This behavior is excessively paranoid about clearing the flag so now it's extra sticky. You can only clear the is_crashed flag when:
- changing modes
- starting to execute a takeoff wp (if mission/index gets reset while still in auto)
- while in takeoff and throttle is suppressed it's held false
behavior that was removed:
- clear flag when starting to execute any nav cmd (reached next wp)
- if while crashed, you "start flying again" (non-sticky)
due to zero-ish vibration in SITL, the is_flying check thinks you're not flying, and this thinks you crashed. This is a SITL specific problem but it brings to light that *if* someone actually had a very low vibration aircraft we don't want to crash them. So, the vibe test has been removed for now. A better test is a variance on the accel
The realpath tool may be missing on a Ubuntu distro and we use it in
sim_vehicle.sh. Make sure it's also installed as part of sitl packages.
On Archlinux it's part of the coreutils package which is part of the
base group. Therefore it's guaranteed to be available.
Fix#2730.
Previously this decision was based on the WPNAV_RADIUS parameter which is unexpected (and undocumented) behaviour. Better just to hard-code it to 2m and remove the dependency on this parameter.
There are implementations for float, Vector2f and Vector3f for the Low
Pass Filter and the *2p filter. I tried to implement these filters
with one common template implementation. This implementation saves
some lines of code and reduced the redundancy. One could save even
more code if the currently overloaded isinf/isnan functions and checks
can be removed.
Signed-off-by: Daniel Frenzel <dgdanielf@gmail.com>
add crash detection, allow disengage via param CRASH_DETECT
improved is_flying behavior
take off, landing and hard-landing improvements
add stillness check to is_flying and log it
minimum airspeed is determined ARSPD_FBW_MIN*0.75
very strict check that all axis are not vibrating much at all
new param: INS_STILL_THRESH used to be a vibration threshold for different platforms
// @Description: Threshold to tolerate vibration to determine if vehicle is motionless. This depends on the frame type and if there is a constant vibration due to motors before launch or after landing. Total motionless is about 0.05. Suggested values: Planes/rover use 0.1, multirotors use 1, tradHeli uses 5
Param 1 of CONTINUE_AND_CHANGE_ALT now denotes which direction the
user expects the plane to travel when changing altitude:
0 = no expectation, command completes when within 5 m of altitude.
1 = climb expected, command completes at or above altitude.
2 = descent expected, command completes at or below altitude.
Param 1 denotes which direction the user expects the plane to
travel when changing altitude:
0 = no expectation, command completes when within 5 m of altitude.
1 = climb expected, command completes at or above altitude.
2 = descent expected, command completes at or below altitude.
This method checks for consistency between accelerometer readings and switches to the unit with the lowest vibration of the difference exceeds 0.3g
The threshold of 1.7 m/s/s corresponds to a maximum tilt error of 10 deg assuming one IMU is good, one is bad and the EKF is using the bad IMU.
fillycheezstake
Tom Pittenger
Andrew Tridgell
mirkix
Leonard Hall
ziltoid2
Randy Mackay
Lucas De Marchi
dgrat
squilter
Michael Day
Paul Riseborough