use signals from multiple systems at different frequencies.

I’ve used VRS RTK for outdoor robotics applications and as a geographer I’ve got to say, 1cm accuracy gives me a partial.

Massively impressive! A very exciting paper https://arxiv.org/abs/2009.12334v4 from the first two authors shows how Starlink can be fused with ordinary GNSS to get extremely impressive +56 dB anti-jam advantage (see Figure 1).

Elon on Twitter was saying it can be made much more accurate than GPS, just that it's not a priority for them.

My guess is that the reason they terminated work with the UT Austin researchers is that they've signed a commercial deal with someone else that hasn't been announced yet.

GPS-independent PNT (position/navigation/timing) is a significant area of both market and military/civil government interest right now, and has been for a while. They won't be indifferent to it at SpaceX or at any other organization that operates or plans a constellation.

Much more valuable would be to add software to the sats that turn them into a globe-sized radio telescope.

In principle, the terminals would also be able to participate.

Elon says at lot of things when the day is long

This is straightforward to do if you have some reference ground stations listening to the satellites. That's how WAAS works. If you have ground reference stations receiving and broadcasting offsets, you can derive navigation info from any set of long-range RF sources which have some pattern to which you can sync.

The U.S. Army has "pseudolites" which do this for short distances.[1] These are used as a backup to counter GPS jamming.

[1] https://www.army.mil/article/169033/Pseudolites_preserve_pos...

You can even do this with simple AM radio stations. Tune an aircraft's ADF to the radio station frequency, get a bearing to the transmitter and then do the same with another station. From both bearings you can figure out where you are as long as you know where the transmitters are.

It's conceptually similar to what they did here with the Starlink signal but using very different technology.

I mean I think both directions here make sense: SpaceX doesn't want to invest engineering effort on this when they're not even profitable and there's no profitability to be seen from navigation, while researchers can do a little reverse engineering and make something workable without the need for that engineering investment.

If the DoD really wanted it bad enough, it could be very profitable. I take the situation to mean that while there may have been some interest from the DoD, it wasn't enough to make it worth SpaceX's time.

I don't think the DoD and GPS are enough revenue to offset the massive CapEx of building Starlink. Their ROI lies in providing internet to the whole world, so its's reasonable they're not allocating time/money on this.

And being laser-focused on a particular market/model is a critical part of any early business succeeding.

There's no profitability from navigation? Starlink could charge a massive premium for high resolution GPS.

But GPS is already so widely deployed and you can't really compete with free.

If someone starts taking out the 24 GPS satellites - a new anti-satellite weapon, or an interceptor satellite like the X-37 is speculated to be - being able to do it off Starlink's thousands as a backup would be extremely appealing to the US government.

(After all, this is the same organization that's happy to pay out the nose to keep ULA alive to have redundant launch options.)

GPS satellites don’t sit in LEO . All ASAT weapons demonstrated so far by any country has been only for LEO based satellites.

That would be why I said "new", yes.

Given that we can get satellites to GEO, we can presumably get one that goes boom up there, should we be so inclined.

It is not that simple, the high orbits mean it is very very difficult for solid propellant based ICBM variants to reach with a meaningful payload, also given the orbits it is much easier to perform early detection and make evasive maneuvers.

Also remember even if you did launch on liquid rockets, It is not easy to launch them in quick succession or with stealth. Liquid rockets needs a lot of time to load fuel, needs a ton of auxiliary equipment( cryogenic fuel storage at-least) and have limited shelf life once loaded which means they are limited to few known sites which you can quickly take offline after the first attack/launch.

Solid rockets/ICBMs can be launched from variety of platforms some of them very mobile and also given their lower footprint - Silos and other sites can be hidden and missiles moved around to retain the strike capability.

ASATs are primarily designed to target low flying spy sats[1], taking few key ones out can at the right time can potentially deliver few hours of advantage in a battle before alternative sats can be rerouted. Communication and navigation sats on the other hand are designed to operate even during normal times when few go offline without loss of operational capabilities .

[1] Also this is much easier to justify domestically and globally than say communication ones even if they were dual use.

The US military would come down on them pretty hard.

We already have premium high resolution GPS. Its for military use only.

I don’t think that’s true any more. IIRC, the high resolution codes were disclosed to the public in the late 90s / early 00s, and there have been no efforts to rotate them since.

Extra precision can be achieved with fixed-point augmentation signals, which I believe is common at airports and construction sites. I would assume the at militaries similarly augment signals in theaters of war. But that’s different than some separate high-resolution mode.

At this point, so many civilian services depend on the high-resolution data that I’d be pretty surprised to see GPS going back to a two-tier system.

A citation for the above:

“In May 2000, at the direction of President Bill Clinton, the U.S. government ended its use of Selective Availability in order to make GPS more responsive to civil and commercial users worldwide.

“The United States has no intent to ever use Selective Availability again.”

https://www.gps.gov/systems/gps/performance/accuracy/

There is already a 2 tier system. Clinton turned off "Selective Availability" but kept the P(Y) code for military only use. The P(Y) code now has been supplemented by an M code for military on the newer generation of GPS.

The way GPS was designed to work originally is two sets of codes. Course Acquisition (CA) code which repeats at 300ms intervals with a deliberate bias was designed to bootstrap the receiver into locating its position in a 1-week Precise (P) long code. This bias was selectively worse in different regions and I think still can be changed in war zones. Happy to be educated further here!

What the US government did was remove the bias from the CA code so it could be used for precise positioning. The military still uses P codes as well. I believe there is a small gain to be had but it’s due to frequencies.

Since then there have been several more advances, mostly to broadcast local augmentation signals. Wide Area (WAAS) and Ground Bases (GBAS) are common in receivers.

L5 is a new band to help solve multipath error in urban areas.

Most receivers also have remote autonomous integrity monitoring, where it can predict its own area of probability (by using groupings of 4 in 5 satellites), and if it’s too large for the intended use case, alert the user. Also with 6 satellites it can calculate combinations of 5 satellite groupings to work out (and exclude) faulty satellites. This is Fault Detection & Exclusion (FDE).

Mobile devices will also download their own separate high resolution almanac and ionospheric data over the internet which is superior to the low data rate GPS almanac. It can also use known cell locations to approximate its position. Combined, this enables rapid (hot) signal lock immediately onto the correct satellite code & Doppler shift frequency, which is why your mobile gets a fix in 3 seconds, versus your car which takes minutes.

> What the US government did was remove the bias from the CA code so it could be used for precise positioning. The military still uses P codes as well. I believe there is a small gain to be had but it’s due to frequencies.

The precision of positioning from the P code is 10 times greater than the C/A code (about 30cm vs 3m). This is due to the wavelength/'chip length' of the code signal which is modulated onto the carrier wave (10.23 Mhz / 29.31 m wavelength for P code, 1.023 Mhz / 293.1 m wavelength for C/A code). Positioning precision is limited to about ~1% of the chip length by signal processing.

This is correct, but I'd like to add that at some point the errors from which frequency and code you use is no longer the dominant factor in the position error. Depending on where you are, either multipath errors (eg due to reflections from buildings or mountains) or athmospheric errors (ie due to the radio signal being distorted in the ionosphere) start to dominate.

Selective Availability being phased out and is no longer a feature on the most recent GPS satellites:

https://en.wikipedia.org/wiki/Error_analysis_for_the_Global_...

> We already have premium high resolution GPS. Its for military use only.

So what, throw down an RTK pod and now you have like centimeter level accuracy. That's stuff anyone can buy for COTS drones, I'm eyeing on it for my DJI drones.

The only thing needing realtime in-flight accuracy of that level without an RTK pod is weaponry and maybe cars outside of road (because on road, they can augment GPS with camera data and road mappings).

Even without RTK, the new dual-band receivers have 30cm accuracy. That's enough for any nefarious activity I can think of.

> The US military would come down on them pretty hard.

Or maybe they'd just buy an exclusive contract.

> The US military would come down on them pretty hard.

Showing a clear lack of understanding how the US government works. The US military can't actually do this.

high resolution gps already exists. lookup WAAS

https://en.wikipedia.org/wiki/Wide_Area_Augmentation_System

"Currently, WAAS satellite coverage is only available in North America"

There is also the European version of this EGNOS, besides there also private Providers like TerraStar

WAAS is nice but dual-band receivers provide similar accuracy levels without needing to see that satellite. Ublox quotes on their ZED-F9P datasheet 1.5m accuracy for GPS+GLONASS+Galileo+BeiDou, increasing to 1m accuracy if using space-based augmentation (WAAS, EGNOSS, MSAS, GAGAN). Even without a SBAS, if you can do RTK, you can get 0.01m accuracy. (The corrections from a nearby reference station are even better than what you can get from a satellite.)

RTK is probably more available than people think. My state offers a public network of continuously-operating reference stations; you can sign up for a free account and then do whatever RTK madness you desire. https://cors.dot.ny.gov/ if you happen to be in New York.

(I realize now that I really wanted to reply to the person complaining about not being in North America, but oh well, maybe they'll find this.)

But the state-run reference stations are generally too far away from you to get that sweet sweet 1cm accuracy...

It's amazing what you can do with consumer-priced gear these days. I set up a Sparkfun ZED-F9P breakout board as a fixed beacon on my roof, and then their "RTK Facet" as the rover to do precise measurements to create a map. I could have done the basic thing I needed to by hiring a surveyor or eyeballing things with a tape measure, but this is much more general.

The GNSS software world seems to be a mess though, ripe for a paradigm shift. For example, QGIS seems to be based on flat projections with transformations rather than 3d-native - from what I can tell, QGIS seems to consider the "degree" to be a unit of length measurement! This leads to ridiculous things like being able to accidentally measure a nonsensical "cartesian" distance between two points that differs from the actual distance by a factor dependent on latitude.

I've still got to tidy up my own pipeline that lets me do things like turn N (point, distance) samples into a single point. I would have thought that type of operation would be common, but thinking about how surveyors work I guess they're usually locating points optically, rather than trying to position a GPS receiver at the point to be measured.

Yeah the software is ... interesting. I'm actually less than a mile away from a reference station, but don't get 1cm accuracy because I don't have a much of a sky view from my apartment's window. (But I'm moving soon and have a great sky view. And am closer to the reference station!)

Another thing you might find interesting is that you can generate a report on how good your reference station is. It's actually in Sparkfun's documentation, so you're probably aware, but if not: https://learn.sparkfun.com/tutorials/how-to-build-a-diy-gnss... Specifically the part where you collect data with u-center and upload the results to https://webapp.csrs-scrs.nrcan-rncan.gc.ca/geod/tools-outils... for analysis was very interesting.

GIS software is indeed messy, but not because the people writing it don't know what they're doing. You seem to be approaching things from a Cartesian perspective, but GIS almost always works in spherical coordinates because the earth is (approximately) a sphere. It makes a lot of common operations easier too.

Imagine you go on a road trip (along the surface of the earth). How far have you driven? In spherical coordinates, that's just changing two angles. In Cartesian coordinates, it's an ugly mess. Doesn't hurt that it's a lot easier to measure angles in surveying than distance.

However, certain GIS systems like QGIS and arcGIS are designed for making maps and have to display things in a 2D space. Thus, they have a projection mapping the spherical coordinates to Cartesian canvas coordinates and back again. This leads to unintuitive behavior, but it's mathematically hard to do better.

Now, the user interfaces and the terminology and the subtly disastrous inconsistencies between different data sources? Hot flaming garbage, all of it. These aren't problems with the underlying data models though.

> You seem to be approaching things from a Cartesian perspective, but GIS almost always works in spherical coordinates because the earth is (approximately) a sphere

Actually no, I'm complaining about the exact opposite. I want to be working in spherical native coordinates, but QGIS seems to treat "degrees" as just another fixed unit of length measurement rather than an angular measurement!

For example I had the measurement projection set to "cartesian", which I would have expected to either give me the linear distance from (X,Y,Z) to (X,Y,Z), or the linear distance between (Lat,Lon) on an approximation of the earth's surface (either sphere or ellipsoid). Instead, it was treating (Lat, Lon) as if they were (X,Y) coordinates on a flat map and doing Pythagorean theorem on the angular measurements, resulting in the longitudinal distance being off by a factor of cos(latitude) !

I can see no paradigm in which such a result would ever be desired, apart from QGIS fundamentally working in terms of linearized projections, with WGS84/spherical coordinates being added on as an afterthought.

> Starlink could charge a massive premium for high resolution GPS.

but then they'd have to make it reliable. We are also not sure how much more (if at all) accurate it is.

sure the bandwidth of the downlink is much higher, and louder than GPS, but the accuracy of the clocks on the satellites is much less. More importantly they are not characterised, so we arn't sure how much they drift due to temperature (both from sun and other effects.)

depending on the navigation type, visual positioning might be better/faster/more accurate. For "military" purposes, silent autonomous navigation without radio sensors is pretty appealing. Using satellite imagery, its perfectly possible to make an accurate, robust visual navigation system

for urban areas, "VPS"s are far quicker and more accurate, but require network access to work practically.

If one way or another Starlink starts providing precise positioning data it will turn their satellites into a valid military target in case of a war. E.g. if China invaded Taiwan, they may take down the part of the constellation that passes over China (if they can cheaply mass produce and launch interceptors).

They could easily saturate the whole LEO where starlink operates with schrapnel - and any damaged sat would add more schrapnel, as described by the "Kessler syndrome".

There's nothing easy about what you propose. It requires launch capability similar to what SpaceX has, i.e. state of the art.

Given that they have a working space station, I'm certain China has enough launch capability to make the Starlink orbit hell with several well-placed rockets undergoing very dirty disassembly in orbit.

(In fact, among the things that makes that scenario less likely is the fact they have a space station and would like to keep visiting it without worrying about passing through a Kessler cloud).

A space station doesn't mean you have a lot of launch capability.

Its and its not clear how you standard rockets can easily be used as anti-sat weapons, and if that makes finical sense.

US had a working space station for over a decade without the ability to put people onto that space station.

This is true, but seems irrelevant because China is self-crewing its station.

A Kessler cloud would not persist in orbit at the altitude of Starlink satellites for very long.

And building and maintaining a Kessler cloud at that altitude seems like it would be… not effective with just a handful of rockets due to the significant atmospheric drag.

Kessler syndrome is a real concern, but only at higher altitudes where atmospheric drag is negligible.

> Given that they have a working space station, I'm certain China has enough launch capability to make the Starlink orbit hell with several well-placed rockets undergoing very dirty disassembly in orbit.

Yes and their "working space station" is also below that orbit. So they'd need to blow up their own station as well. Also the international space station as well.

That would be quite dumb. Whatever you put in that orbit will decay within a few years.

Given the current situation, USA/SpaceX is in a far better position to quickly repopulate LEO with satellites when the debris has fallen down. In that case, they may send up satellites armed with weapons that can shoot down anything being launched into orbit to create debris again.

I don't think entering into that kind of conflict with USA is a winning proposition for China unless they have their own Falcon 9 or Starship-like rocket.

> In that case, they may send up satellites armed with weapons that can shoot down anything being launched into orbit to create debris again.

What happened to the outer space treaty?

Ripped up as soon as you started intentionally trying to create Kessler syndrome...

In some hypothetical hot conflict where a major belligerent nation had already started taking down satellites and/or polluting orbits, the treaty will have already become a 'dead letter'.

As long as they're not using nukes I don't think compliance with the OST would be an issue

Let me guess: You're one of those people who said that the US Space Force "violated the treaty" because it "weaponizes space", right?

The treaty bans weapons of mass destruction in space. It does not ban military activity in space at all.

> Given the current situation, USA/SpaceX is in a far better position to quickly repopulate LEO with satellites when the debris has fallen down

It's far cheaper to heft tens of thousands of ball bearings into orbit vs a single satellite. You can give them a nice spread so you have a space shotgun that ruin an orbit for years at a time. The decay is a bonus to the attacker[1] because they can a go all out during wartime, without impacting their long-term space-faring program.

1. The decay also allows the same armaments to cover a larger vertical slice of the orbit.

You mean like Project West Ford? https://en.m.wikipedia.org/wiki/Project_West_Ford

I was thinking having the projectiles being aerodynamic may increase decay time, then realized you can have the launch vehicle use solar power to fire the metal spheres using a railgun mechanism. This way, you can get vastly larger areas of denial by firing at an angle perpendicular to that of travel. You can have an every-shifting debris field if you fire 2 ball-bearings (port and starboard) every 10-50 meters to make that entire altitude unusable[1] - not just a specific orbit.

Edit: 1. The perigee and apogee will differ for projectiles fired from port or starboard based on launch vehicle inclination and the resulting relative speeds to earth. It would be a nightmare to track and avoid the resulting mess.

> USA/SpaceX is in a far better position to quickly repopulate LEO with satellites when the debris has fallen down

Until china threatens musks other factories and he voluntarily decided that defense isn’t a good business for SpaceX. Or maybe he’ll do what he did with Ukraine and advocate for china just to end the war (probably due to business risk of Tesla et al).

Tesla might be the most profitable business Musk is involved in, but SpaceX is nearest and dearest to his heart. It’s worth more to Musk than one factory which, after Berlin and Austin finish ramping up, will represent just one quarter of their global factory capacity.

IMHO, I think would be more likely that Musk would divest his entire ownership of Tesla than do anything which would compromise the long-term mission of SpaceX.

Do you think you know the long term mission of SpaceX? It certainly is not colonizing Mars, no matter what Elon claims. Nothing SpaceX is known to be working on is anywhere close to adequate for that.

My own guess is it is to get the US military dependent on him. Then he can write himself billion-dollar checks. NASA is just for practice.

I don’t think he’d have that luxury. Especially not if a situation snuck up as they often do in war. Not that I think any of this is likely.

25% of their factory output is a lot. Sure it’s an amount that won’t leave the company in ruin, but it’s a lot. Teslas growing stock price basically funds all of spaceX so divesting is unlikely.

Realistically, SpaceX filled the mission musk had. It proved that space is viable with better tech. They’re really beholden to the government and he won’t be able to shake that.

It is simply wrong on the evidence that Tesla’s stock rise funds SpaceX. It is a profitable business and funds itself. And SpaceX doesn’t hold any Tesla stock as far as I’m aware.

My point is that he would divest ownership of Tesla before allowing himself to be blackmailed over SpaceX. I’m not saying that he wouldn’t care if the factory in China shut, though I am saying he might prefer that over SpaceX being neutered.

> It is simply wrong on the evidence that Tesla’s stock rise funds SpaceX. It is a profitable business and funds itself. And SpaceX doesn’t hold any Tesla stock as far as I’m aware.

No but musk owns Tesla shares and as they rise he can invest new cash into it.

He could, but so far he hasn't, AFAIK. You might be thinking of Jeff Bezos who pours something like a billion dollars a year into Blue Origin.

You said "Teslas growing stock price basically funds all of spaceX" and now you're completely changing your story to something else which is also speculation based on nothing. It would be appreciated if you actually admitted that to yourself and perhaps in an apology here.

Kessler syndrome is not "easy" to create at 550km. The orbits decay way too fast. Whole satellites decay in 5 years or less; smaller debris decays faster due to higher surface area to mass ratio. And the orbit doesn't have to decay all the way to 0 for the debris to stop being a threat to Starlink; a few km is all it would take. I haven't seen a lot of calculations about this but my belief is that even intentionally creating Kessler syndrome at 550 km would be infeasible, and it certainly won't happen by accident.

The point really isn't to completely shut down that orbit, just deny it for however long you strategically need. Throwing up a huge flak cloud using magnetized cheap metal tinsel on an old ICBM isn't only easy, but also has actually been done before, albeit for different reasons

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

You won't be able to cause Kessler syndrome promptly. It takes a while for the collision debris to build up, and it takes a while for individual satellites to get hit. Space is big and even huge debris clouds will take a long time to hit something. It's not going to help your invasion next month or anything like that. Very impractical as a weapon of war, even if it was feasible which it isn't.

Space is big. Really big. You just won't believe how vastly hugely mind-bogglingly big it is. Any given orbital plane is larger than the entire surface of Earth, including the oceans. Combine that with the fact that explosions and collisions introduce randomness to the orbital plane of each piece of debris, and you’ve got a lot of hurdles to overcome in order to get to some short term low earth orbit Kessler.

That just makes it more likely. If you think war is going to last awhile, or you have a severe disadvantage, you shut down access to space for a few years. The cost of such an attack is lower BECAUSE it resolves itself rather quickly.

Though the issue is that it is substantially harder to create the Kessler Syndrome than people claim.

I disagree that it is at all likely. It would take months if not years worth of launches to fill an orbital shell with debris and nobody's going to do it before starting a war. It's practically a declaration of war on its own. It won't happen at the beginning of a war because nobody starts a war with the intention of having it drag on for years. If you're winning a war you're not going to do it and if you're losing a war you're not going to have the resources or time to do it nor the ability to do the launches without getting your launch pads destroyed. Furthermore it doesn't shut down all access to space, as you can launch through the cloud to a higher orbit.

> I disagree that it is at all likely. It would take months if not years worth of launches to fill an orbital shell with debris

Aren't we in agreement?

>> Though the issue is that it is substantially harder to create the Kessler Syndrome than people claim.

Starlink already is a valid military target in time of war.

Also, the way the orbits are established and the fact that Taiwan is roughly equatorial, I thin every satellite eventually passes over it. Starlink is not geostationary

I would be more worried about the undersea fiber optic cable. There's 13 cables that provide the vast majority of internet into the country and they're deep underwater where they could easily be destroyed with no defense.

See Nord Stream 1 and 2 for an example.

I'm guessing it's a bit easier to repair/splice a fiber break than repairing gas pipes

Harder if they roll up the cable, or just drag the cut end a mile away.

Every part of the constellation passes over China. The Starlink satellite constellation is practically invulnerable to physical attack because it would be prohibitively expensive to hit thousands of individual satellites with interceptors. And SpaceX can keep launching them, 50 at a time (potentially hundreds at a time once Starship is working, along with an increase in constellation size to 30,000+).

Physical attacks on the ground stations are more feasible from a physics perspective, but now that the satellites have laser links you'd have to take out ground stations all over the world to completely cut service, not just locally.

The way to attack Starlink would be hacking, either of the command and control system or the user terminals. Failing that, then jamming, and/or anti-radiation missiles targeted at the user terminals. Russia is known to be trying hacking and jamming already (recall that they were already successful in hacking Viasat at the very beginning of the war). I haven't heard about them locating user terminals by their transmissions but I'd be shocked if they aren't trying that too.

China already has an extensive network of fixed location coordinated signal jammers deployed across the country. It is regularly used to overpower Radio Free Asia and Voice of Tibet broadcasts, and has conducted exercises blocking GPS.

But a much more likely response would be equipping local police with direction finding equipment for signals in the 10-12 GHz band and bashing in the skulls of anyone found with a terminal.

The military threat of Starlink to China is as a way for the Taiwanese defense to communicate during an invasion. In peacetime they can prevent Starlink from operating in China simply by threatening Elon Musk with the loss of his Shanghai factory.

Starlink seems to be protected by the fact that it could launch satalites cheaper than the adversary. That advantage will hold until the opponent gains the ability to launch small satalites at a low cost. Eg. Opponents launches and pre-positions mini suicide satalites 100 at a time in orbit, leading to a weaponization of space.

Or it might be easier to just pressure Musk's other business interests.

Not really. To destroy satellite you need suborbital rocket with much lower speed. Usually it is a missile launched from airplane. Even WW2 V2 missile could propably do it, if it had navigation.

Even so, SpaceX can launch 50 satellites in one launch, and each suborbital interceptor can only hit one. And the "if it had navigation" part is doing a lot of work in that sentence; a satellite interceptor kill vehicle is not easy to build. Are 50 of them cheaper than one Falcon 9 launch with a 15x reused booster? I think "maybe" is the only possible answer, but probably the cost of killing all the satellites that way would be around the same order of magnitude as putting them up in the first place. You would need to build multiple whole factories just to make the thousands of interceptors you would need.

I wonder how hard it is to take out multiple starlink satellites right after launch, when they're all on essentially the same orbit?

It's an interesting idea. They are deployed and start spreading out within minutes of launch. You might be able to take them all out in the first couple of orbits. I expect it would be much harder after that. SpaceX could possibly defend against this by varying the orbit so you don't know where to place your interceptors, and maybe by having the satellites boost apart from each other sooner.

But you'd also need to take out most of the on-orbit satellites before it would be useful to blow up the replacements.

Seems like it’d be easier/more effective to take out the launch vessel.

Sure, but you'd have to be relatively close to the launch pad. The advantage of doing it in orbit is you can do it from the other side of the world.

Since 500 terminals a month are being destroyed in Ukraine, per Elon, it seems clear they are targeting them.

I would have assumed most advanced militaries would have a maser or similar capable of cooking civilian solar panels at 500km by now.

Recent actions of Russia in Ukraine have shown how vacuous the term “valid military target” is, to the point that there’s very little sense in discussing it.

(One may argue that along with that it also exposed how empirically inadequate are a lot of other terms, institutions, conventions, and rules we’ve become accustomed to relying on for maintaining our peace, security, and ensuring that bad guys don’t go unpunished, but i guess that would be a digression for another time).

The meaning behind such terms, absent methods of enforcement when they are violated, exists only for those warring states that choose to respect them - the list that likely does not include any of the realistic opponents we consider today, such as Russia, or North Korea, or Iran, or unfortunately, China.

There are a few technical reasons I don’t see this as a real threat:

1. There isn’t a particular part of the constellation that passes over China. There are probably a few launch groups that never or rarely pass over China, but a majority do. An adversary would have to destroy or disable a few thousand satellites.

2. Anti-satellite weapons aren’t nearly plentiful enough and given the ground support required, I’d be surprised if launching more than a few per day is feasible. A counterattack would come too quickly.

3. Precise positions don’t help that much. Even knowing a satellite position to ~2m still requires some active tracking on the interceptor. It’s not much benefit over knowing the position to 1km.

4. The debris created would be catastrophic and likely to damage the ISS even at Starlink’s low altitude.

All of the first three apply to GPS as well even though it’s only ~32 satellites.

Any communication systems are valid military targets in case of war. So much so that they are usually among the first targets.

But there are also other considerations. In your example I doubt China would want to provoke the US by shooting American-owned satellites out of the sky, and I doubt Starlink would be that important anyway.

TV satellite signals have been used like this for decades. I sat through a company pitch on this and one of the advantages was it (TV, not Starlink) penetrates buildings and for example underground parking lots way better than GPS signals.

I've been wondering how precise Starlink could get with positioning if they put some effort into it. Maybe add some dedicated hardware to the much larger 2.0 batch of satellites. If they felt they could surpass what is available via GPS, I could see them integrating Starlink uplinks into Teslas to aid their self-driving efforts. Adding always on data would be a bonus.

Yes! This could be so exciting. Given that starlink satellites are much more numerous and much closer than GPS satellites, I wonder what the lower bounds are for possible accuracy. Imagine a globally available positioning system with 1cm accuracy and 30 updates per second... (compared to 1 time per second for GPS) This could be a revolution for VR, robotics, logistics and so many other fields.

And for cruise missiles.

(Maybe you should read why GPS is not more accurate: it's not a technical limitation, it's done on purpose.

Also try to read up on Chesterton's Fence.)

This changed over 20 years ago, GPS no longer has artificial limitations for accuracy.

You can't get commercial GPS receivers licenced unless you restrict the altitude and speed they operate at, but a dedicated self-built receiver technically doesn't need to have these restrictions. There are SDR+software projects that do this, which could technically be used for ICBM guidance with no restrictions.

The restrictions are for high speed and high altitude, not accuracy. You can use the GPS module at full accuracy if you aren't really high or going really fast. Model (high-powered) rocketry trackers occasionally run into this, but usually not an issue since you need the location data to find where it lands, and it's going pretty slow under chute.

More precision would require an atomic (hydrogen or cesium) clock on each satellite, and periodic adjustments for relativistic effects, just like regular GPS and other GNSS systems.

An intermediate improvement might be a small/lightweight rubidium clock synced to GPS to improve stability for when GPS is unavailable to the satellite.

It's a neat trick, but I wouldn't rely on it for the same reason I recommend against relying undocumented behavior in an API: if Starlink isn't intended to be a global positioning system, then they can change anything in the protocol at any time to improve its utility for its intended purposes as the expense of its utility for GPS.

That's annoying if your RPC library is now broken because you assumed order of unrelated events wouldn't change. In a GPS it can direct you off a cliff.

This is good advice, for sure, and I've received it from wise colleagues in the past.

Where I land on this is: Undocumented behavior can be useful if you're doing something which is short in duration, and narrow in purpose. Don't build a product off it, though.

I guess in this case, it could be useful for a fallback positioning mode for the military or something.

More and more devices already have GPS L5 chips which is accurate to a foot

The problem is altitude. I can't get my watch or phone to give me proper altitude anywhere. Within a meter would be amazing forget a foot.

https://barbeau.medium.com/tl-dr-dual-frequency-gnss-on-andr...

https://barbeau.medium.com/crowdsourcing-gnss-capabilities-o...

https://docs.google.com/spreadsheets/d/1jXtRCoEnnFNWj6_oFlVW...

Try a WAAS or more general an SBAS receiver, they give accurate altitude information. Accurate enough to land an aircraft with.

WAAS is fascinating. Interesting that older Garmin watches used to have WAAS but because it's a "north america only" ability it seems to have been discontinued after the first Fenix/Epix

https://www8.garmin.com/aboutGPS/waas.html

WAAS is North America only, because that's the specific name of the local system. The generic term is SBAS, which works in Europe, Japan, India, Russia, China and soon also Australia and parts of Africa. It's used by modern airplanes to do GPS (GNSS) approaches, so expect it to be rolled out everywhere over time.

Some systems/countries are listed here: https://en.wikipedia.org/wiki/GNSS_augmentation

> And so we [SpaceX] have to focus completely on staying out of bankruptcy. We cannot afford any distractions.”

Looks like Elon got them to build it for free.

I had an inkling that Todd's team would be behind this. They do awesome work!

Oh probably shouldn't be sending those the iranian revolution then.

Well… if it is that easy, might as well make the sequences proprietary again (change them for security), but it should not cost very much after all to add such functionality. At least not nearly as much as Elon thought.

I think the article makes it clear that SpaceX does not care much about security here. It's more that they don't have (engineering) bandwidth to actually put something together and sell/support it.

Focus is a very real thing in business, starting an additional product that has a minimal overhead can still end up ~2-3 years later as a massive cash drain with a complex hierarchy of workers with a very small ARR. That's why you take on customers before building it.

The one note I will mention about SpaceX and caring about HW/terminal security is the subsidized cost. SpaceX is retailing the terminal I bought for $500, while I understand the hardware all-in is north of $2000.

In these scenarios you will usually see the vendor default to locking down the platform/hardware, if for nothing else to prevent people from buying it and re-purposing it due to subsidized components included.

Do Starlink sats have the capacity to store IPFS data?

I'd assume not, at least no more than any other network router has the ability to do so.

If it did; which wouldn't take much - it sure would change what he was building.

I'm posting too fast so I'm editing this: If Elon slapped storage into each of those sats he could use a system like IPFS to store data and create say... a world wide censorship-proof social media company that not even governments could block (without direct action).

I... What?

That's how GPS works, it was initially intended for something else --- was it for detecting nuclear blasts?

I thought the idea for GPS went back pretty much to Sputnik. People realized pretty quickly they could track it by timing its signals from a known location on Earth and then also quickly realized you could reverse that and figure out where on Earth you are once you have enough timing signals from known locations in space. It just took a bit for the computers to get small enough and for us to put up the satellites.

That's correct and led directly to the US Navy's doppler-based Transit GNSS that operated from 1961 until 1996.

Navstar / GPS emerged as a joint project in 1973 based on individual services' research into a better form of GNSS.

Not for detecting nuclear blasts, but rather for directing them. The primary motivation for Transit/NNSS, the first satellite navigation system, was to provide location information to missile submarines so they could calculate firing solutions for SLBMs. The submarine needs to know where it is so it can tell the missile where to go.

Detecting nuclear blasts from space came shortly later with Project Vela. Those satellites didn't depend on something like GPS because the position of each satellite at any given time could be calculated from its known orbital parameters; no need for radio navigation.

> The submarine needs to know where it is so it can tell the missile where to go.

That reminds me of "the missile knows where it is because it knows where it isn't"

https://www.youtube.com/watch?v=bZe5J8SVCYQ

“OFDM is all the rage,” says Mark Psiaki, a GPS expert and aerospace professor at Virginia Tech.

Did a Virginia Tech professor really say "it's all the rage"?

Reply on news.ycombinator.com