It's a tough problem to solve because you're up against the laws of physics and the very boring (and often counterintuitive) "Antenna Theory". Bluetooth is in the UHF band, and UHF isn't good for penetrating anything let a lone concrete rubble.
To penetrate rubble effectively you really want to be in the ELF-VLF bands, (That's what submarines/mining bots/underground seismic sensors use to get signals out).
Obviously that's ridiculous. Everything from ELF to even HF is impossible to use in a "under the rubble" situation because of physics[1]. Bluetooth (UHF) might be "better than nothing" but you're losing at least 25-30 dBs (which is like 99.99% signal) in 12 inches of concrete rubble. VHF (like a handheld radio) can buy you another 5 inches.
Honestly I think sound waves travel further in such medium than RF waves.
[1]: Your "standard reference dipole" antenna needs to be 1/2 or 1/4 your wave length to resonate. At ELF-VLF range you need an antenna that's 10k-1k feet long. You can play with inductors and loops to electrically lengthen your antenna without physically lengthening it, but you're not gonna get that below 500-200 feet. The length of a submarine is an important design consideration when deciding on what type of radio signal it needs to be able to receive/transmit vs how deep it needs to be for stealth.
That means LoRa and cellular network won't help either... That's good to know.
Yeah, sound seems like a good option. There's two points I'm struggling with though: 1. Disasters often are accompanied with a lot of noise which I'm afraid would drown or interfere with the other sound waves. 2. Mobile devices might not be able to detect the sound waves that are broadcasted.
I'll need to do more homework about those.
We'll have more flexibility with recovery for sure, so maybe that can address the shortcomings of the sound signaling...?
It still has to be cheap-ish though. Because first responders already have MUCH better tools [1].
Again, thank you so much for this comment. If you have any other suggestions or feedback, please don't hesitate to open an issue on github. Your input is awesome.
[1]: https://www.dhs.gov/archive/detecting-heartbeats-rubble-dhs-...
It does. If anyone is buried under the rubble "why don't they just tweet/SMS/mesh-network they are stuck" might be the first question someone would have. A simple answer is that the nearest GSM tower is also under rubble. But that's not enough. Even if the nearest GSM tower is 100% fine (or their next LoRa mesh peer is also fine), you need to be less than 12-15 inches of concrete (in the most optimal/reflective way) to have any signal at all. At the end of the day you're up against physics. 3 questions physics mandates:
1. What is the wavelength you're transmitting?
2. How far is your receiver?
3. What is your medium?
The keyword "Antenna Theory" or "Antenna Design" is a good rabbit hole for this sort of thing. There is this funny saying among radio enthusiasts "everything is an antenna if you squint hard enough"[1].
There is also something that cooks call "The Human Element"[2] when they're trying to evaluate the appeal of some foods to people. The "Human Element" in an "under the rubble" situation is that you want something that a normal human being would have on them when the sudden situation of an earthquake happens and they find themselves under a rubble.
I don't know what that is. I live in a place that is pretty far removed from earthquakes as a major problem. Geologically speaking, we get a "major earthquake" every 500 years or so. However, that's an average and it's been 600 years for us but the max has been 3,000 years and the min has been 200 and the mean means nothing to the individual. However, a device of some sort that can emit low frequency, low bandwidth[3], long range message seems like a useful thing. I don't know if that would be the type of device every human would have "on their body" all the time. But maybe??
[1] Everything that's electrically conductive is an RF antenna. If you get buried in the right place next to a 100 feet long rebar, you could (theoretically) send a signal that anyone in 100 mile range could hear provided you had the right equipment of course.
[2] In cooking, it is this elusive thing that explains how some cultures/people love certain tastes or textures that other cultures don't. Slimy foods in Middle Eastern or Japanese/Asian cultures vs Western Pallet for example.
[3] A device that needs to send "S.O.S" needs maybe 1byte/second or even less. You can't communicate any meaningful information in a seismic-pulse-signal type device, but "SOS" is certainly doable.
Yeah the human element is the strictest here. From the little research/thinking I did, BLE felt like the most "commonly available" tool we have. But I'm wondering if manufacturing a cheap wearable could also be an option.
I definitely need to do a longer/deeper dive for this project. I overlooked it in the past, thinking people don't need it. But it seems like I just dropped the ball in distribution previously.
So yeah, I imagine concrete is going to be even worse than a shrubbery, that said, might get some interesting propagation paths if there's metal surfaces in the rubble to reflect waves.
And I can only eirp 1000mW. I mean, legally. I can't eirp anywhere near that can you imagine the losses?
It looks that for 600m you could use a vertical wire with a lot of radials :) Maybe with a balloon?
But 2200 would be fun indeed.
The problem with bass sounds is ostensibly they're nondirectional. So the ideal sound for a phone would probably be the loudest pink noise they can make without blowing the speaker out.
What about Rydberg sensors for VLF earth penetrating imaging, at least?
From "3D scene reconstruction in adverse weather conditions via Gaussian splatting" https://news.ycombinator.com/item?id=42900053 :
> Is it possible to see microwave ovens on the ground with a Rydberg antenna array for in-cockpit drone Remote ID signal location?
With a frequency range from below 1 KHz to multiple THz fwiu, Rydberg antennae can receive VLF but IDK about ELF.
IIRC there's actually also a way to transmit with or like Rydberg antennae now; probably with VLF if that's best for the application and there's not already enough backscatter to e.g. infer occlusion with? https://www.google.com/search?q=transmit+with+Rydberg+antenn....
I imagine such fancier tools would be less available among common folks, and more among first responders.
NASA already has the tech to detect heartbeats under rubble using radar [1]. No additional equipment is needed by the rescued. The problem is emergency response can get overwhelmed in large disasters.
If Rydberg sensors would be more common, and new tech is added to mobile devices, this can seriously shift the playing field.
I will look into this, because we need out of the box solutions. Thank you!
[1]: https://www.dhs.gov/archive/detecting-heartbeats-rubble-dhs-...
FWIU there are contractor-grade imaging capabilities and there are military-grade see through walls and earth capabilities that law enforcement also have but there challenges with due process.
At the right time of day, with the right ambient temperature, it's possible to see the studs in the walls with consumer IR but only at a distance.
Also, FWIU it's possible to find plastic in the ground - i.e. banned mines - with thermal imaging at certain times of day.
Could there be a transceiver on a post in the truck at the road, with other flying drones to measure backscatter and/or transceiver emissions?
Hopefully NASA or another solvent company builds a product out of their FINDER research (from JPL).
How many of such heartbeat detection devices were ever manufactured? Did they ever add a better directional mic; like one that can read heartbeats from hundreds of meters awat? Is it mountable on a motorized tripod?
It sounds like there is a signal separation challenge that applied onboard AI could help with.
I just went through Helene in Asheville last year and it was painfully obvious that our cloud overlords have overlooked the offline disaster use case. Basically, when you’re in a situation where you desperately need technology to help, you’re on your own. I was imagining that tools like this would be great, but without the cloud, I was helpless.
Maybe instead of trying to get users to install this, it could be a proof of concept to show what’s possible, and to say to Apple and Google: install this basic lifesaving tool on every phone by default?
I'll work towards that. I'll need to figure out how to prove the poc is useful.
Apple's "Connect to a Satellite" touches on that, but it's not available for most of the world or on old devices. :|
Admittedly doesn’t work under rubble, but should work in the usual/general case.
I am saying this because I think your target market may not be people stuck under rubble, but large scale industries, construction site workers, miners, firefighters, who can install the app beforehand. Cheers.
It's a local version of https://www.findmespot.com/
Sadly today, money and connections are needed for r&d and distribution. And having a market that buys a tool like this in certain circumstances can seriously contribute to finding/building publicly available solutions.
Heavy duty wearable tags and a dense network of wall-mounted relays pinging them and triangulating 24/7 can be the only solution. Sure you can use BLE beacons for that, and this is being solved with multiple competing products as we speak (look up "locating objects in warehouse"), but specifically for the human safety applications, they would have to be imposed by the regulators. Workforce is seen as disposable and nothing short of laws and regulations is going to force employers to improve safety.
In Australia, open-cut mining dominates coal production, accounting for over 80% of the total, while underground mining's share has decreased to around 15%. Open-cut mining is generally more efficient and cost-effective, allowing for higher recovery rates and larger-scale operations.
It has been estimated that more than two-thirds of the world’s yearly mineral production is extracted by surface mining. There are several types of surface mining, but the three most common are open-pit mining, strip mining, and quarrying. These differ from one another in the mine geometries created, the techniques used, and the minerals produced.
Like anybody else in the modern world, it's hard to pry smart phones away from mine workers ...
Miners warned to limit smartphones in the pit (2013) - https://www.itnews.com.au/news/miners-warned-to-limit-smartp...
---
Fldigi is a free and open-source program which allows an ordinary computer's sound card to be used as a simple two-way data modem. The software is mostly used by amateur radio operators who connect the microphone and headphone connections of an amateur radio SSB or FM transceiver to the computer's headphone and microphone connections, respectively. This interconnection creates a "sound card defined radio" whose available bandwidth is limited by the sound card's sample rate and the external radio's bandwidth. Such communications are normally done on the shortwave amateur radio bands in modes such as PSK31, MFSK, RTTY, Olivia, and CW. Increasingly, the software is also being used for data on VHF and UHF frequencies using faster modes such as 8-PSK. Using this software, it is possible for amateur radio operators to communicate worldwide while using only a few watts of RF power.
It's all real slick, but apparently WSJT-X software (made by an academic, not meant as a slur) is slicker, and nearly everyone uses that now. It isn't designed for "rag-chewing", it's designed for propagation testing and "contesting"; it's real dull after you've made dozens of contacts really far away.
Lately when i actually transmit with fldigi, i will use a web SDR like kiwi to "listen" to myself. Since the radio has a soundcard, and my PC has a soundcard, i can route websdr to one "modem" and my radio to another, and decode my own transmissions.
that's also boring!
Ideally, there would be an option to have your phone "listen" in the background and notify you if anyone needs help. I haven't added that yet.
For now: let's say I'm in a neighborhood. An earthquake happens, a building collapses. I have the app installed and I'm fine. So, I open it and I get close to each building and check if there are any survivors under it.
You'll keep broadcasting, until your battery runs out. It's not heavy on the battery, so it can broadcast for a while.
I'll clarify that in the readme.
The more important problem right now is signaling. Recovery is a much more flexible problem to solve.
Something like that could be useful if anyone with an iphone going nearby the signal was enough to pick it up, without them requiring to have the app installed.
TLDR; Apple devices scanning in the background require a lot of workarounds [1].
Another comment suggested proposing this to Apple [2]. If this were built by Apple internally, it would be game-changing. A little delusion never hurt anyone :seenoevil:
[1]: http://www.davidgyoungtech.com/2020/05/07/hacking-the-overfl...
[1] I used to be terrified of the concept of an Avalanche. However, at the end of the day you're only exposed to an Avalanche if you choose to be or live next to a large mountain in the winter. It's not even like Hawaiians living next to an active volcano. All you have to do (in terms of living) is to move away during Avalanche season. Hawaiians atleast have the excuse of "it hasn't erupted in 10 years"-maybe. THe other example is if you're a skiing/winter-sport enthusiast (like me). And you simply balance your chance of an avalanche vs the reports and "how much fun it sounds to ski with your friends"
[2] If you find yourself in an avalanche situation, you need/must do your best to "swim" to the top of the avalanche. There is gear that can help you do so outside of beacons or "signal based" notifications. Once you've established all those unfeasible, you're left with the same tools a "sudden earthquake victim" is up against. You still have the upper hand because a simple handheld transceiver (from like Icom, Kenwood or Yaesu) will have 100x the range through snow vs one through liquid water or concrete.
[3] Physically speaking: ice is eventually going to behave like water under enough depth. I think.
[4] Avalanche Gear example: https://alpenglowsports.com/collections/avalanche-safety
I know that only because recently I have been getting a notification on my Pixel 9 Pro that Satellite based emergency services are going in and out of coverage for some reason. I never even knew it existed until my phone told me it was down lol.
At this moment they only cover some countries, but it might be well worth focusing on building out this tech instead of trying to make something new.
It's hard to find a one-size-fits-all solution. I don't know if satellite coverage would still be available under rubble. It's definitely valuable in other situations.
The country coverage is a huge part. My country will likely not get satellite coverage in the next 5 to 10 years. Such countries is where disaster response is worse and infrastructure is more prone to failure.
This is a bit more extreme, but some type of triangulation from multiple rescuers could be useful in closing in on a spot.
This is really interesting and thanks for sharing.
For the specific case of burial under rubble it might be better to work on sensing sounds rather than using radio broadcasts. If the person can move they may be able to tap on concrete with another piece of concrete. Using three or more phones (plus BLE or WiFi for timing coordination) placed on the rubble it might be possible to triangulate the location of the tapping. While there are professional (expensive) versions of this (like the heartbeat detector mentioned above), giving the capability to everyone could be useful. Adding the BLE emission/detection couldn't hurt and when the trapped person runs the app it could give them instructions to maximize their chances of detection. For tapping those instructions might be to tap three times once every five minutes to minimize the physical energy required. That would mean the three sensing phones would need to wait 5 to 10 to 15 minutes to do detection and acquire a location. Tapping on stones could work by itself since people can put their ears to the rubble and listen for the tapping, so the main benefit of the app might be just to tell the person what they should do in their situation (tell the trapped person to tap 3 times every five minutes, tell the rescuer to listen for 3 taps occurring every five minutes, providing a timer to indicate the 5 minute intervals could be useful too). Five minutes might not be optimal, some research would be needed on that, and it may depend on the energy level of the trapped person so perhaps the sensing cell phones should be left in place for hours or days. Ideally all phones would have this as an emergency app that would provide advice and help with BLE beacons, timers, cellular calls, sensing, and whatever in any situation.
In general this seems like a difficult problem and worthy of some extended research.
I find the "instruction app" a little limited because a person needs to be conscious, they need to be calm and collected to remember to open the app, and their phone needs to be accessible.
In its current state, providing instructions is better than just broadcasting BLE which will be weakened and lost by rubble.
In my country, a lot of first responders still rely on asking people to scream if they can hear them. Anyone unconscious would be discovered days later...
There has to be some way of providing an automated signal of life.
The comments about the physics limitations are pretty awesome. They make me wonder if we can assist it in any way... I learned about passive resonators recently. I know they have a short distance transmission, but maybe if they were part of the building material?
I need to start by learning more about the physics. I know I'm limited by it, but maybe there's window we can walk through instead of the door.
Suggestion: You can possibly try if you can use OpenHaystack [1] to send some "unknown tracker found nearby" alert to any nearby iPhone even if the receiver device doesn't have your app installed or use the FindMy network to send arbitrary data for your app's communication. [2]
In my situation, GPS was being jammed. That made GPS not useful at all. I tried my best to avoid it for that and battery usage.
But this initial reading and turning it off is just genius. I added to the list. Thank you so much!
WiFi doesn't fare any better than BLE under rubble unfortunately, but WiFi is going to be a useful technology to build disaster response tools in general
I'm not a fan of how Apple managed to make background BLE advertising and scanning work only with its own devices.
Can voice be transferred? Message be recorded and rebroadcasted?
What about making the device vibrate SOS pulses. The person at stake might not have the strength to tap or bang, may be able put the phone on a metallic surface and send vibrations?
Those two questions are spot on.
Voice cannot be transferred yet. Re-broadcasting might require a bluetooth meshnet.
Vibrations are a great idea. They are not that battery intensive. They also can be detected by 3rd party tech.
Both were added to the list of ideas. Thank you!!
I haven't done any effort there, because I want to get this into hands.
Improvements are definitely needed with location, but the signal loss as you mentioned is higher priority and restricts what can be used for location.
I'm planning to do my homework about alternatives to help mitigate the signal loss. And from there, the options for location might be clearer.
Some of them might require you open the application the first time ever while on internet.
There are many out there (on the App/Play store) that support both use cases you mentioned, I just can't list any off the top of my head because it's been a while since I used one.
I don't know the amount of battery that is necessary to do that.
To cut time, I didn't modify the volume programmatically. So the user would need to have their volume set at 100%.
This needs improvement, especially considering my phone is always muted.
Likewise, the battery usage needs better testing.
Ideal conditions is likely open spaces with no interference. Walls and concrete lower that as you mentioned.
I'm hoping it can help if people are standing right next to the rubble, whether BLE or the "siren" sound.
I have no guarantees though. I haven't properly tested because I haven't thought of a low cost option yet.
I don't know if phones support LE Long Range. I will look into it [1]. Thank you!!
It of course depends on the disaster, but in case of war this app can be used to find survivors for purposes other than saving them.
Great idea though.
It'll be easy to find a collapsed building site. The person rescuing just needs to be close to the site with the app. Otherwise, the signal might not be noticed.
+1 regarding finding survivors for malicious purposes. I tried my best to lower the surface area of that happening. It's still possible in edge cases, like opposing soldiers finding the people transmitting SOS, but they need to be extremely close to do that.
If the person doesn't respond to an "Are you conscious?" alert, maybe allow responders to "toggle" certain actions like the flashlight you mentioned.
A meshnet for communicating back-and-forth between devices. Permissions might be tricky though.
Cross-platform bluetooth between Android/iOS is too annoying sadly. If LoRa can help address those issues, it would simplify things a lot.
Although it has the obvious limitation that you would have to be within arm's reach of it before you were to be trapped in a collapsing building, and sound can travel sufficiently under rumble.
> {Code-and-Response, Call-for-Code}/DroneAid: https://github.com/Code-and-Response/DroneAid
> "DroneAid: A Symbol Language and ML model for indicating needs to drones, planes" (2010) https://github.com/Code-and-Response/DroneAid
> All but one of the DroneAid Symbol Language Symbols are drawn within upward pointing triangles.
> Is there a simpler set of QR codes for the ground that could be made with sticks or rocks or things the wind won't bend?
I would change your license so it can’t be stolen by Apple. At the very least they need to buy it or hire you as employee.
I will also try to create a monetized version for specific use. Maybe that can help fund building a solution that's publicly available.
Both of those will increase the chances of it being adopted by smartphones, like Google and Apple. That's why I switched from GPL-3.0 to MIT.
I appreciate the kind words and the awesome suggestion. Thank you!
Would be nice if the readme included the current method to detect disaster and the nature of the "SOS" signal. Is that something Bluetooth has a behaved protocol for, or is it really just chirping?
+1, I usually have it off, but toggle it on when I'm in a bad setting. It can be made more accurate with time, especially with onboard ML models.
It is a Bluetooth protocol (BLE Scanning and Advertising, if you'd like to learn more about it).
I'll be sure to update the readme. Thank you for the great questions!!
Yeah, false positives are the largest problem. The barometer sensor helps reduce them, but it's not available on all phones.
Machine learning can help a lot here, but I haven't looked for publicly available data yet.
Well thought out. Excellent work. Thank you for making this available as open source!
Side note -- I wonder if it's possible to have a companion system to scan regularly for the signal at home for the elderly. Then notify caregivers from the companion using available wifi/cell if the SOS is detected. Since it's so well behaved it seems like even having a regular wellness check mode would be possible. Having a companion system for relay would allow the app itself to never need WiFi/cell and have independent care notifications protocols that wouldn't need to be maintained in the app.
That's a really cool idea. Kudos!! A companion system should be relatively easy. It's also two single devices communicating with each other which makes things much simpler to build as well. This is worth looking closer into.