London Underground hosts tests for 'quantum compass' that could replace GPS | Hacker News Reader

London Underground hosts tests for 'quantum compass' that could replace GPS

theguardian.com

198 pointsbeejiu134 comments

I remember doing laser cooling of rubidium atoms in physics class. It's really cool how it works:

* Rubidium will absorb photons of a certain wavelength and re-emit it in a random direction. Since photons have momentum, it will get some net momentum change from absorbing photons all coming from a particular direction, but the re-emissions are in random directions so they have net zero momentum.

* Shine laser light at it from all directions but at a wavelength slightly longer than the wavelength at which it will absorb. Now, if the rubidium atom is moving, the light hitting it head on will be Doppler shifted into the wavelength that it absorbs, slowing it down, whereas the other light wouldn't affect it. So, no matter how it is moving, it will slow down.

Rubidium is a good material since it has an S shell on the outside as though it were a big hydrogen atom, but it is so massive that it is a lot nicer to work with for this application.

>> It's really cool how it works

Pun intended? Please say yes.

Maybe I'm missing something but hitting a gas just above absolute zero with lasers from every direction doesn't sound as something that could be miniaturised.

Most technology has a limit like this (for instance CPU transistor size in the past). However once this is a proven technology, money will be invested in solving these issues.

Theoretically you might be able to do it with some VCSELs, not sure though...

Maybe not down to a pea size, but possibly down to an orange size. Lasers are tiny, and "every direction" does not involve many of them, it involves some fiber optics, I suppose. The bulk of the device would be thermal insulation, and a volume for liquid helium + some piping to let it evaporate, and to replenish it.

You don't need any helium or nitrogen here, cooling happens only by laser cooling and evaporative cooling from magnetic or optical traps. The atoms are perfectly insulated in an ultra high vacuum. Electronics still take the bulk of the volume here, as does the laser system. While the lasers themselves are tiny indeed, the light needs to be manipulated before reaching the atoms. And yes, it involves quite a lot of fiber optics :-).

For many purposes, "every direction" is only three directions in our 3D space.

It's pretty amazing what's slowly but surely becoming possible. This company, for example, has successfully made chip-scale "fiber" ring gyros with incredible performance that would have been very large heavy units 5-10 years ago: https://www.anellophotonics.com/products/x3

There's a whole bunch of laser cooling science/applications: https://en.wikipedia.org/wiki/Laser_cooling

Does it mean that using this mechanism you can only achieve cooling for a narrow temperature range ?

As I assume you cannot shine laser light at a spread spectrum to capture all the possible ΔT

I imagine at higher temperatures, the lasers that hit at an angle would have the correct doppler shift.

Thank you for this explanation. I've often wondered how laser cooling works and never found a clean explanation.

Great explanation :)

If I understood this correctly, would this allow you to make an extremely good estimation of direction of velocity? If you want to use dead reckoning to estimate position, how would you know the abs of the velocity?

I'm way out of my depth here but I believe this would function as a very sensitive accelerometer. The rubidium would be still relative to the lasers so when you accelerated, it would have inertia and take some time to catch up with the rest of the device. You'd then integrate the acceleration to find the velocity and integrate that to find position.

Yeah, it is an extremely sensitive accelerometer which it calls out about halfway through the article.

"At the heart of the quantum compass – which could be ready for widespread use in a few years – is a device known as an accelerometer that can measure how an object’s velocity changes over time."

Thank you for this, it's the first question that entered my head when I read the article.

>Now, if the rubidium atom is moving, the light hitting it head on will be Doppler shifted into the wavelength that it absorbs, slowing it down, whereas the other light wouldn't affect it

Spooky.

It's like having a Maxwell's demon for photons sitting on top of every atom!

Journalists get it wrong, again. Sigh. It will never replace GNSS. It's quantum inertial navigation (QIN) that double-integrates acceleration like another method, requiring external position, heading, and velocity (re)calibration and drifts without them. It has absolutely no idea where it is from only itself.

From the article: "This information, combined with the starting point of that object, allows its future positions to be calculated."

It doesn’t go into the details of recalibration, I’ll give you that.

Because it can't. There are always cumulative errors in INS. There is no way around this without external references, but then it's no longer inertial navigation and it's inertial-assisted navigation.

An underground navigation system based on triangulation of UWB cells would be a better solution than some nonstarter project the size of a refrigerator that requires liquid nitrogen.

> There are always cumulative errors in INS.

Is that necessarily true of this quantum thing? I know nothing about it except this article, theoretically if it kept track of exact Plank lengths or something, then there would be no errors to accumulate, right? Lots of the things that seem intuitively true break down in weird ways when dealing with quantum effects.

There is no magic involved, so there will still be errors and limitations.

https://physics.stackexchange.com/questions/679991/which-is-...

TL;DR-TL;DR: says the opposite of your implied claim, "atom-gyros are set to outperform light-based gyros"

TL;DR: this is a StackExchange question with 1 answer, noting it is indeterminate if a quantum gyroscope would be more accurate than a laser-atom-based one.

It looks like you rushed through and missed that in this context, TFA is describing an atom gyro.

That leaves conversation at a point where either A) we assume the scientist interviewed knows what they're doing, or B) following your unstated lead, assume they're a crackpot and the whole article is irrelevant because they're untrustworthy, and thus in an ideal world, there's 0 comments on the article.

Problem is always going to be sampling frequency.

All accelerometers tell you is the direction of the acceleration vector (ie how speed is changing and in which direction). You still have to add the individual vectors to derive where you actually are.

And if you don't sample fast enough and your acceleration has frequency components at frequency comparable to your sampling, the acceleration you measure may not reflect where you actually are (ref Nyquist sampling theorem)

Imagine sampling at 1hz, and you just happen to have a bump every 1 sec (eg your wheel happens to have a flat spot and is turning at 1Hz), followed almost instantly later by a bump in the opposite direction. Your sampling only sees (say) the +ve components, misses the -ve and accrues a bunch of error.

If you can sample fast enough, you can minimize this sort of error, but you can't really make it go away.

Oh, btw, if you make it work well enough you're considered munitions for export control purposes, so limits the number of countries you can sell to. Same reason civilian GPS units stop working somewhere around 1200mph

These are all excellent elucidations of classic mechanical principles making it hard. I'm not sure they're enough to make me say the scientist/institution in the article a priori has it wrong, especially because it's not a one-off dude just messing around.

Thank you for this comment, it brought me back to signal processing lectures from many years ago.

> theoretically if it kept track of exact Plank lengths or something,

Sure, just show me a way to measure something in terms of "exact Planck lengths" and we're money.

stares at TFA

Not really sure what to say, here.

Actually, in a train context, inertial is pretty damn high accuracy because you already have the world's best odometer showing (A) how many times your known size wheel has turned on your known track; and (B) you can test many axels simultaneously to ensure no slippage (I guess you wouldn't test the wheels as they move faster creating a higher frequency sampling requirement for ~no benefit whereas a dedicated axel feature could live within an environmental enclosure protected from dust and grime); (C) you are constantly doing exactly the same trip-segments over and over again.

So for a train with an offline positioning requirement, I'd suggest that an odometer based solution is close to ideal.

Re odometer - don't the wheels of a train slide a little bit? I know the train is big pile of heavy iron, but still. My naive thinking is that when braking or during a rush start some extra distance can be covered.

Don't know, just thinking out loud.

If you are averaging axels across multiple carriages slippage is unlikely to amount to much. Moreover, because you are traveling known segments, you can reset your position at each station or known intersection/detectable segment terminus, so you're going to have zero accrued drift at that point. It's a perfect deployment scenario for an odometer based solution. If you want to be higher confidence, sensor-fusion with an IMU, laser TOF/LIDAR, camera, ambient light sensor, radio signals, or MEMS microphone can verify position. Et voila! - no need for GNSS.

Not only do they slide, the "known size" changes not only due to wear but also due to variations in the gauge from metre to metre along the track.

Generally trains do everything they can to avoid slipping. They have anti-lock breaks and traction control just like your car, and stuff your car doesn’t have, like “sanders” which pour sand onto the track just in front of the wheels.

Regardless of material, dynamic friction is always lower than static friction. So for maximum acceleration and breaking it’s important to ensure you wheels stay in “rolling” mode of interaction, and don’t slip.

There is no liquid nitrogen involved here. The instrument from the article is actually rather big, current generations of quantum IMUs are roughly half this size with lots of room for miniaturization.

One big advantage of these atom interferometers is that they actually don't need to be recalibrated because the reference is the wavelength of the lasers which can be controlled with extreme precision.

A big disadvantage is however the limited repetition rate, which is on the order of only 1 Hz at the moment. Currently, combinations with "classical" IMUs seem most promising, and there is lots of interest in these devices for applications in planes, cars and spacecraft.

Surely it can only calculate its current position, calculating its future positions would seem to require a divination machine.

>>> It will never replace GNSS.

It will have to. That is the point. This isn't about better in-car navigation. The big money behind quantum gyroscopes is the potential to guide submarines/aircraft/missiles in times of war when the GNSS systems are down or otherwise unreliable, just like the best of traditional gyroscopes. Dead reckoning is a legitimate means of navigation, but there are also some aspects where actual replacement of GNSS might happen. An extremely sensitive gyroscope could probably determine latitude based on the earth's rotation (Foucault Pendulum). Then layer on a detailed map of variations in the earth's gravity and/or magnetic fields and one might be able to pinpoint a location absent external signals.

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

That's an extremely niche application unlikely to be scaled down to anything smaller than a backpack like laser ring gyros. As such, this type of positioning gear isn't for consumer use and is targeted mostly for underground surveying.

(I worked for Trimble.)

...yeah? I don't think anyone has claimed they're developing this for consumer use.

... that could replace GPS

That's the implied application, which is wrong.

... and GPS was initially developed for - right, the military.

GPS can and was miniaturized. LRGs and QGs aren't for retail use because they're very expensive, bulky, and largely unnecessary.

F(T) != F(T+1)

I, for one, am glad to see people working to innovate along well-trodden paths.

Yes, for their relevant applications. If you were a commercial spelunker or were the Ukrainian army needing to lob missiles into Russia that were impervious to encrypted GPS jamming, then you would gladly welcome your quantum overlords. Neither LRG or QG are ever going to be made into MEMS devices shoved into an iPhone. Its application is dead reckoning. For the use-case of navigation underground useable by everyday people, Skyhook-like services that rely on 5G UWB microcells are the most likely evolution beyond relying on Wi-Fi SSIDs and conventional cell towers for tower-assisted GNSS.

Why does it matter if it's not for consumer applications? GNSS is used for many more applications (and arguably more critical) than consumer applications (agriculture, mapping & surveying, aeronautics, shipping, etc.)

Also just want to mention that, yes, integration errors accumulate when using intero-receptive sensors but if errors are small enough (white noise, various biases, sample rates, quantization, etc.) from the inertial sensors an odometry solution might be adequate until an extero-receptive sensor can localize the sensor within an external frame.

This can shift the discussion from solving a problem that has no solution (i.e. how do I integrate a signal with white noise without any error) to an engineering problem (i.e. what error parameters allow the odometry to be accurate within x% over some timeframe).

There was interesting work DARPA was sponsoring around the above idea that you can read more about here: https://www.darpa.mil/program/micro-technology-for-positioni...

>The end goals of the TIMU program are the demonstration of a single-chip IMU which maintains an accumulated position error of less than 1 nmi/hour with device volume of less than 10 mm3 and power consumption of less than 200 mW.

(My job is related to estimating location of things).

nmi - Nautical mile (1 nmi/hour mentioned above)

Today the international nautical mile is defined as 1,852 metres (about 6,076 ft; 1.151 mi).

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

When one of the potential customers in this "extremely niche" space has an $800 billion annual budget, you don't lose sleep over it.

Underground surveying or jam-proof guidance for long range ordinance?

Fair point, this would fit quite snugly on a ship or a sub.

the real ones will probably smaller.

I wonder will having more than one improve accuracy.

Never go to sea with two chronometers, take one or three.

Man with one clock, always knows what time it is. Man with two clocks... never quite sure.

You could combine it with another IMU using a Kalman filter and improve accuracy that way.

So it doesn't solve the issue inherent will all inertial navigation approaches (error accumulates very quickly because of the double integration), it just has less error?

Seems like an incremental improvement at best.

Every time a train arrives in a station, its exact position becomes known.

Not to mention wireless networks like the ones for the airtag or the wifi endpoints mapping google has could be adapted to share positions of "right here, right now" anonymously.

Which means that if only one device knows where it is, you can calibrate, and if several know, you can even correct for mistakes.

If the system is precise enough, you only need to calibrate once in a while. Being at home/office, on the local wifi, once, could be enough.

Besides, nobody wants the GPS to go, but it's a nice alternative that can't be jammed by enemy forces and can be used for hiking, diving, etc.

Yes, but these are approximate.

I mean the exact position, down to a millimeter, when e.g. a photo sensor on a train passes past a LED mounted below the station platform. The LED's light is modulated so there can be no mistake, and its location is precisely known.

A meter precision is fine for what most people use GPS for. If the device is as precise as they say, you won't add much imprecision to the original calibration, so you will be in the right tunnel at the right time.

Reading the article, I was under the impression that this was to be used for stuff like submarines, or vehicles in the arctic... stuff where finding one's position might be tricky.

This is absolutely overkill for a train where you can just rely on a calibrated odometer to know your position.

I've been looking for some data on estimated drift from this, but no specifics. That said, since it's looking at wave properties of cooled atoms, I'm assuming we're talking "wavelength" accuracy as opposed to more conventional inertial nav systems.

My point perhaps is that though this is "nothing new", it's (probably) way more accurate than anything before it. So the "pain points" of recalibration and drift are extremely minimized.

But how large is the error? If it's 1m per day, that would be fine, wouldn't it?

> It has absolutely no idea where it is from only itself.

Neither do GNSS satellites. You can't solve this problem with any method without giving each device a starting reference, and then either continually updating it or enabling some mechanism for multiple devices to vote/agree on reference(T+1).

This new tech might allow other improvements, such as the QIN device being a relatively-static "hub", and the user wearing smaller, cheaper accelerometers that connect to it regularly to reset their starting positions.

> Neither do GNSS satellites

Is it a similar problem? I hadn't thought about how GPS satellites know where they are. Do they need regular correction?

There's some natural drift, and about yearly boost to keep them in their desired orbits

https://space.stackexchange.com/questions/30752/do-operating...

As far as I know - while their onboard clock is pretty good - they continuously get corrections from ground stations. (Once per day.)

https://gpsbeam.com/ground-stations/

the "navigation message" broadcast by the sats is described here https://gssc.esa.int/navipedia/index.php/GPS_Navigation_Mess...

> their onboard clock is pretty good

They use atomic clocks, so it's accurate to something like 1 part in 10^16, or about 1 second in a billion years! A friend of mine is working on the next generation, which will be even more accurate.

In contrast, the force of gravity experienced by the satellite is known to much less accuracy. In fact, changes on a monthly basis due to rainfall/rivers and tides. The GRACE satellites measure the change, although I couldn't find any information on how accurate their measurements are, except that they can measure the distance between the two satellites to within a micrometer (10^-6), so substantially less precision than the clock!

Not quite that accurate. Apparently it’s more like 3 seconds per million years for a rubidium atomic clock, and 1 second per 3 million years for a hydrogen maser. (Caesium atomic clocks are somewhere in between)

ESA use both types onboard Galileo satellites:

https://www.esa.int/Applications/Satellite_navigation/How_th...

Amongst other things, the clocks need to be corrected for relativistic effects.

Due to the net effect of both kinetic and gravitational time dilation, clocks onboard a satellite advance more quickly than they would on Earth (when observed from Earth).

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

Yes, satellites are tracked by ground stations, and receive an update every 24 hours with information about their project positions in the future.

GPS satellites transmit this data in the form of an almanac which includes all the high level parameters for estimating the location of every satellite, and ephemeris data which allows you to calculate the precise location of the satellite, when used in tandem with the almanac.

The almanac doesn’t change too often, but ephemeris data is only valid for a few hours. The satellites recalculate the ephemeris data themselves, and normally have a few months of needed data stored on board, just in case they can’t get updates. But the expectation is they’re updated every 24 hours.

This is also why, in a zombie apocalypse type scenario, GPS would become inaccurate past the point of usability within a few weeks, maybe a few months max.

Even as somebody relatively familiar with inertial navigation, it took me almost the entire article to figure out that that's what this is doing.

I really wish the words "inertial navigation" or "dead reckoning" would have occurred at least once in the article, but obviously "quantum navigation" sounds much cooler and as an added bonus makes it sound more like magic than technology.

Especially with the word "compass" in the title, I figured they made an accurate map of the earth's magnetic field and would use that to avoid needing any satellites. Then I opened the comment thread...

Come to think of it, the earth's magnetic pole shifts so I guess this isn't possible? Either way, "London Underground tests new dead reckoning system" would have been more accurate it sounds like

Regarding pole shift: you maybe could have multiple static measurement stations spread around which measure the current values at known positions and synchronize the data once per day to the trains. I’m not sure how you could even determine position just from magnetic field alone, surely there are multiple positions with the same magnetic conditions?

Maybe, I don't know enough about magnetic fields to say much. In my head they're directional so you'd know how it's aligned and can measure how strong it is, and that might function similar to a gravity map, especially if you know where you're coming from: there may be two places on earth with the same amount of gravity, or in this case, the same magnetic field strength, but if those places aren't next to each other that may still be tolerable depending on the system requirements.

This work proposes a more accurate means of dead reckoning using a quantum-level source for inertial measurement in an Inertial Navitagation System (INS). Useful for robotic navigation underground, inside RF-shielded structures, on other planetary bodies, or underwater. The issue of localization error drift reduction from INS's seems very much like a Moore's law challenge, though I haven't mapped out the advances in intertial measurement accuracy/precision by year to see if the projected rate is similar or not. In any event, pretty cool; we'll see if this one pans out.

The science is cool, but I wonder if a specialized popular science magazine of yester-year would do a better job at presenting it? "We made a high-precision accelerometer that uses lasers and quantum effects; let me tell you everything about it." Right now, it reads more like "Dr. C has this very peculiar quirk when he goes to the London underground. Admittedly, it is not as bizarre as his habit of drinking coffee instead of tea, but going to the underground carrying an aluminum tube instead of a newspaper is just a tad flummoxing..."

So far as I can tell, it's this accelerometer: https://www.imperial.ac.uk/news/188973/quantum-compass-could...

(2018, I'm unclear what's changed since then?)

Russia launched a satellite capable of destroying GPS satellites and suddenly funding became available?

There has been constant development trying to improve accelerometer and gyros.

There are even commercial products, for example: https://m2lasers.com/quantum-accelerometer.html

They claim 10E-7 m/s/s precision, which is is about 60cm of drift in an hour, and up to 375 meters after a day.

Obviously this won't work without re-calibrating to a known location at least a few times per day.

Also changes in the status-quo of international politics: Russia has been chronically sabotaging non-Russian navigation systems in neighboring countries, beaming jamming and spoofing signals across the borders.

Russian airplanes aren't affected because nobody has chosen to retaliate in kind... Yet.

There is no need to jam GLONASS. Russian planes are affected using other means--sanctions have cut them off from access to parts and maintenance. It's more impactful because all planes are affected by this, not just those that fly near the border of NATO countries.

I feel like that is mixing separate issues. The sanctions aren't because of the jamming and spoofing. If some other country started foisting the same pollution/sabotage onto its neighbors, those neighbors would still have a problem deciding on how to respond.

The main point is that the bar for someone breaking your navigation system has been lowered. It's not just something you'd expect in a small area or for a short time around an imminent violent confrontation, chronic disruption in a peaceful country is now A Thing That Happens.

That, in turn, changes the engineering considerations for how robust a product needs to be.

The bar has been crossed. Finland and other airlines flying in Northern Europe are suffering from GPS disruptions that make landing at smaller less-equipped airports impossible:

https://www.reuters.com/world/europe/finnair-pauses-flights-...

https://thebarentsobserver.com/en/2024/06/gps-interference-p...

Disruptions happen over a large area from Finland down to Poland and southern Sweden:

https://x.com/auonsson/status/1772005350134210751

Russians definitely increased their jamming activity after sanctions were imposed. The point I was making is that you do not need to make operation of Russian airplanes more difficult by jamming GPS/GLONASS signals, it is sufficient to deny access to maintenance and parts.

> The point I was making is that you do not need to make operation of Russian airplanes more difficult

I'm saying that "make flights difficult" is kinda-besides-the-point. The real goal is to impose some kind of return-pain to make the offender stop. Poetic symmetry is merely an ideal bonus.

In other words, it may be true that "the ideal retaliation is not very effective because of other circumstances", however that is not the same as "there is no need to retaliate at all."

Planes have inertial measurement units which are the primary source. GPS or GNSS are just backups. You can switch these off with zero consequence during the flight. It may cause your satellite TV and internet to stop working but no critical systems will be impacted.

Russian airplanes are commercially purchased from the same sources as the rest of the world and are generally outfitted with identical equipment.

The impacts to flights has to do with certain types of GPS coordinated instrument approaches for landings. Most of these runways have alternative approach strategies that can be used during jamming or other unavailability.

> Russian airplanes are

The simpler answer is that Russia is not directing the sabotaging signals into their domestic flight paths.

Highly accurate dead reckoning basically. I wonder how it compares to the other method of mapping subterranean patterns and matching to them.

Related: South-pointing chariot, 5th century BC

https://en.wikipedia.org/wiki/South-pointing_chariot

With high enough accelerometer accuracy, you need to start adjusting your subtracted gravity vector with maps, right?

This will become more and more common, see https://rmi.org/insight/the-cleantech-revolution/

Direct link to PDF: https://rmi.org/wp-content/uploads/dlm_uploads/2024/06/RMI-C...

So, basically an extremely accurate dead reckoning system?

It's a shame all the cool stuff is all hidden in a white box, but I suppose it does look a bit like a Hollywood depiction of a homemade nuclear bomb so the British Transport Police would be getting a few panicked calls if it were in a perspex box.

There's also no step-free access at South Kensington or Gloucester Road, so that must be a fun struggle for a grad student!

Is the idea here to accurately track the acceleration that the sensor is undergoing? So, if you started at a known location with velocity=0 and integrate the acceleration data wrt time twice, you get the relative change in position and can thus know your new location?

Looks like it, given (FTA):

“The aim of the Imperial College project […] is to create a device that […] does not rely on receiving external signals.”

and

“At the heart of the quantum compass – which could be ready for widespread use in a few years – is a device known as an accelerometer that can measure how an object’s velocity changes”

So, it’s advanced dead reckoning (https://en.wikipedia.org/wiki/Dead_reckoning).

I wonder whether they really don’t use external signals at all or occasionally use them to correct for accumulation of errors. For example, the ability to detect that velocity w.r.t. earth is zero would help with that.

(Of course, for the London subway, the fact that trains tend to follow the tracks can help, if the system knows the track layout. That probably is the simplest way to prevent accumulation of errors, but then, do they really need such a complex accelerometer?)

They're just using the subway as a convenient moving lab. The real-world application of this technology is for military submarines and marine drones.

My guess is that calculating the initial velocity is hard, so it will drift slowly. Perhapa keep the device on a table for a minute to ensure v=0, or sync it at every subway stop.

Russian shenanigans was my first thought too, but the real game changer is probably:

>under ground and under water

Tunnels, fiber cables, pipelines, submarines, autonomous subs etc would all benefit from less location guessing. (Tunnels tbf you can do with lasers already)

I realize no one would be allowed to tell me if I am correct, but this tech sounds like it is perfect for military use. I am guessing there is already a pretty polished version of this roaming the oceans now.

so this is how by the time of Dune they'll say the compass is english instead of chinese like i was told?

If this was developed in the US this would immediately be ITAR restricted

What are you implying? That the UK technology will be publicly available for others countries to use?

mind the gap!

That smells like bs to me.

Why? It's just an extremely accurate accelerometer. It still needs to be giving a known starting position. It doesn't replace GPS so much as augments it. Granted, the article isn't very clear on that.

Yeah i understand the concept. But i am under an impression that those guys are just milking an easy source of dumb money this way.

Reply on news.ycombinator.com