I found a video UCL put out a few years ago on what the article seems to be discussing: https://www.youtube.com/watch?v=xcqkXkWZhbM

Tangent: I was about to pedantically comment that the correct spelling is "ded reckoning", but TIL that's not true (assuming Wikipedia can be trusted) : [0]

[0] https://en.wikipedia.org/wiki/Dead_reckoning

I was under the impression that you can't avoid drift due to fundamental properties of calculus, but it sounds rad if I'm wrong.

Pretty sure you can't avoid drift due to fundamental inaccuracies in our ability to take measurements. Even GPS satellites synchronize with ground stations to correct their own drift, making the ground stations the frame of reference. And then the ground stations are all drifting relative to each other because of continental drift.

So in the event that your synchronization signals are being jammed, dead reckoning (where you are your own frame of reference) seems like the best you can do.

If drift/noise is low-enough, it can be used for a long-enough period to be viable.

Ring gyros and precision accelerometers have been used in inertial navigation systems for ages. This is likely another step along that path.

Making an accelerometer precise enough to be useful for dead reckoning on a time frame more than a few hours looks like a big challenge. On a ship another way is available - speed of a ship relative ocean bottom can be measured using Doppler effect (which can be used for dead reckoning). And it likely is more accurate.