Hmm... I'm not sure .. would need to think about it some more. Best I can think of would be two separate rules, one for each sensor, setting a variable or virtual switch once each one reaches the timeout points (15 / 30 mins). In a third rule, once both switches / variables are on, do what you want to do and turn them off.
Not ideal, so hopefully someone has a better idea.
I’ve got a very outside of the box idea that might work, but am completely unsure if it is the best practice. My thought is this…utilize the Zone Motion Controller app to create two virtual motion sensors. Select Motion aggregation but only select one sensor. Have the activity timeout be your length of your stay for the respective sensor. Do this twice, once for each sensor, so that you have two virtual sensors that mirror your intended triggers.
Once that is done, you need to write two rules to have it work. Take advantage of the new conditional triggers or use required expressions based off the virtual sensors. For instance:
That is not how triggers on Hubitat works. Hubitat is event driven, if you trigger on an AND two events would have to occur at the exact millisecond to make the rule run. The OR in the triggers is the correct thing for this platform.
Your AND belongs in the Actions section of Rule Machine. You probably want a conditional action, but there are likely other ways to accomplish what you want to.
THEN do something
There are other posts from the app's author bravenel, that explain the AND situation better if you want to research why this works this way.
Edit to say:
Your triggers likely would be a simple "motion inactive" for these sensors. The 15/30 minute inactive period would be in the Actions section in the conditionals.
Another way would be. Trigger on both with an OR. And then set a condition action with an AND saying if sensor 1 has been inactive for at least 15 minutes and sensor 2 has been inactive for at least 30 minutes then run the rule.
To expand on @neonturbo 's point, and likely other's thoughts, the HE platform, and in particular RM, is based on events, which can now include a single device remaining in a single state for a period of time, but not more than one device in a state for different periods of time.
Your circumstance and/or alternate options in the HE platform are the more likely paths to finding a solution I expect.