AP_NavEKF3: added two more EK3_OPTION bits

[tridge]

these allow for easier GPS-denied testing of EKF3, and allows for replay testing without the SetLatLng code being active.

Note that the gaps in the bitmask are for pending range beacon features, we don't want to fill in the gaps so that existing replay logs for that feature remain valid

[rmackay9]

I wonder why we couldn't just use the EK3 source switching feature? there are 3 slots and you can setup one to not use GPS

[tridge]

I wonder why we couldn't just use the EK3 source switching feature? there are 3 slots and you can setup one to not use GPS

if this was being used for end users, yes, but as a development option to test the behaviour of EKF3 under GPS denied environments it doesn't produce equivalent behaviour.

[rmackay9]

@tridge,

This seems like a very rare edge case so I wonder why we need a parameter to set it. How about an auxiliary switch and/or a mavlink command?

[priseborough]

@tridge,

This seems like a very rare edge case so I wonder why we need a parameter to set it. How about an auxiliary switch and/or a mavlink command?

With parameters it is possible to change the behaviour when doing tuning and comparison of processing options via EKF replay.

[rmackay9]

Hi @priseborough,

OK, but can't auxiliary switches also be changed easily as part of tuning? I think the recording of the EKF state in the DAL is separate from how the user configures the feature.

Sorry, I don't mean to be difficult but I saw another one of those usability complaints on our YouTube channel and it reminded me that we need to be careful when we expose additional complexity to our users.

[rmackay9]

sorry, I'm adding a "request changes" because I just want to be sure we're thinking through the UI exposed to the user

[tridge]

OK, but can't auxiliary switches also be changed easily as part of tuning? I think the recording of the EKF state in the DAL is separate from how the user configures the feature.

there are really 2 parts to this PR:

• make it possible via an aux switch to disable GPS in EKF3 but not in EKF2, while leaving the src selection parameters alone. This is essential for doing GPS-denied flight testing where you fly on EKF2 but run EKF3 in GPS denied mode
• make it possible to set this state right from startup, to guarantee that the EKF3 code has never ingested a GPS sample until you move the aux switch to enable GPS in EKF3. We don't (as far as I am aware) have a way to set the initial state of an aux switch. The EK3 code initialises pretty early, often well before the user has properly connected to the vehicle to trigger an AUX switch

These features are being used as part of the testing of the beacon GPS-denied functionality, but Paul and I thought we should PR them as they are useful capabilities for anyone else trying to do GPS denied testing