Presence sensing sporadic and unreliable; Fixed

That only works if many others report the same issue that is also reproducible to help us narrow down a bug in Mode Manager. However, if the issue is not being reported, then the likelihood of your problem getting resolved is none. I suggest creating a thread and sharing more details about your set-up. It sounds like the problem is not the geofence but the Mode Manager settings, in your case.

5 Likes

You could try setting up a virtual presence in HE that you share with google home, then use the Home/Away routines in google to toggle that presence. This is an interesting solution I had not thought of before. Might be worth a try as a backup to Tasker. I am betting that it will not work if the internet is down though. (Tasker will at least work when I arrive regardless of internet as long as my router is up. However, it won't detect away as it would have to send through the cloud - Confirmed this last week when the "bomb cyclone" took out our internet)

*** Edited to add, the virtual presence created by the Combined Presence app is BOTH a presence and a switch. If you can't toggle presence in Google home, you can toggle the switch. ****

2 Likes

This thread started as an iOS geofence issue, but derailed into an all inclusive geofence problem. Unfortunately Tasker doesn't exist in the rather tightly controlled operating system that Apple offers. But for folks heavily invested in the Apple universe, that use the new Hubitat HomeKit integration, one must wonder, why users look for other solutions than native Apple presence sensing, which based on general feedback is as good, if not better than Google's own presence sensing. Not suggesting that users shouldn't rely on Hubitat features, but might be a worthy alternative for those who have issues with Hubitat's geofence feature that cannot be easily resolved.

4 Likes

@bobbyD I think a common denominator were seeing in this thread is that the app does infact see when we're in or out of our geofences. The problem seems to be that the hub isn't reacting to the geo information from the app. It seems many are reporting the problem lies in the communication of that "in or out" data from app to hub.

Which part of this conundrum are you seeing as the problem? App to hub communication, or the app knowing whether the phone is in or out of the geofence?
I think this is where people are confused in regards to all of this as it seems os changes, battery optimizations, etc shouldn't affect app to hub communications.

3 Likes

I can tell you that part of my issues were at one time related to the location services on the phone itself. Google Maps, Waze, and the HE app would show my location properly. But, I later found that even though the location coordinates were accurate from a GPS standpoint, the calibraton on my compass was off. I had read an obscure post about it when I was researching out of frustration back in my smartthings days. What I found out was that Play Services, which is what the apps rely on for location, does not work properly when the compass is out of calibration. They recommended an app (GPS Info and Status). I got that app, and sure enough, even though my location showed properly on all the apps, my compass calibration was red nearly every single time my presence was inaccurate. When I ran the compass calibration, it would update my presence correctly in the ST app within 5-10 seconds.

If you have a magnetic case, a holster with a magnet clasp, or use a magnetic mount, that will through your compass calibration off.

But, that is only one of the issues. The other issues are also if it is putting the app to sleep, if play services or apples version of that are asleep, and if you have your own location services configured properly.

Well, that is a discussion that requires further input. If the app marks the presence sensor correctly, but the hub doesn't react, it means the geofence feature works as expected but something else may be blocking the communication with the hub. As first troubleshooting step, one should look at the Mobile App device and check if the status is reflected correctly in the hub's web interface. If it is, then the problem is likely with Mode Manager app settings (or whatever app is used to trigger the hub's reaction). If the status is incorrect, then one might want to look to see if the hub can communicate with the phone. Testing push notifications is one way to test that connection is working.

1 Like

@bobbyD

I detailed that as the case in another post on Christmas Eve/Chrismas Day and tagged support

It was incorrect

Tested

All other notifications received correctly.

In summary:

  • Location moves in and out of Geofence
  • Presence does not update on the device to reflect that position
  • Notifications to phone work normally
  • It worked up to the point of the Hubitat phone app updating to 1.8.2
  • It failed the day of that app update (nothing to do with iOS update which was many days previous)
  • At which point it never worked again despite logging out/in on the app and selecting the appropriate phone. Did this on both mine and my wife's iPhone (iOS 16.2)
2 Likes

That issue is under investigation and will provide an update on your thread when I have any news to share. However, the general feedback on this particular thread has been that geofence feature never worked and should be removed from the mobile app, which is simply inflammatory rhetoric that doesn't help anyone.

5 Likes

Not sure that’s a fair or helpful synopsis of this thread. I think most are highlighting unreliability/inconsistency rather than claiming it’s a feature that has never worked. Not sure too many people have suggested it should be removed either.

IMO, presence sensing is central to home automation….so it’s frustrating when it doesn’t work reliably (natively)
I think, at least in part, this thread is a sounding board for that frustration.

Just my view…

5 Likes

Can't claim to have read the entire thread, but as long as that frustration is channelled in a constructive way, then it can be a helpful avenue to get things moving forward.

1 Like

I can understand that. I too rely on presence sensing, although I am not an iOS user on regular basis, Android presence sensing has its own challenges. The path forward is to make presence sensing more reliable and I can assure you that this at the top of our mobile engineers priorities.

2 Likes

I suppose we’re probably getting pedantic re: the meaning of ā€œmostā€ or ā€œgeneralā€ but there absolutely have been claims that it ā€œneverā€ worked, and requests to ā€œremoveā€ the feature from more than one person in this thread. And IMO that’s having an impact on the overall tone.

3 Likes

And as Bobby pointed out, this is another generalization about some other posts in this thread that isn’t really actionable on its own.

A common denominator would be something shared by all. Is this what everyone is experiencing? Or is it what a few people in this thread have reported?

2 Likes

Maybe someone should make a poll :person_shrugging:

2 Likes

I think some very good information is coming out of this thread. I know I'm armed with a few more things to try, though I'm not looking forward to parsing through this thread again to find them, lol.

One thing that I never thought about before Bobby brought it up are the challenges faced by the engineers in regards to privacy. I definitely appreciate the effort made to provide geolocation without signing away our souls. Kudos!

2 Likes

I cannot agree with you more. That’s why I use multiple methods of presence detection.

Along with a virtual counter. So ā€œpresentā€ requires one of three methods to indicate presence. While ā€œnot presentā€ requires two of three methods to indicate absence.

Edit: I should add. I have never had the three methods not concur. And there's usually less than a few seconds difference in the time at which they indicate I am present.

2 Likes

You can always send me a private message. I kept a close eye on this thread, so I should be able to point out some things you may want to try. :wink:

4 Likes

Polls are only as good as the sampling method used to get responses. A poll will spit out a number at the end, but that number isn’t necessarily more reliable than the posts made by users in this thread.

Sorry, I’m really not trying to be a dick. Just highlighting the difficulty support probably has when trying to get specific, actionable feedback, or identify trends/separate the signal from the noise.

6 Likes

As do I, Combined Presence using ST, HK and HE. Does the trick most of the time. I think many on this thread are just yearning for HE to handle that reliable, by itself.

It’s reassuring to hear from @bobbyD that the engineers are working on this. This reassurance is something I have not seen too much of in the forums over the years (but perhaps that’s my lack of reading!)

I’m more than happy to help troubleshoot this by providing logs or whatever (I’m no expert on troubleshooting such things, I’m not a HE guru)…just need someone to tell me what they need..

I’m sure better fixes will be found if we all work together.

2 Likes

How representative would be to have 100 users who have had some issues and came to the community to look for answers, vs. tens of thousands that never visit the community? Also, just browsing this very thread, the problem cannot be pin-pointed to one source. Some issues appear to be related to the geofence, other issues appear to be related to multiple mobile devices created on the hub side, while others appear to have issues with Mode Manager. That is in essence why it is hard to get it right, but our engineers have proved over the years that nothing is impossible for them to fix.

3 Likes