Michael Lyle

Hangtime no longer applies to failsafe

2

stabilization.c is gettin' to be a mess.

A) observe cross-axis coupling B) compute a mixer that eliminates, say, half of that coupling (because we don't know causation and because phase margin, etc)

1. We take out the bias from the gyros at the sensor level... using perhaps old / stale data... and then put it back in. 2. I think the way...

The offsets used in baro pressure are not applied to the GPS. This means that offset between sealevel and home do weird things with the INS. Relates to #443

Synchronous gyro interrupts on flyingpi

10

depends on ~~#1714~~ , ~~#1762~~, and ~~#1779~~ Still a little hackish (as this call is not universal between stm32 and posix yet, so bmx055 can't work on STM32). But this...

- [ ] Autotune support - [ ] Automatically increment version numbers - [ ] Auto-connect to bootloader considered harmful - [ ] Crashes when you enter a letter on...

Attitude estimates often poor in fast-forward-flight or in fixed wing

2

Shows up really blatantly on OSD, etc. taulabs/taulabs#2021 etc

Vehicle setup wizard improvements

3

- Checkbox to use "naze"-style motor mapping - Brushed motor output type.

in-flight mag calibration

1

Get the scales right, ignore biases from currents, etc. Fit to a sphere while flying. At first offline (requiring a commit) but ultimately while-nav-runs.

* UAVOs have a very strange compiler attributes (alignment 4, packed) * The UAVO manager stores them with uint8 alignment trailed onto non-multiple-of-4 structures * The callback interface passes this...