As space threats mount, U.S. lags in protecting GPS services | Hacker News Reader

As space threats mount, U.S. lags in protecting GPS services

nytimes.com

12 pointsJumpCrisscross28 comments

Open articleView on HN

https://archive.ph/M9OPc

Navies still train in using the sextant:

* https://www.npr.org/2016/02/22/467210492/u-s-navy-brings-bac...

* https://www.cbc.ca/news/canada/british-columbia/canadian-nav...

* https://en.wikipedia.org/wiki/Celestial_navigation

Note that there are 'machines' (Nortronics NAS-14V2) that can look at the sky and return lat/long:

* https://timeandnavigation.si.edu/multimedia-asset/nortronics...

* https://maritime-executive.com/blog/automated-celestial-navi...

For anyone else wondering: celestial navigation (by the stars) apparently works during the day as well, as the stars' infrared light remains "visible" [0]:

> The new systems use infrared rather than visible light for locating stars, allowing daytime navigation. The stars shine just as brightly in the day sky as they do at night, but their light is masked by sunlight scattered by the atmosphere. The scattering is strongest at short wavelengths... But glimpse that same sky with a filter that allows only infrared light, and the sky suddenly becomes dark - and filled with stars.

[0] https://www.popularmechanics.com/military/research/a36078957...

That's how ICBMs do it!

> Intercontinental ballistic missiles use celestial navigation to check and correct their course (initially set using internal gyroscopes) while flying outside the Earth's atmosphere. The immunity to jamming signals is the main driver behind this seemingly archaic technique.

Source: https://en.wikipedia.org/wiki/Celestial_navigation

The folks at the Resilient Navigation and Timing Foundation link to articles on the topic:

* https://rntfnd.org/blog/

* https://rntfnd.org

They recently linked to China's white paper on building a resilient infrastructure:

* https://rntfnd.org/2024/03/01/patton-read-their-book-chinas-...

including a ground-based eLoran infrastructure which is set to be further extended by 2026:

* https://rntfnd.org/2023/11/28/china-eloran-used-for-critical...

Last year, Australian farmers lost automatic steering for their agricultural machinery for a significant time when a single geo-sync satellite, which broadcast a GPS correction signal over the continent, failed.

https://www.theguardian.com/australia-news/2023/apr/23/austr...

Previous HN Discussion: https://news.ycombinator.com/item?id=35643705

Wow, I remember hearing about the I-4 F1 outage affecting shipping and aviation, but I didn't realize it was also transmitting a widely used SBAS!

That seems like a massive single point of failure, given how easy it should be to make that available over at least one other satellite in the region.

The thought of space warfare is a bit scary in how it will affect modern life. Reading this I realized I have an implicit assumption that, in my lifetime, someone very well could blow up some key satellites. That would have a meaningful impact on day-to-day life and the economy.

Hopefully countermeasures and/or back up systems are figured out that add sufficient defense & redundancy before we get to the point where that happens.

Of all the things on my worry about list if any combination of the unholy trinity of superpowers start messing with eachother for real, gps/glonass coverage, Elon Musk and TV is really far down the list.

No google maps is really gonna suck during the nuclear holocaust! /s

Blowing up satellites does not imply we'd be living in a nuclear holocaust.

The day Henry Freaking Kissinger got abatted for "appeasement" I started to hoard bottle caps for some Fallout larping... it is like people have gotten psychotic.

It's a real worry. It's amazing what's connected to GPS that ought not to be. I just happened to notice that modern broadcast TV transmitters have integrated GPS/GLONASS receivers.

We should have a general prohibition against telecom and power infrastructures being dependent on GPS. The NERC looked at this for the power grid in 2016.[1] There was concern, but not enough was dependent on GPS at that time to be a serious problem. This needs to be looked at and tested regularly.

Telecom has been looked at, too. 5G has problems. The relevant report is paywalled.[2]

[1] https://www.naspi.org/sites/default/files/2016-09/nerc_exten...

[2] https://www.atis.org/press-releases/atis-issues-report-gps-v...

> I just happened to notice that modern broadcast TV transmitters have integrated GPS/GLONASS receivers.

For modulation schemes relying on accurate timing, it's very common to use GPS to keep an accurate clock.

>It's amazing what's connected to GPS that ought not to be.

Is there a practical alternative for high-accuracy synchronisation? You could make the case for a backup shortwave receiver, but I'm not sure that's any more reliable than GNSS.

> You could make the case for a backup shortwave receiver, but I'm not sure that's any more reliable than GNSS.

GPS L1 power when received is -128.43 dBm (-158.43 dBW):

* https://apps.fcc.gov/els/GetAtt.html?id=110032&x=

* https://support.spirent.com/index?page=content&id=FAQ14116

* https://support.spirent.com/SC_KnowledgeView?Id=FAQ14116

Meanwhile (e.g.) Loran-C is -60 dBm:

* https://www.qsl.net/df3lp/projects/lfscan/index.html

So GPS signals are below the noise floor, but Loran-C are 50 dBm are above it:

* https://www.prc68.com/I/A2100F.shtml

So while something like Loran-C didn't have global coverage while it was still active:

* https://en.wikipedia.org/wiki/File:NGA-Atlantic_Loran.png

* https://en.wikipedia.org/wiki/File:NGA-Pacific_Loran.png

* https://en.wikipedia.org/wiki/File:Loranstationscrkl.jpg

* https://timeandnavigation.si.edu/multimedia-asset/loran-day-...

it did cover large populations centres, as well as major marine shipping and air traffic routes. Another reason why it didn't the coverage wasn't global was a lot of transmitters (e.g., the continents of South America and Africa).

With regards to timing accuracy, a paper fro 1992:

> The accuracy of time comparison by means of the reception of common time signals, namely, LORAN-C and Global Positioning System (GPS), is examined. Differences of UTC from the emission times of these two radio wave signals measured at four Japanese stations and one U.S. station are analyzed and compared with the clock comparison reports published by the Bureau International des Poids et Mesures. The accuracy of the time comparison via LORAN-C signals of the northwest Pacific chain is about 3.61 μs and that via GPS is 0.14 μs. The accuracy analysis reveals the existence of systematic errors at the individual stations. In LORAN-C, the most serious error comes from those in the delay time calibration of receiving instruments and estimation of the propagation delay over land. Calibration of the clocks by a portable clock improves the accuracy of the GPS method to the accuracy of the GPS signals.

* https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/92RS...

* https://doi.org/10.1029/92RS01010

eLoran can get to the 100s- to 10s-of-ns range:

* https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7697629/

I think CDMA was also heavily dependent on it, since it was using S-CDMA (unlike UMTS/3G).

I think it is important to separate the use of GPS as a distributed clock from using GPS as a navigation system. These are different concerns and there are many commercial systems that use GPS for one or the other exclusively.

The US military has never used GPS for navigation because it relies on state-of-the-art INS almost exclusively. GPS is for building accurate maps, not navigation. It wasn't expected to survive a serious war when it was designed. A challenge for the commercial market is that 1) INS is relatively expensive, whereas GPS is almost free these days, and 2) state-of-the-art INS is considered a cornerstone military tech and US R&D in this area is highly classified and controlled. This makes it a solved problem for the US military but not for anyone else. For many commercial navigation cases you can solve this with sufficiently advanced localization tech (basically, AI) but the tech isn't there yet.

Using GPS as a shared clock source could be solved in most cases by choosing a different clock to use. Before GPS was cheap and ubiquitous, other clock sources were a thing and frankly should be quite cheap now if there was a big enough market for them.

Is that really true? INS/inertial navigation systems accumulate errors over time, GPS is still required for them, its not a replacement?

Advanced INS drifts very slowly, it only needs to meet a precision target for a finite period of time. Military INS only accepts GPS corrections within the (classified) drift model and larger military platforms crowdsource position from a network of INS units distributed throughout the hardware to lower the noise floor. For this reason, the amount of deflection you can induce by spoofing GPS undetected in military systems is measured in meters, which for high-precision work might be enough. The trope that you can make a military system fly off in a completely different direction by spoofing GPS is fiction.

There is a fair amount of circumstantial evidence that state-of-the-art US military INS tech can exceed the precision of non-spoofed GPS for long durations. Some new military system specifications no longer accept GPS corrections.

It is important to note that US military INS is a different animal than what is widely available in the commercial world. It is exotic tech and a closely held secret that has been under continuous R&D for half a century. The ubiquitous laser gyros we use now in high-precision applications were developed for US military applications in the 1960s and eventually de-classified.

Very interesting.

I hope some of that progress will eventually trickle down into e.g. civilian aviation – GPS jamming seems to be a growing concern there.

> Some new military system specifications no longer accept GPS corrections.

Curious: are there specific MILSTDs / STANAGs / whatever that you can cite for this?

> The US military has never used GPS for navigation because it relies on state-of-the-art INS almost exclusively. GPS is for building accurate maps, not navigation.

Huh, interesting. What's the purpose of the (encrypted) Y-code then?

Sure, an INS is essential for situations in which GPS is denied in whatever way, and it would be a problem to not regularly do exercises with it disabled in order to make sure no dependencies creep in. But why not make use of it in peacetime or in a conflict with an adversary not capable of denying GPS?

> other clock sources were a thing and frankly should be quite cheap now if there was a big enough market for them.

Do you have any examples? The only thing that comes to mind for me are the various LF time beacons that some radio controlled clocks still use, like WWVB, DCF77 etc., but these are presumably orders of magnitude less accurate since you need to know your position and propagation conditions at high precision to determine propagation delay.

Article can’t be read without signing up.

That’s all NYT links for the last 5 years or so. Annoying, but someone usually posts the archive link soon.

You can just copy/paste the URL into http://archive.is yourself

Library for my city has passes for patrons.

If one doesn't want to use archive links, might be good to check their local library.

Reply on news.ycombinator.com