- mavlink in PX4/Firmware (56106be4805cb2f70d22ad31b87fbd9774f0d0f3): 99e82cad70
- mavlink current upstream: 0909631552
- Changes: 99e82cad70...0909631552
09096315 2022-05-26 Hamish Willee - Clarify and cross reference home position everywhere (#1836)
38c1a99f 2022-05-25 Julian Oes - WIND_COV: clarify description and unknown value (#1845)
H7 Only supports 2 not 3 CAN interfaces.
CanInitHelper passes in in the run time configuration of
the number of interfaces. The code was ignoring these!
This type (23) doesn't specify a motor number, so it can't be properly handled.
There are duo (19) and quad (20) tailsitter types that still work in simulation.
Signed-off-by: Silvan Fuhrer <silvan@auterion.com>
Based on feedback that very often the battery is used down too low.
I observed this happens consistently when the cell voltage is properly
load compensated. The default load compensation before #19429 was very
inaccurate and resulted in unpredictable estimate.
After that if there is a usable current measurement and the battery is
within expected tolerances of the default internal resistance the
compensation is pretty good and 3.5V is too low for an empty compensated
cell voltage. That was seen in various logs where the compensated
cell voltage was already dropping fast after 3.6V.
In case the voltage is not load compensated the vehicle estimates the
state of charge a bit too low which is safer than to high
especially for a default configuration.
internal PX4IO safety_off state is removed and replaced with a normal safety button event. From this 'commit' commander is taking care of the PX4IO safety.
The baro observation noise parameter is often over-estimated in order as
a measure to mitigate temporary offsets in the readings due to wind
gusts or poor pressure compensation tuning. The side effect is that the
innovation sequence monitoring based on normalized innovation struggles
to detect an offset in the state because the innovation isn't
statistically significant.
To counter this issue, a simpler check is added to trigger the process
noise boost when the innovation has the same sign for a long period of
time.
We use lockdown to prevent outputs like motors and servos from being
active in HITL simulation. This means that we can't treat the lockdown
flag as a flight_terminated, otherwise we can't arm in HITL at all.
I got multiple times the feedback now that a consistent delay
is helpful and makes sense but the default delay
is too long
for low battery action where you're trying to come back in time
and possibly the emergency reaction kicks in while the critical action
is executing which leads to a longer accumulated delay.
- mavlink in PX4/Firmware (87c73145b36a835b1635de0498a5613a7af5cafc): a1cb2c0e71
- mavlink current upstream: 99e82cad70
- Changes: a1cb2c0e71...99e82cad70
99e82cad 2022-05-19 Hamish Willee - Deprecated GPS_INJECT_DATA (#1842)
bf3df03d 2022-05-16 Hamish Willee - WIND_COV - accuracy units are m/s (#1844)
a73d4864 2022-05-11 Hamish Willee - development.xml FIGURE_EIGHT_EXECUTION_STATUS - add a note about it being set at lower rate (#1841)
