Smartphones May Soon Provide Earthquake Warnings (sciencemag.org)

sciencehabit writes: When it comes to an earthquake, just a few seconds' warning could make the difference between life and death. But many earthquake-prone countries lack the seismic networks that would give their citizens the lead time to find cover or shut down critical utilities. Now, a group of enterprising engineers is looking at a substitute network: smartphones. Using smartphones' built-in accelerometers, researchers have invented an app, released today, that they say can detect strong earthquakes seconds before the damaging seismic waves arrive. MyShake, as the app is called, could become the basis for an earthquake warning system for the world's most vulnerable regions.

Comments

Power

[MobileTatsu-NJG]

I actually wouldn't mind buying a usb doohickey that has an accelerometer that sends info to a centralized place to detect and report quakes. I dont, however, think I want to spend battery-life on my phone for it. I mean, wouldn't it need to frequently check the GPS?

Re:Power

[Anonymous Coward]

They don't need a high quality fix, at least not right away, I would assume it just waits for the phone to stop moving, then gets one fix, and watches the accelerometers, or just deals with the lower quality cell phone fixes.

Re:Power

[rasmusbr]

I actually wouldn't mind buying a usb doohickey that has an accelerometer that sends info to a centralized place to detect and report quakes. I dont, however, think I want to spend battery-life on my phone for it. I mean, wouldn't it need to frequently check the GPS?

I think cell-tower positioning would generally be good enough. If you have thousands of users then the positioning errors will tend to cancel each other out. You probably don't need to update the position very often either, since most people spend most of their time not moving very fast or very far.

I believe the main problems with regards to battery use are:
1. Recording a continuous stream of accelerometer samples without keeping the CPU awake.
2. Measuring the power ad other relevant characteristics of the accelerometer signal without using the phone's main CPU, only waking up the CPU if the signal meets your criteria, so that you can pipe interesting signals (and only interesting signals) over the network to your servers for further analysis.
3. Find a good match between false positives and false negatives so that you don't make network requests too often. The cellular radio on your phone is quite power hungry. The cellular radio in transmission mode is a lot hungrier than the GPS.

Basically, you need to be able to run code on a low power CPU/MCU that the accelerometer delivers data to and then have it wake up the main CPU when you detect something that looks like an earthquake signal. If you're lucky the manufacturer already has that code in place for you, so that you can just tell the system to wake itself up if it detect such and such a signal.

Re:Power

[rasmusbr]

But yeah, you could also just take an old smartphone, run it off of a car battery and have the app keep the CPU and WiFi on at all times. The amount of data that you would send if you sent a 50 Hz signal without any compression is on the order of 2 GB per month.

Re:Power

[Sowelu]

If only 5% of users have their CPU and data currently in use, then why wake up other phones? You don't need data from literally everyone who's running it, and it doesn't need to be going all the time. Besides, I bet a lot of phones wake in SOME way when they start moving.

The only useful signal is when a phone goes from "long-term stationary" to "moving": if someone is carrying their phone, actively using it, or riding a vehicle, data from them is worthless. Send one signal when the phone first starts moving after, oh, 2+ minutes of being at rest, and only alert if you see a large spike around a single cell tower within a five-second window. Data use would be very low.

Re: Power

[Anonymous Coward]

I can't understand how they deal with the information bottleneck. Having a million cellphones trying to communicate simultaneously must effectively have the same effect as a DDoS attack..

Re:Power

[rasmusbr]

Yeah, come to think of it, if you have a large enough installed base then you could just rely on the phones that are connected to a charger at any given time.

Re:Power

[langgp]

According to this http://phys.org/news/2016-02-myshake-app-smartphones-worldwide-seismic.html MyShake only briefly activates GPS after it detects an earthquake.

Re: Power

[Anonymous Coward]

Each phone on a network has a tiny continuous link just to keep the connection alive and manage signalling.

It's what sms travels over. Net 0 additional bandwidth.

Re:Power

[sumdumass]

You need a control which means lots of other phones. Otherwise you could get false positives.

Imagine this scenario. It's the night of the big game. Everyone stays up to watch it. You plug your phone into the charger and place it on the night stand next to the bed. You then start having sex and the motions cause enough vibration to trigger the app. Now imagine everyone doing the same. Right about the time you are about to finish you get a text saying we're all going to die.

Re:Power

[Anonymous Coward]

You can download software that uses your laptop accelerometer and BOINC to monitor them http://qcn.caltech.edu/sensor/

Re:Power

[rnturn]

"I think cell-tower positioning would generally be good enough. If you have thousands of users then the positioning errors will tend to cancel each other out. You probably don't need to update the position very often either, since most people spend most of their time not moving very fast or very far."

Seems like the place to put the accelerometer is on the cell tower. Power wouldn't be a problem on the tower. Plus, wouldn't putting it in my cellphone generate a slew of false alarms whenever I jog up the stairs to my office, run across the street, or any number of everyday activities that would cause my cellphone to get jostled?

Re:Power

[KGIII]

What I'm missing is how will they know if it's an earthquake or just a car driving by? How will they know if it's in a pocket or not the big one? I didn't read the article, I'm no heretic, but it looks like they just want to install this on the cell phone that you're using as opposed to an old, no longer used, cell that can be put into a stationary position.

Making use of old cell phones for something like that makes some sense to me. This? No... I'm obviously missing something because this makes absolutely no sense to me.

Re:Power

[rasmusbr]

I'm not an expert, but I imagine they would do at least these three things:

1. Correlate waveforms recorded by multiple phones in multiple places, i.e. look for a single fingerprint in the waveforms. If it's a bunch of different waveforms then it's probably not an earthquake.
2. For those waveforms that match the fingerprint: plot the velocity that the waves would have to be moving at in order to cause the waveforms in those particular places at those particular times and see if the speeds are consistent with the speed that earthquakes travel at and if the velocity vectors diverge from an epicentre. If the speeds are too low or too high or if the vectors do not diverge from an epicentre then it's not an earthquake,
3. Compare the waveforms with known waveforms from historical earthquakes, i.e. see if it looks anything like an earthquake.

In order to fool the system you would have to have multiple people in multiple places doing the same "silly walk" at just the right times.

Re:Power

[KGIII]

I kind of figured that? I am not an expert, so I don't know either. It's just that I don't see how they're going to get meaningful data (with enough "at rest" devices perhaps?) and separate the signal from the noise. It is a bit interesting but, at the same time, I've oft repeated that it's kind of important that we look to continued uses for retired smart phones - I think that might actually be a more viable method. Slap it on a shelf, connected to the wireless, and then use the one in your pocket as the device that gets the notifications? I'd articulate further but I suspect you'd be more fluent in that than I would.

I just don't see it as feasible, at any meaningful level, with the devices in pockets, in cars, with movement, etc... There's a whole lot of signal and noise issues. It might, and this is why I'm not an expert, be enough noise that it's easy to filter out (high signal to noise ratio) but that just seems improbable - without this system having years and years of massive amounts of real-world data, correlation, and some VERY big compute power behind it. (Distributed computing on the retired smart phones might be feasible.)

Again, I'm going to speculate that you'd be far more adept at refining the above than I would be. If not, I'm not the most articulate, I'll try to elaborate just so long as you don't hold me to very high standards.

Re:Power

[rasmusbr]

Yeah, the best way would definitely be to organise an effort to set up dedicated hardware.

Most developing countries are probably capable of doing that, especially now that you can pretty much get a used smartphone for free. Build a smartphone app, get some used smartphones from a recycling center, install the app on all of them, put one smartphone on a stable surface (such as a basement floor) in every government office around the country and make sure the office workers know how to maintain the system (keep the phone connected to the charger, to the wifi network, restart the app if it crashes, etc).

But the article that we're discussing is about a research project at a university, which means that part of their job is to look into other ways of doing things. Maybe the crowdsourcing app could supplement the government-organised detection system.

Re:Power

[KGIII]

Seeing as we're now off on a tangent, I've actually pondered how to make use of all the old cell phones. It seems like it might be fairly trivial to use them as distributed compute devices. I'm thinking a project (I was interested because of my son's project) that would use the devices to sequence genomes on endangered flora and fauna from the South American rain forest.

Given that they're already DC, it seems like it might even be reasonable to power it with solar. They're "disposable" (or at least treated like that) so they'd all be donated. They could scale out reasonably well. One could even make some sort of rack that had its own solar power on it, was weather resistant, and had batteries (even just deep cell marine batteries), and provision them and people could put them on their property and then feed the data back into the system. They could be housed in a fairly small footprint. They don't really generate a lot of heat - but that'd have to be looked into. When one breaks, we're throwing so many away (as a society) that it could just be sent out to recycling.

Hell, done well - there's the chance to make a small profit recovering the materials from within the phone itself. If the organization was set up as a non-profit then people could write their donations off. It'd be easy to just assign a few arbitrary values, they could include a SASE and the receipt could be mailed back to them. If it works and has X amount of resources it's worth $5. For Y it's worth $10. For X it's worth $20. Something along those lines.

I figure I might as well mention that again - I'd mentioned it in the past, probably a year and a half ago, and actually had a couple of people email me about the idea. it seems like it should be fairly trivial. It's actually quite a bit of compute power and it won't be fast but it'd be cheap. I was told that such could even be quite specifically targeted and optimized as the architecture is fairly similar across a bunch of them. I'd have to dig through my email archives to find out the various suggestions.

Hmm... I'm not really seeing any major downsides. The whole thing could be run as a non-profit and I'd not be surprised to find some university interest. I've still got some friends in academia. Maybe it's time to write something a bit more formal and approach them. I'm not exactly being all that productive, I don't need money, and it might turn into something of value?

Re:Power

[rasmusbr]

Sure, if you can find an efficient way to power them. I guess the chargers / AC/DC-converters that phones ship with are probably horribly inefficient.

Also, what happens when your cluster of phones all try to access the Internet at the same time?

The total world production of smartphones and tablets is now over 1 billion per year and many of those units have multi-core chips. The number of CPU cores that get discarded or recycled each year will surely top a billion within a couple of years. That is a significant resource if you can find a way to use it efficiently.

Re:Power

[KGIII]

I'm thinking power over DC - probably straight from photovoltaic. It'd be trivial (I think) to power them with batteries that are charged from solar power - they could even retain their internal batteries.

Only one, probably a hub, would connect to the net at a time - they'd communicate with each other on their own network. They'd not all be connected to the 'net at one time - they'd all communicate with each other and then, I'd guess through a wired or wireless hub/access point, communicate with the 'net as a whole. They're all in close proximity, perhaps bluetooth can be used to shunt data between them and then simply have a "master" phone which does the connecting to the hub after assembling the data.

And yes, that's what I was thinking. It's actually quite a lot of compute power that's just being chucked into the landfills. This might even convince more people to responsibly discard their old equipment. They'd then be centralized and would make recycling more efficient. They're low-power devices. I really think there's something to this but I lack the expertise to go it alone. I guess I could muddle my way through it and get it started? I'm not really that adept and my programming skills aren't the greatest but it doesn't seem like it'd be all that difficult. It seems like a good use for things - it doesn't even need to be for sequencing genomes. There's all sorts of things that can be done with them.

I've used an old phone to play around with before. I've even run some NAS off of a phone, used one as a wireless hub that was more an access point, and have even had an HTTP server loaded on one at one time. I never got around to getting a database and PHP up and running on it but I suspect that would have worked.

So like Amber alert - everyone goona disable

[Anonymous Coward]

Except Amber parents.

Re:So like Amber alert - everyone goona disable

[sunderland56]

Hey, the Amber alert system is really useful. If you kidnap a kid, you know you're golden until you see the alerts.

Mexico City: Got it for quite some time already

[gwolf]

Most of the public buildings in Mexico have had "alerta sísmica" for well over five years already (don't remember how long). Last year, a large-scale rollout was made so there are audible alert boxes all over the city. Of course, we are lucky to be ~500Km away from most of the tectonically active locations in our region, while unlucky enough to be sitting atop one of the most "interesting" soils for seismic activity amplification... So it's a very unique combination!

Re:Mexico City: Got it for quite some time already

[gwolf]

FWIW, forgot to mention, the typical reaction time we have is around one minute. Usually too little to have everybody evacuated from premises... But we have history on this regard, and it has proven a very important development.

Re:Mexico City: Got it for quite some time already

[Anonymous Coward]

Have you experienced it in action ?

Re:Mexico City: Got it for quite some time already

[Anonymous Coward]

Usually too little to have everybody evacuated from premises

That may be a good thing. Many people are killed by falling debris when they flee buildings during an earthquake. The better strategy is to find shelter under furniture where you will be protected from falling objects (including, in some cases, the floors above you collapsing).

Re:Mexico City: Got it for quite some time already

[tlhIngan]

FWIW, forgot to mention, the typical reaction time we have is around one minute. Usually too little to have everybody evacuated from premises... But we have history on this regard, and it has proven a very important development.

One minute is not a lot of time, but if you get a warning, you can do a lot of things in that minute that will save lives and reduce the disaster.

For example - gas pipelines can be depressurized and isolation valves closed, so any fires that start from the gas is limited to a very small amount of gas.

Elevators can be set to stop at the next available floor so passengers can get out. (one minute is more than enough for an elevator to go to a floor and evacuate its passengers).

Electricity can be shut off as well to limit the amount of fires that get sparked. Emergency power off functions may be activated in data centers to save critical data and perform quick power off to reduce fire risk.

Trains and public transit can be slowed down or stopped - depending on the system, it may not be possible to evacuate the trains in time, but once stopped, there's a much lower chance of derailing

And humans can go and hide in the strongest part of their building for safety.

One minute is not a lot of time, but a lot can be done to minimize the amount of risk and even reduce post-earthquake fires.

I say we combine the two

[Krishnoid]

Let's strip a smartphone down and attach it to the animals that go squirrely before an earthquake. Ground motion detection enhanced with animal senses -- bears further investigation, I bet.

Re:I say we combine the two

[Anonymous Coward]

animals that go squirrely

bears further investigation

I'll have the duck and cover please.

xkcd

[plaut]

obligatory xkcd reference: https://xkcd.com/723/

Re:xkcd

[kaur]

This one also applies:
https://xkcd.com/937/

Re:xkcd

[Opportunist]

I was honestly pondering writing an app like this. A few RAD tools I know have an almost complete wizard-made solution for it, just add some flashy graphics and maybe a configuration gimmick.

Re:xkcd

[myid]

Re. the cartoon - I guess that could happen, if you typed fast enough. One time at work I was standing next to my boss A, as he was talking to B on the phone. During the phone call, a quake whose epicenter was closer to B struck. B got all excited, saying "Hey, we're having an earthquake!" Then the shaking hit where A and I were. So A found out about the quake before he felt it.

Re:xkcd

[PPH]

USGS already does this. They monitor tweets for the word 'earthquake' in a number of languages. And they got a few seconds advance notice on the 2015 Nepal earthquake before their seismographs picked it up.

Re:xkcd

[angel'o'sphere]

Yeah, perhaps we indeed live in a morphogenetic connected world.

I was thinking about an Tsunami App yesterday. As I'm in an area that was hit by a Tsunami 2004 ... it would need to rely on SMS so, as I for my part don't have internet in this area of the world (via the phone service).

This is nothing new

[stephanruby]

This is nothing new. There apps that already do this. For instance, this one does it and it's not called MyShake. And I bet there are many more apps that do exactly the same thing if you just look for them.

Re:This is nothing new

[NotInHere]

Earth quake warnings in fact were a thing even before the concept of "apps" were invented (yes, probably the apper troll considers this luddite now): https://en.wikipedia.org/wiki/...

So yeah this now gets "upgraded" to the smartphone age: if you don't install this app and register with your email, and let the app monitor your activity, and whatever it does, you won't get an earthquake warning in advance and are probably more likely to die. Yeah, quite cool app.

Of course, several times

[gwolf]

Every 3 or 4 months

A few seconds?

[PopeRatzo]

I do my best to avoid places with earthquakes, but if I ever go to an earthquake zone, what the fuck am I gonna do with those "few seconds" warning that this app would provide me? Unless I can somehow use those seconds to board a hot air balloon, what can I possibly do with a few seconds if the earth is about to open up, besides take the opportunity to plant a big wet goodbye kiss on my own ass?

I admit that I've been fortifying myself with spirits in order to work up the courage to go out into the cold to walk the dog, so maybe I'm missing something.

Re:A few seconds?

[Quinn_Inuit]

Well, some people are only a few seconds from a stairwell or could at least get under a sturdy desk.

Re:A few seconds?

[Known Nutter]

what the fuck am I gonna do with those "few seconds" warning

We get this question quite a bit in the Bay Area. Here are the common responses:

Public: Citizens, including schoolchildren, drop, cover, and hold on; turn off stoves, safely stop vehicles.

Businesses: Personnel move to safe locations, automated systems ensure elevators doors open, production lines are shut down, sensitive equipment is placed in a safe mode.

Medical services: Surgeons, dentists, and others stop delicate procedures.

Emergency responders: Open firehouse doors, personnel prepare and prioritize response decisions.

Power infrastructure: Protect power stations and grid facilities from strong shaking.

For more: http://earthquake.usgs.gov/res...

Re:A few seconds?

[angel'o'sphere]

If you are so scared about earth quakes, why are you not even knowing the basic measures you can take?

http://www.shakeout.org/dropco...

Re:A few seconds?

[Alumoi]

Update your failbook page? Tweet you're about to die? Upload your last selfie?

Re:A few seconds?

[PopeRatzo]

If you are so scared about earth quakes, why are you not even knowing the basic measures you can take?

I'm not afraid of earthquakes, because I don't frequent the hell-holes that have earthquakes.

Re:A few seconds?

[angel'o'sphere]

You can have an earthquake on every single point on the earth.

Re:A few seconds?

[PopeRatzo]

You can have an earthquake on every single point on the earth.

Thanks a lot. Way to spoil my weekend.

Re:A few seconds?

[angel'o'sphere]

Well, the simple things are standing in a door frame or hiding below a table.
Good Luck :)

Re:A few seconds?

[PopeRatzo]

Well, the simple things are standing in a door frame or hiding below a table.

Can I just build a giant table and door frame over my house?

Re:A few seconds?

[angel'o'sphere]

Haha :D

Ofc you could, rofl. But: would it help if the house below it shakes?

LOL

[Archfeld]

I'd like my smartphone shaken, not stirred :)

Japan has a sensor network

[oheso]

We get alerts on our cell phones with at most a couple of seconds' warning. In theory that gives you time to duck under a desk or into a doorway. In practice it's always been: Beep! What the fuck is that? Earthquake!

Re:Japan has a sensor network

[khchung]

In practice it's always been: Beep! What the fuck is that? Earthquake!

Well, at least that would help you know right away it *is* an earthquake, and not your building collapsing or you got dizzy, and allows you to react correctly.

Re:Japan has a sensor network

[Anonymous Coward]

In practice it's always been: Beep! What the fuck is that? Earthquake!

Well, at least that would help you know right away it *is* an earthquake, and not your building collapsing or you got dizzy, and allows you to react correctly.

Or Godzilla.

Re:Japan has a sensor network

[angel'o'sphere]

Actually either the App had its unique sound or if it was SMS based you would assign a unique ring tone to the sending number.

App requests way too many permissions

[Quinn_Inuit]

My wife was at the conference where this was unveiled and she came home excited about it. When we went to install it on our phones, though, we discovered it requested a surprising array of permissions that you wouldn't think it would need, like information about your contacts list. I think I'll hold off till they scale that back to something more reasonable.

Re:App requests way too many permissions

[drinkypoo]

If you have a rooted Android phone you can use Xposed Framework to get AppOps Xposed, which brings back the AppOps feature that lets you control atomic permissions on Android. Then you can deny all the privacy-stealing permissions for an app and still use it.

Re:App requests way too many permissions

[Quinn_Inuit]

Good idea. I keep forgetting that I can get Xposed for Lollipop now.

Re: Mexico City: Got it for quite some time alread

[Anonymous Coward]

I can't remember how many times my cube-farm walls have caught and held-up the few floors above me that were collapsing in an earthquake. Those things are a godsend.

Ummm, Welcome to Japan?

[Anonymous Coward]

Welcome to the 20th century everyone.

Flash Mob

[PPH]

Tomorrow at noon, everyone jump up and down.

Re:Unless phone time is set to 1 Jan 1970

[sunderland56]

Hey, even a bricked Apple phone can tell you when there's an earthquake. Just leave it on the edge of any convenient table.

News at 11

[invictusvoyd]

A porngraphy website whose "myShake" and "Ishake" (on ios) apps are quite popular , has filed copywrite infringement suite on a recent seismography app .

Re:News at 11

[Anonymous Coward]

Maybe phones with running porn apps aren't the ideal data providers for seismology. You know, false positives.