VA Smarthome Project


(Jeff May) #1

I am new to SmartThings but am really excited about what it will offer. I am working at the VA Hospital developing smarthomes to monitor and provide care for patients. The project is focused on patients with brain injuries, but is really applicable to all sorts of patients. The idea is to monitor a patient at home using various sensors and tracking technologies to make sure that they are safe, taking their medication, exercising / walking around, eating, bathing, getting to their appointments and all sorts of other things.

Continuing the discussion from What projects are you currently working on?:

Monitoring Elderly Parent - Need a little help and advice
(Greg) #2

Can I ask which one?

(Jeff May) #3

Working at the VA hospital in Tampa.

(April Wong) #4

WOW! This is amazing, and extremely exciting. Thank you for what you’re doing to take care of our Veterans!

(Tim Slagle) #5

Super cool man!


Welcome to the community! It sounds like a very exciting project!

I am myself quadriparetic, wheelchair dependent with limited hand function. So I’m very interested in various forms of home automation, although my issues are all physical, no cognitive stuff.

For example, right now I’m considering a couple of wall switches based in part on how easy they are for my service dog to use. :blush:

In fact, the first major problem SMartThings solved for me was the ability to use a motion detector as a touchless switch so I can unlock the deadbolt on my door to let someone in. I can work this bolt manually with two hands, but it’s tiring. Being able to do it with a hand wave makes every day easier.

I also have a light set to turn on as a persistent reminder that it’s time to feed the dog. It’s a small table light, not obtrusive, and not something I have to respond to immediately. I like it much better than an audible alert. I don’t have to respond instantly, but I won’t forget either.

Jawbone’s UP24 integrates with ST. I don’t have one, but a number of community members do. @smart has one and really likes it. If the person will wear it, that gives you an activity monitor.

Anyway, lots of possibilities!

(Convinced ST will never be unbroken…) #7


Really neat to hear of this project, but I am not sure SmartThings is the right choice for critical mission tasks. You have to ask yourself what is the worst case consequence of having a missed/delayed event or notification in your system design. If it is more grave than “minor inconvenience” I would suggest you look into something a bit more hard wired.


BTW, I agree with @scottinpollock about mission critical issues. I use SmartThings for convenience, but there’s always a backup option for if the system fails.

I don’t use SmartThings for my security system, that’s hardwired with professional monitoring. And I don’t use SmartThings for a panic button, I use a GPS Cellular device which I can also use away from home.

On the other hand, as a quad a lot of things that are mildly inconvenient for a more physically able person are exhausting and time consuming for me. So even if the convenience option fails 20% of the time, my quality of life goes way up.

Belt and suspenders, but SmartThings is a pretty snazzy set of suspenders. :smiley:

(Geko) #9

I agree with @scottinpollock. ST is fun and games for casual use, but you have to be very careful if you’re going to use it for security or healthcare related applications.

(Jeff May) #10

Yeah, I used to work on the Space Shuttle propulsion system before coming here so I definitely understand about critical systems and redundancy. Here are some keys to what we are doing…

  1. Agree - Never trust something like this for a critical system where likelihood X consequence is unacceptable.

  2. There can be redundancy and conclusive evidence (i.e. camera and we track the actual PATH that patients walk).

  3. In many of our cases, we are just adding something to make what we already had better.

(Jeff May) #11

What I meant by #2 is that you can assume if you have data, then the action happened. You just cannot assume the inverse. If I see all the data evidence that a patient took their meds I can assume they took them (from a system reliability standpoint). If I don’t see that data (or the system goes down) then some action or intervention may need to be taken.

What I meant by #3 is that in many cases, a patient is already going to be staying at home (with or without a caregiver) but now you can have more confidence in knowing what is happening. There is about to be an EXPLOSION of elderly that will HAVE to stay at home (just not enough facilities) so we will need better ways to monitor and care for them.

We are using more than just SmartThings, but one of our use cases was a quadraplegic whose wife and caregiver passed away and nobody knew for 3 days. Our system would detect no motion or activity and a nurse could call or follow up. Could go on and on.


BTW…for all the usual accessibility reasons, I prefer iOS to android for most small devices. (Love my ipad!)

I really, really want one touch voice control of SmartThings. There are no plug and play options yet except a very limited Samsung Gear S Smartwatch capability that only does one or two things. But I don’t expect any Home Automation system to have a solid voice option until summer 2016.

Meanwhile, though, one of the community members here has created an android phone smartapp called SharpTools which does offer some real voice functionality. Not quite enough to make me invest in android devices, but if people are already using them, definitely something To consider.

I know not everyone has the ability to voice, but it’s something else to keep in mind.

(Jeff May) #13

One of our biggest uses is keeping patients on task. It is not critical that a patient brushes his teeth, or makes his bed, or bathes or even eats (assuming that a caregiver or clinician sees him on some regular basis). Even the meds, we use a tray where a caregiver can see if any meds are still in the pillbox etc.

Where we are testing this is actually in apartments in the hospital. Even if it just stayed here (where we have nurses) one of the things we want to see are whether the patients are taking showers.


Several people are using humidity sensors very successfully with SmartThings.

@jody.albritton uses them I think to turn on a humidifier.

@rayzurbock has an app that turns off the bathroom light when there’s a period of no motion in the bathroom, BUT extends that time when the humidity level is rising, on the assumption there’s someone in the shower.

Very clever. Not for me, since my first question is always “Will the lights be on after a fall?” But, hey, “your use case is not my use case.” :blush::umbrella:

Anyway, he has a lot of experience with ST and humidity detection during showers.

(Brian S. Lowrance) #15

Thanks for the mention. I’ve severely neglected my Bathroom Light Control SmartApp. It works for my situation (most of the time). I have some feature requests, mainly ability to control multiple lights such as a vanity with 4 or 5 smart light bulbs each with their own on/off “switch”. Right now it controls 1 switch for the light and can turn on another switch when it switches to humidity control (for a fan, heater or other light if desired) then turn that switch back off when the humidity drops back down. I’ve been focusing much of my development time on the Big Talker SmartApp. I’ve learned a lot during development of Big Talker. I’ll soon go back and revisit the feature requests for Bathroom Light Control.


I know you’ve been busy! :blush:

I just figured you might be able to share a little of your expertise in using the multi sensor for a bathroom humidity reading as a “shower on” indicator (as opposed to trying to use a flood detector) since that might be particularly helpful for @jmay33 .

Especially since using the multi sensor gives you a whole lot more choices about where to place the sensor in the bathroom, right?

(Brian S. Lowrance) #17

Ah… Understood. I am using the Aeon Multisensor that has sensors for motion, light level and humidity. These sensors only read / report Humidity by default every 8 minutes to save battery. I’ve noticed it only seems to send information if there has been a change though. This makes Bathroom Light Control’s detection, averaging and forecasting very limited. I’ve also read about being able to make a custom Aeon Multisensor device type in ST with special “Configure” options to adjust the reporting period as desired down to 1 minute. I’ve not spent much time on that, but I was not successful in the brief attempt I made, plus I don’t think a lot of people would adjust it and so I wanted to write Bathroom Light Control to work with it as is. When someone enters the Bathroom, the motion sensor picks them up and turns the light on. The SmartApp starts the “Motion” timer, which starts each time motion ends but restarts anytime motion is detected again. Once the timer finally expires from no motion, it turns the light off. However, if a rapid humidity rise is detected, I switch to a different timer “Humidity” timer, which then turns the light off after motion is not detected for an extended period (10? 15 minutes; user configured). I try to detect that someone is in the shower within 1 reading. To do so, I average the last x (default 4) readings and detect a humidity report greater than 1 whole percent rise from the average point or 1 whole percent fall from the last reading. I tried a more sensitive .5% but that was too inaccurate. 1% up from the average or down from the last reading seems to be the sweet spot, the issue I see is if someone sets the light off period based on motion to 5 minutes and the humidity off period to 10 minutes, the person could be in the shower and an 8 minute humidity reading not occur to switch the off timer method from “motion” to “humidity” and thus allowing the light to turn off. My whole purpose is to turn a light on when someone enters the room and turn it off when they leave. When they enter the shower and close the curtain, the motion sensor can no longer detect them, so I want to extend the lights run time to prevent it from turning off on them, then set it back to normal once the app no longer thinks they are in the shower. If configured properly this should allow you to place a multisensor in a bathroom across the room, away from a wet zone and allow you to control lights. My mutlisensor is above my doorway, it detects a person about 2-3 steps in and turns the light on. It consistently switches to humidity while someone is in the shower and back to motion control after they exit/leave the room.

(Brian S. Lowrance) #18

I suppose another method would be to detect person entering the bathroom via motion and start a “no motion” timer, but provide a button that could be pressed (rather than humidity readings) to signal shower/bath timer and then extend the “no motion” timer to 30 minutes or so.

(Jeff May) #19

I would have considered motion detector that can catch over the shower curtain. We are testing the use of a ST moisture detector in the shower and I am really interested in trying the IRIS moisture sensor (It has a probe so easier to position and electronics box is out of the shower and it is $20 cheaper).

Aeon multisensor can only refresh in 4 minute intervals with 4 minutes being the min.

It would be nice to get a flowmeter to put inline with showerhead and sink faucet. for what we are doing it is important to know things like this.

(Paul) #20

You could put a temperature sensor wrapped in insulation around the hot water supply pipe. If the temperature is above 100°F, someone is in the shower.

That’s a bit less work than a flow meter, but a bit less accurate of course.

I’m going to try something similar to detect if I need to run my hot water recirculation pump when I enter the bathroom. Just haven’t had the time or the spare temperature sensor yet…