←back to thread

158 points nizarmah | 1 comments | | HN request time: 0.208s | source

A couple of months ago, I built this app to help identify people stuck under rubble.

First responders have awesome tools. But in tough situations, even common folks need to help.

After what happened in Myanmar, we need something like this that works properly.

It has only been tested in controlled environments. It can also be improved; I know BLE is not _that_ effective under rubble.

If you have any feedback or can contribute, don't hold back.

Show context
daveguy ◴[] No.43552300[source]
Okay, what is "disaster detection"? I'd really prefer my phone not chirp noise over Bluetooth channels if I don't notice and respond to a notification in time.

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?

replies(1): >>43552356 #
nizarmah ◴[] No.43552356[source]
1. Disaster detection is a user toggle-able option that uses motion sensors to detect extreme shifts. 2. SOS signal is a BLE broadcast along with an alarm-like audio (generated from a sine wave).

+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!!

replies(2): >>43552405 #>>43557853 #
dylan604 ◴[] No.43552405[source]
It would be interesting to see if some of the earthquake advanced warning notifications could be integrated too. at least then it would be an "official" warning. seems like false positives would be possible with the device's motion sensors even with an "i'm okay" type of cancel.
replies(1): >>43552417 #
1. nizarmah ◴[] No.43552417[source]
I LOVE this. Thank you! Huge +1, even though it's not going to be easy.

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.