Binary State False Positive Handling

,
1

Hi all, first post! I use Frigate with object recognition to determine, for example, when a car is in my driveway. The binary sensor goes from ‘cleared’ to ‘detected’ when a car is detected in my driveway zone. The problem i have is that, while a car is sitting in my driveway and is generally ‘detected,’ the object recognition sometimes falls below the threshold for probability that the object is, in fact, a car… at this point, the state goes to ‘cleared’ for a brief moment, before going back to ‘detected’.

I have an automation to notify me when a car enters the driveway, but this notification is triggered randomly as it briefly goes to a ‘cleared’ state and back to ‘detected’ when a car is sitting in the driveway. I can set the automation to ensure that the ‘detected’ state lasts at least 5 seconds before triggering, but i don’t see a way to ensure that the automation does not trigger if the ‘cleared’ state was not a false positive ‘cleared’ - for example, I’d like an IF condition that the ‘cleared’ state lasted longer than 5 seconds before the ‘detected’ state can trigger the automation.

in summary: car is sitting in driveway for one hour. ‘Detected’ state for an hour (automation triggered an hour ago). ‘Cleared’ state for one second (false positive object recognition probability hiccup). ‘Detected’ state again… automation triggers again undesirably. Fix: need to ensure ‘Cleared’ state lasted X seconds for automation to run…

Any ideas?

2

Create a template condition, use this template:

{{ now() - trigger.from_state.last_changed > timedelta(seconds=10) }}

The automation will only continue past this condition if the previous state of the sensor (‘cleared’) had remained in that state for over 10 seconds

3

that’s awesome, thanks… will test it but it seems to be exactly what i was looking for, thanks again!