I'm one of the leads on this project! We've been flying under the radar for a while now, since everyone working on this has a "day job" (mostly ATLAS/CMS physicists), and we have essentially no funding. But we are making progress and hope to eventually do a full-scale release and produce some science! I'll try to answer any questions in this thread.
Speaking of radar, many phones have FM radio receivers, which could conceivably look for radio signatures of cosmic rays. No idea to what extent the hardware is made available to software, but I just found another reason to procrastinate.
Also kudos on actually going through and doing something with this; I remember talking about using cell phone ccd's as detectors a few years ago with other grad students, but of course it never got past the chit-chat stage (I remember wanting a dosimeter in my pants to ring at me if I got too close to one of my radioactive sources).
People have approached us with this idea! I don't think we have much/any access at the API-level to radio signals. However are talking with another group which is looking to build an inexpensive RF array about possibly integrating the two networks. There is also another group which is creating a dedicated radio device that plugs into the headphone jack of your smartphone, but in this case you need a special piece of hardware and the phone is basically just acting as readout/internet access. It's a cool idea but we're trying to show that it's possible to do science with the existing hardware that people already own!
Ha, we had this idea quite a long time ago (2008) when "apps" really started becoming a thing.
I think this works if you were in a densely populated area and a significant fraction of the population was listening, but that was always a pipe dream.
On a related note, when a 10^19 or above event hit a densely populated area, there's a decent sized ionization column near the center of the shower that could likely lead to cancer. Theres probably 100 to 1000 or so events like that in an area the size of the Bay Area.
Edit: I took out a part talking about noise after I glanced at the paper, but I would still be worried about it because the area of the detector (aka camera sensor) is roughly equivalent to the area of the electronics (ADC, CPU, Memory, etc...) in a smart phone, whereas most surface detectors have a huge scintillation footprint compared to the electronics footprint.
Actually the active area of the CMOS sensor is really quite small. Like ~0.1cm^2. It is the main reason we need so darn many phones (and, as you correctly point out, an appropriate density of phones) in order to have sensitivity.
The reason it's still interesting is that at the very very highest UHECR energies, like above 10^21 eV, the density of ionizing particles is so tremendous that it's almost impossible for a (tiny) camera not to see it, if it's close to the core. This is great because unlike the Pierre Auger observatory (which is already about 100% effective at detecting showers above 10^18 eV), we become better and better at detecting events the higher the energy. So while P-A has pretty much constrained the UHE spectrum up to about 10^20 eV, the limitations on its size (as in, how much area it covers on the ground) means it can't do any better, since the higher-energy events become so rare that the probability of them happening over their patch in the desert is quite small.
The hope with CRAYFIS is that, given global coverage, we can be the most sensitive measurement of these ultra-rare events at energies above what P-A has been able to observe.
On a related note, when a 10^19 or
above event hits a densely populated
area, there's a decent sized ionization
column near the center of the shower
that could likely lead to cancer.
Theres probably 100 to 1000 or so
events like that in an area the size
of the Bay Area.
Ha! Oh man. That'd be the best alert a smart phone could ever show me.
*beep*
You've just been irradiated by powerful
cosmic rays from outer space. Your risk
factor for certain forms of cancer has
increased by 0.013%.
I really would want to know if I was suddenly showered by cosmic rays, and my location within the crowd-sourced area-of-effect, but I wonder if an app could transmit it's sensor noise to a central server, and receive accurate results aggregated from all other app instances very quickly. I figure there'd be a lag of maybe a couple of hours or more.
Reminds me of this other site, which provides air quality stats:
So, to answer my own question, based on the graph of the rates at the wikipedia article[0] it looks as though you might see 100 to 1000 events per year, over an area the size of the Bay Area.
Yep, sorry for not responding, but about 100 per year. IIRC, 10^19 events have an electronic ionization column of about 6 feet wide close to sea level, which would be could be significant. I've been out of this area of physics for almost 7 years so my memory may be a little fuzzy.
Is this something that reasonable people should try to protect themselves from? I spend most of my time indoors. Will my house's roof protect me, or is this the kind of ray that penetrates everything?
Is there some simple insulation one could install to mitigate the risk? Specifically around one's sleeping quarters, where disproportionate time is spent.
Do cosmic rays disproportionately come from one direction? I realized that I'm somehow thinking of them coming down from straight up. However, they could come in at any angle I suppose, or even upwards through the earth if they're capable of passing through it. That being said, if atmosphere or matter degrade or deter cosmic rays, then I imagine they probabilistically come from straight up, since that requires passing through the least atmosphere.
Not much you can do to shield these events, it's mostly highly-penetrating gamma rays and muons. Maybe spend a lot of time a few dozen meters underground or the the basement of high-rise buildings. If it helps, you basically have to be within a meter of these things to really experience significant radiation (rough estimate; I don't know what dose actually corresponds to cancer). At a rate of 1/km^2/yr, that means the odds of one of these hitting within 1m of you in a year is 3e-6.
In other words, you can expect to get hit once about every 320,000 years. So I wouldn't worry about it personally. However it's interesting to note that the same logic implies that about 22,000 people are affected globally each year.
As for directionality, they can come from pretty much any direction, but as the angle deviates from zenith (straight up), the particles have to travel through more atmosphere, so the intensity is decreased. Neutrinos can travel upwards through the earth, but you don't have to worry about those causing health problems (for the same reason they are able to pass right through the earth).
yeah I was referring to the specific case where the observer is roughly close to x_max for a 10^19 or above event and quite literally in the ionization column. You can get rid of almost all of the electronic and gamma components in with about 10 feet sand, 5 feet concrete or 1 foot of lead, but that of course doesn't stop the muonic component.
(I worked on HiRes/Telescope Array for 4 years on muon detectors and bi-static radar techniques for UHECRs, but that was quite a while ago and some of my memory might be slightly off)
Not sure how much energy would be deposited, but I'm sure it's not insignificant. It's rare enough that no one person could expect it to happen but probably frequent enough that it does happen at least once a decade in a densely populated area.
We read about them! Very cool project, especially from high-school students. A lot of people got onto this idea about the same time. See also CellRAD, which is an app that is supposed to act as a dosimiter to detect when you are near radioactive sources. It's been known for quite a while that silicon pixels can detect ionizing radiation (in fact, that's why we use pixel detectors in the inner tracker layers of state-of-the-art particle detectors such as ATLAS and CMS). The problem with the pixels in the CMOS camera of your smartphone is that they are very small (in both pitch and depth) and were optimized to respond to optical photons, rather than charged particles.
We have played around a bit with implementing these radiation-sensing features, but the main goal for the CRAYFIS project (and which sets it apart from other similar projects) is to create a distributed network of phones that all work together to act as a giant telescope for UHECR events.
We've invited a small number of beta testers for the android platform, mostly to test that our website (which functions as our DAQ) is working and to get a baseline dataset to characterize the kind of challenges we'll have in processing the data.
AWS gave us a rather generous grant of credit for their services, and we spent a good amount of time developing a system for DAQ/live-metrics that should scale to at least a few hundred thousand users. The devices have to upload their data every ~3min, so it's a fairly significant amount of traffic. Beyond that we'd have to put either a lot more time into it or find "real" scalability experts, which means we'd need to get some $$$ first ;)
The main thing that has caused our beta rollout to stall is the fact that we have not yet been able to successfully calibrate our response to muons, which are an important component of the UHECR shower. Whereas we were able to calibrate the response to x-ray/gamma-ray photons fairly accurately using common radioactive sources, there is no easy way to generate muons under controlled circumstances (you basically need a particle accelerator). We are applying for beam time at Fermilab, so hopefully we will be able to do these measurements soon.
In parallel, we are also trying the "bottom up" approach, by developing detailed simulations to model how energetic particles interact with CMOS pixels. For this we're using GEANT, which does a great job of simulating how particle deposit energy into materials, but the hard part is modeling how that turns into actual charges that get read out by the CMOS electronics. We're getting close, but with nobody working full-time on this project and no funding, most development has happened in spurts and bursts whenever someone happens to have time away from their "day job".
In order to look for cosmic rays we have to scan HD video frames at ~30fps, which is a ton of data. So that means we have to very aggressively throw away "uninteresting" data before selecting candidates to upload to our servers (in physics jargon this is called triggering). Since the muons are not currently understood, we are afraid that we cannot design an effective trigger. Which means that if we continued the beta roll out, we might blow through all of our AWS cash and then realize retrospectively that all the data we have taken is useless. Hence we made the decision to pump the brakes on the beta testing until we can resolve the instrumental questions.
P.S.: If anyone really wants to give the app a try (unfortunately android only at the moment), feel free to shoot me an email directly. My email is my hn username at cern.ch.
Unfortunately not open sourced (yet). The main reason is because we're physicists, not web security experts (although I do have some background in this). I was afraid that if we open source the code and make the protocol way too obvious, the bar will be just too low for some troll to spam our data acquisition servers, making it hard for us to do science. So for now we're kind of relying on good-old "security through obscurity".
In the future we hope to harden the online system further, and eventually open source the code for both the android/iOS apps, as well as the online website which functions as a high-throughput distributed DAQ with online analytics.
Have you thought about trying to partner with someone to do a Google Summer of Code (or something similar) for this? It seems specific and targeted enough that it might have a chance of getting support.
That's a really cool idea, I don't think we've even considered this before. If you have information about how to get this kind of thing going, please get in touch! (My email is my hn username at cern.ch)
We're here! Unfortunately I'm terrible at social media and especially blogging, so from the outside we must seem pretty dead. See my response to GP, but basically we've pumped the brakes a bit after a small initial beta roll-out in order to ensure we are accurately modeling the detector response to muons. This is essential to ensure we're not throwing the baby out with the bathwater when deciding which data to keep (i.e. upload to the server). Since we scan HD video at ~30fps to look for the cosmic ray signatures, we have to very aggressively throw away uninteresting data in order to keep the data rates reasonable.
If anyone here is really interested in giving the beta test a go (android only for now), feel free to email me directly (look elsewhere in the thread for my address). It would be good to hear some feedback from the pro hackers here on hn :)
