KPIs/Metrics for PX4 Autopilot

Hi All,

I am trying to find documentation that explains how is the PX4 Autopilot performance being evaluated, for example:
What are the defined tests? are they manual/automatic?
Are there specific environments that must be used? (outdoor, indoor)
What are the defined metrics that are being used?
What is the quality procedure being performed before shipping out a new stable release? (for example: latest v1.13.1)
Is there an official standard for Autopilot SW to enable comparison between different Autopilot solutions?

Would appreciate any help with gathering the data :slight_smile:

Hi @rroche,

Could you help with getting the information (docs/links or anything else)?

Thanks,
Ronen L.

Hi @rroche @hamishwillee
Kind reminder on this post :slight_smile:
would appreciate any help/direction on this topic

Thanks,
Ronen L.

There are a whole bunch of automated tests that get run every time a changes is submitted. As I understand it almost all of them are outdoor since PX requires a position estimate for most modes, and providing that would require offboard (companion computer) support.
These essentially sanity test that a change will not break any existing behaviour.

I don’t manage the release but my understanding is that a version of the mainline is taken well before the release and flight tested by members of the dev team until critical bugs are fixed. I am not sure how much formal testing takes place or how this is managed.

Is there an official standard for Autopilot SW to enable comparison between different Autopilot solutions?

No.