Zone x Occupancy Falsely Detected, overall Occupancy detected, when Zone x Target count = 0

[clintebbesen]

As per the image below, Occupancy is showing as Detected, however there is no one in the room, no motion. Zone 1 shows occupancy Detected, count of 0.

There is a target being tracked, I am standing in the next room, outside of the Zone 1 bounds. This can be seen in the Zone configurator in the image below.

[clintebbesen]

Ive tried to start using another sensor in a different room, having the same issue.

I opened the zone configurator for debug purposes, as you can see it only shows 1 target, I am in Zone 1, however Zone 3 is still showing Occupancy detected, even though I have left that zone 5 minutes ago, and the zone 3 target count is 0

[clintebbesen]

I have a third device with the same issue, this one has had the false trigger for the last half hour holding the light on, I have left the room for 20 mins and come back and its still showing the same false trigger. This one is showing two active targets in the zone configurator when the two other times mentioned above it only showed 1 active target. On the device page in HA it is only showing a target count of 1 though, also realised now the Zone occupancy is wrong there too. The zone configurator shows the target being in zone 1, but zone 1 shows clear occupancy, and zone 3 falsely shows occupancy.

[clintebbesen]

a 4th Presence sensor is doing the same thing now, realised the bathroom light had been on for a few hours.

Target 1, 2 and 3 active are all off, target counts are all 0, overall occupancy is clear, but zone 2 and 3 have occupancy on. Different to the other times its happened in other rooms, this one shows all 3 targets as off.

Is there a way to do some internal checks and if its own values don't agree with each other then do a reboot? I would assume there should never be a case where overall occupancy is clear, but one or more of the zones show as detected, this I assume would be considered an error state, and could trigger a reboot or something?

Is there any more details I can provide to help narrow down the bugs in the detection code?