Dead man's switch

Share your Projects!
21

If you are reading this message, that means that I am already dead.

** furiously tries to recall emails **

** curses wildly **

** pulls the plug on server erasing itself **

** reconsiders whether this was ever a good idea in the first place **

Anyway, um, I’m just suggesting people think twice, or perhaps three times, before implementing something like this. There are sooooo many ways it can go wrong.

See, e.g., https://arstechnica.com/civis/viewtopic.php?f=23&t=905876 and https://www.wired.com/2002/06/dead-men-tell-no-passwords

22

It’s great to see that this topic can spark some interest in use cases and imagination.

As stated before this, as many automations, is very reliant on the individual setup and use case. At first I had the same thought as you @hazymat, sending out an informative email. I quickly decided against this practice since I don’t believe that the ppl close to me checks their email that often. As I live in a apartment complex making a lot of noise seemed to be the better option - however I don’t see a problem in doing both.

NOTE: I think it is very important to prepare ppl If you plan to send out an emergency email. This should not come like a weird surprise, they have a right to know that this could happen and what should be done in that case. This approace makes false alarms even more of a no-no. (Im stil testing my automation, so far no falls alarms).

@Michael_Davydov I like all the thought you put into this, and yes - a simple approach is the safest in my opinion. However I don’t know if there is any fitness trackers that allow you to pull data? - and what happens whens if the fitness tracker is taken off? Having a button could be a good solution in some cases - but what happens if one had a fall in the shower … The switch should not rely on you being able to activate it, a 911 call would be more appropriate in that case :wink:

Thanks for some interesting reading @wixoff, however im not sure if you have read the topic :sweat_smile:
I think your first link is a great example of why testing is important and falls alarms are a no go, as you hints at.

However toasting servers and deleting embarrassing internet history should perhaps not be the primary goal here :wink:

23

LOL, of course I read the thread. There are some people for whom this would be really useful and I appreciate the efforts to avoid false positives. But I get the sense that for others this would be an “ah, cool, I can do this” thing that would likely end up with someone getting a large and unpleasant surprise for no good reason.

And besides, I’m not saying “don’t”; I’m saying “be careful.”

1 Like
24

@wixoff Coming to this with thought and responsibility is actually the main idea of my post as well. If this system is implemented poorly it can cause more harm than good. But I don’t think anyone here mentioned wiping servers and passwords? However it might be a nice thing as well, if you have something you don’t want other people to see after you’re dead, but on the other hand… you won’t really care, right?

So for me main goal of this switch, is to alert other people of trouble if I can’t do this myself, and preferably in time for them to come and help me. Or someone else in that case. I even described a few use-case scenarios in which it would come helpful.

@GigabitGuy I’m not talking about just one option. It MUST HAVE multiple activation options. Button is just one of them. But if you can’t reach it, system should be intelligent enough to trigger the alarm itself.

I don’t know of any fitness trackers allowing you to pull data either. But maybe it’s possible to pull data from services. For example, here’s fit-bit’s API Docs, it looks like you can pull heart rate data for a specifit time period: https://dev.fitbit.com/docs/heart-rate/

It looks like it’s even possible to pull data in specific time range, I believe with 10 minute intervals?

So I can imagine a system, which pulls this data every 10 minutes or so, and if it detects that there. I can’t check docs thoroughy right now, but I think there might even be some heart-rate alarms built into fitbit API?

As for when you put it away - I don’t know if fitness trackers can detect this. But if they do, it should be as simple as just turning alarm trigger off. If you put your fitness tracker off, then heart rate won’t trigger any alarms. So just keep it on at times when you think something may happen. From what I know you can wear them 24\7. And even supposed to, to get accurate data over time.

Another layer of conditions could be your activity in social networks and something else. So we come to trigger-condition logic of HASS :smiley:

So, Trigger - very low heart rate.
Conditions -

  • you are listed at home
  • No movements in your home for an hour or two or more
  • You do not respond to notifications (mentioned in my post, emails, sms, pushbullet, actual alarm in your house, anything to get you to push the button “False alarm”).
  • Maybe check social media activity (if you were in facebook a minute or two ago, then you’re likely ok)

BUT. All this should only delay the alarm, especially social media check. Once triggered the only way to turn it off permanently (until another trigger) is you pressing the “False alarm” button (well, could be a hass trigger).

Still, it’s just thoughts yet. Not a manual :smiley: I’m just thinking out loud.

1 Like
25

@wingers1290 @joch Another way to check if your phone is charging is to use Tasker. I have mine set up to send a mqtt message anytime it is charging (and when it’s unplugged), but you could also use http requests if you don’t want to mess with mqtt.

26

Great feedback @justyns , sometimes i would like to be an android user. I’m currently trying to use the icloud component battery_status attribute as a template switch - not having much luck so far, but it should work as well in theory for all us iOS users.

UPDATE:
Using the iCloud component, it can be done in automations like this.

- alias: 'Use the battery_status attribute from iCloud to trigger an automation'
   trigger:
    platform: template
    value_template: "{{ is_state_attr('device_tracker.iphone', 'battery_status', 'Charging') }}"
27

Hello. I am a caregiver. Currently I am taking care of a 63 year old man who has multiple sclerosis. He does not have control on his lower extremities and has very poor dexterity.

Being a care provider and tech enthusiast, the discovery of Home Assistant (thanks to HA evangelist, @bruhautomation) was a way for me to incorporate the two to help the person I am caring for, make life a bit easier for him. With a very poor dexterity, he was constantly reaching for remotes to control a roller blind, control the TV, push a button to unlock a door whenever his PT would arrive, text or call someone, etc… With his Google Home + HA, it makes life easier. So massive thanks to HA and Ben!

His mother (almost 90 years of age) is still very active and independent. She drives herself around to go shopping and see friends and family. The son would call her every single day in the morning to check on her if she is okay as she lives alone in her house.

She came by to see her son yesterday and was asking me of some device that she could use in case of an emergency (she falls down) and notifies love ones that she needs help. Immediately I thought it would be a very simple project to make. Just setup HA and a wearable button then just push that button whenever she falls. Reading this switched off the light bulb that sparked yesterday.

I am posting this long and maybe boring for some to read to say that although this is a technology that is available (very expensive) that can be bought in the market, it is always good to DIY in a case-to-case basis.

I will be following this thread to hopefully get more logic and idea on the implementation. Thanks guys!

3 Likes
28

Hi @charlesdalid, its great to hear from somebody that is actively using the technology in such a way, and in use cases that can really make some lives a bit easier.

Please share your experiences using the tech like that, and feel free to hit me up if I can be to any help (being a former caregiver myself).

29

@charlesdalid Indeed, these are use-cases for it that I greatly appreciate, and look into as well. But I don’t have to look after older people, not yet at least.

But speaking of your situation… It should be easy to make a simple button for a person to press if they fall down. But if they can reach for a button, they probably can just reach their phone and make a call themselves, pressing, well, 2-3 buttons instead of one. The automated trigger should be able to detect if they actually fell or fainted. Is that what is stopping you?

I may suggest an idea here. It’s still harder than just a button, but much easier than what we’re talking about here, considering that it should be a wearable.

Now, right now it’s just an idea, implementation may be impossible or hard, I did not look into it yet. I think it should be possible to use smart-watch or at least smartphone’s gyroscope to detect fall. Using smartwatch or fit-bit - it’s possible to detect heart rate. Now, these two sensors may trigger an alarm. Alarm should them turn on on person’s phone, loud and clear - serving two purposes. 1 - try to wake the person up. 2 - warn him\her that if this alarm is not turned off, it will warn other people that something happened. If no one turns this alarm off in, say, 10-60 seconds, it notifies other people, sends GPS location, whatever is required.

I think this project may actually work. But probably not in relation to HA. As a separate app on the phone or something. The tricky part here is to actually detect a fall, and figure out how heart-rate and whatever else information you can get from fit-bit’s API can help.

Don’t let that light bulg die out :slight_smile:

30

About your idea for a (panic)button @charlesdalid, you could perhaps use a z-wave keyfob as the panic-button, attached somewhere easily accessible to the individual person (on their pants, shirt or on a loose necklace, etc.) … it might be a lot easier to simply tap a button for a lot of individuals then rely on a fiddly smartphone.

I found these Aeon Labs keyfob’s, and they seem to have the most simple (and prettiest IMO) design. . Aeon Labs Panic Button (youtube)

A lot of other keyfobs have 4+ buttons - not ideal in an emergency :thinking:

I hope this can lead to some inspiration for you and others in need. :slightly_smiling_face:

On a sidenote, I just ordered a couple of smartlocks (Danalock V2) on an insane sale (its on a danish shop and gone now sadly), and it got me thinking that in a DMS or panic-button scenario it might be a good idea to have smartlocks unlock the doors for easier/faster access to help. This of course, again, requires a non-false alarms tolerance, but might be worth considering in the individual scenario.

31

Maybe the flic button is an alternative one. A cheap one would be the smart button from xiaomi but only if you already have the gateway :slight_smile:

32

I will add this here just because I think it is interesting. It could be an ideal somewhat passive monitor although it has no direct way to connect with HA.
https://www.aurahome.com/elder-care

33

just a quick thought but have you looked at Bens video hacking a amazon dash button think it was via IFTTT had a few / 10 second delay but maybe an easy solution.

34

Dash buttons are not particularly reliable. I wouldn’t want to trust anyones life to one.

35

I have implemented this I think it’s a great idea but needed it to be fool proof as false positives of sending alerts to family members when my wife leaves her phone at home are not nice.

I had several scenarios where false positives could occur.

  1. At night when asleep well this is a no brainier add condition for night time but what happens if accident happens at night? I added a bed Occupancy sensor to take care of that so if in bed triggers input boolean happy days.

  2. We leave our phones at home, I use Bluetooth tracker so phone at home with no movement would trigger dead man. I have added key chain Bluetooth so this is used to trigger dream man as she always takes her keys. (well forgets them less then her phone)

Is there any other false positives anyone else has come across just trying to avoid worrying the family unnecessarily.

36

add motion detectors.
ass long as there is motion then there is no dead man.

no motion for some time and the phone at home, would be worrying.
the more motion detectors, the more you can rely on it.

but i wouldnt implement such a thing if more then 1 people share a house. the chance that both (all) are in trouble is very small. or just an alert to each other, nowing what can be the false positive.

37

Except for people who regularily forget their phones at home. I couldn’t do that to check if my mom is ok because at least once a week she leaves her home while forgetting her phone. That would then trigger an alarm even though she’s fine.

38

yeah right.
so the phone is not an option for detection in that case.
i then would rely on motion on expected times. or extend the timeframe.

a motion detector near the bed tells if someone gets out and into bed.
a motion detector near the frontdoor tells if someone left the house, if it is the last motion.
a 3th motion detector in the living area or kitchen tells normal activity.

not in bed and at home needs to give movement in the living area at least once in a while.
in bed or not at home tells not to worry.

a few days would be enough to tell how long between motion in the living area is normal.

39

Just read through this topic and it has inspired me to start working on an “Emergency SOS” setup… thanks for the inspiration.