I have been playing with the False Motion Reduction Zone Controller.
The setting of the "Activation Window" does not quite appear to do what I think it does. I read the explanation in the Documentation over about 10 times and my interpretation of the words do not quite correspond to what seems to happen.
Can someone explain the "Activation Window" in another way?
The documentation states: "This is the amount of time all motion sensors specified in the zone must detect movement before the zone will activate."
I read that statement over and over again, and also looked at the timing diagram in the explanation. The timing diagram seems to show something different to me.
A few questions... is it edge triggered, or do the minimum number of devices have to be on simultaneously for the entire duration of the specified window, or do they both have to be on simultaneously for at least part of the window?
I have tried a number of things and have still not quite figured it out. I am still not sure, but it looks like what the words in the documentation say are not quite what happens, or else I am interpreting it wrong.
At least I am getting some exercise, running around different parts of my yard at different speeds to trip detectors.
Just did some testing with this using virtual sensors since I wasn't sure about some of these nuances. To summarize what I found:
- "False motion reduction" requires at least the "minimum activation threshold" number of sensors to generate a motion "active" event within the specified "activation window" timeframe
To be specific:
- The "activation window" starts counting down when the first sensor sends an "active" event
- If your activation window is 5 seconds and sensor A becomes (and remains) active at time 0 and sensor B becomes active at 6 seconds, the zone will not activate, as the "active" events did not happen within the specified time (this is true regardless of whether sensor A is still active)
The last thing is what I wasn't sure about. I still have this test setup configured, so if there's anything else you want me to try, let me know.
That's sort of where my mind is going as well.
I am not sure yet if the required number of events have to overlap or can be sequential within the "window" as I have all my motion detectors set for a 30ish second timeout. Some are hard wired into the specific detectors to have a 30 second timeout(Iris seems to have around a 30 second timeout for instance) and for others I was able to select the timeout period in software. It would be good to know for sure if they have to overlap or not. The timing diagram in the example shows them as overlapping but nothing says that they actually have to be.
Also, as near as I can figure from reading, if other detectors continue to go off once he window starts, I believe the output is maintained in an active mode. I am not sure yet if it is maintained in the active mode until the "activation window" value has been reached after the last detected event ended or if it ends in that much time after the first or after the second trigger occurs.
More experimentation to do tonight I guess.
The activation window is just the amount of time you are giving the selected number of motion sensors to trigger in order to turn the virtual motion sensor on. You can then select other motion sensors to keep lights on in ML. I use the same ones from my group. That way I don’t have to keep setting off all of the sensors within a certain window in order to keep it from turning off.
I wasn't sure how to interpret the diagram, so I just tested instead.
I think I addressed your question in my results. The only thing that matters is the time at which your sensor sends an "active" event. If enough different devices (greater than or equal to your "minimum activation threshold") send "active" events in the specified time period (the "activation window"), then the "zone sensor" becomes active. Whether any sensor that went "active" stays active or becomes "inactive" during that time does not matter.
- Activation window = 5 seconds
- Minimum activation threshold = 2 sensors
Sensor A becomes active at time 0 (and stays active), Sensor B becomes active at 4 seconds = Zone Motion sensor becomes active
Sensor A becomes active at time 0, Sensor A becomes inactive at 3 seconds, Sensor B becomes active at 4 seconds = Zone Motion sensor becomes active
Sensor A becomes active at time 0 (and stays active), Sensor B becomes active at 6 seconds = Zone Motion sensor does not become active
OK... so it sounds like I did not have to arbitrarily make my detectors all be a minimum of 30 seconds long to force a guaranteed overlap then. I will try making them all one second tonight and see if it makes any difference.
Thank you !
Wait, you have motion sensors that will flip to "inactive" after one second? That is crazily low. I could see that being do-able for mains/USB-powered sensors (and beam sensors, if any exist for Hubitat), but that's practically unheard of for battery-powered sensors (and would likely draim the battery quickly in areas with lots of motion). Many stay "active" for 30-60 seconds or even several minutes by default, with some (mostly Z-Wave, it seems) being configurable to lower. Many have a few similar-sounding parameters that do slightly different things.
Anyway, don't want to derail your plans if this works for you or you're already aware of the above...just seemed a little odd. But yes, I agree that you don't need to worry much about the sensors themselves--ZMC uses only the "active" event times and doesn't care how long they stay active. Good luck!
What about all of those messages being sent over the mesh. There is only so much bandwidth. I actually changed my Hue motion from the default of 10 secs? to 30, because I couldn't shut off the light without it coming back on before leaving the room.
You two make good points... I will leave it at 30 seconds then.
Right now I have to identical motion detector zone controllers set up... one with a 1 minute window and one with a 10 second window. I will check the logs later to see which one seems to work better.
And. the massive jump in the settings options... 10 seconds ... and then one minute. My yard is huge. 10 seconds seems a little short, even with a large number of sensors. At the same time, 1 minute will give the bugs a lot of time to do their magic and trip at least 2 devices. I wish there was a 30 second window option.
4 of my "detectors" are WYZE cameras. A bug very close by (in front of the lens) will trip them, as will a bat. In the day time, birds going after bugs trip them constantly. The outdoor light only goes on at night, but eventually I will hook up some sort of audible chime or something during the day. I wont do that now though... too many false alarms.
Ive got both the 10 second window and the 1 minute window algorithms running and I'm watching them go off for random reasons... bugs, cats, stray wildlife...
What I see gives me a thought... maybe both are useful. 2 detectors for the 10 second window and 3 or 4 sensors for the 60 second window maybe ? Not sure... need to play with this and review logs regularly.
You should just get a couple Hue outdoor motion sensors and be done with it. You can set the sensitivity and reset time on them and they also give temp and illuminance readings.
Two of my sensors are Hue sensors. Two are Zooz. The rest are Iris, and there are also 4 Wyze cameras.
I have a 35 acre farm. I am mostly concerned about 2 acres in front of the house.... hence the need for a large number of sensors.
With a lot of sensors come a lot of false triggers.
I am trying to figure out which sensors I like the best, and also trying to optimize the detection algorithm to reduce false triggers and increase positive triggers. It is a bit of a challenge .
I do have all sorts of wild animals in the yard all day and night as well. I would prefer to not detect them as well.
You should have started with that It would have made things much more clear what you were after.
I personally have a couple wyze cameras. I wouldn’t trust them to be as reliable as a Hue or Iris sensor (which I also have) that is directly connected and supported by the hub. The nice thing about the outdoor Hue sensor is it’s big battery that should stand up well to the winter. I like the Iris as well. I have 16 Iris, 3 Hue, and 3 Ring. I can’t tell a difference really between the zigbee sensors, but the zwave plus first gen Ring are very slow by comparison.
Hi Ken, I am using them all right now to learn what works best. I'm sure that summer and winter will be different.
Yes, the WYZE cameras are a little "flaky" . Sometimes they respond better than a motion detector and sometimes they are not the best choice. Hence the reason for using multiple sensors and summing them all together.
I think I have also figured out how to add weighting factors of sorts to each sensor, so if any are overly sensitive, I can give less weight to the overly sensitive ones. I will start to implement that in a couple of weeks once I get more history with the sensors. Having the ability review logs every morning is an awesome feature.
I’m glad to hear it’s working out for you . With such a large outdoor area, you might even want to create several “ false motion reduction zone” zones, and use the zone devices created by the app in an all encompassing “false motion reduction zone”.
yup, that has occurred to me... as soon as I get a little more experience with these sensors, I plan to do jut that. It would allow me to possibly add a feature to audibly alert me to where the motion is occurring so that I dont have to bring up Hubitat to look at things.
If you use Alexa, with the Echo Speaks community app, you could do that in the app.
I am still trying to figure Echo Speaks out.
I saw someplace where it is no longer working but I’m not sure that is true.
Do you know for a fact that it still works?
If so, is there a good Echo Speaks Install for dummies Tutorial out there. It is definitely something I am interested in doing in the very near future.