Open TigerWuu opened 1 month ago
@TigerWuu Thanks for reporting the bug. Feel free to open a pull request for the fix
@Jaeyoung-Lim Hi, I have fixed the wind estimation problem the following is the result. Now the estimations look much more normal.
I modified two repositories to fix this problem, gz-sim and PX4-Autopilot. The current information flow will be like this
However, I have the following problem with PR, I am currently working on
cause I run into the same issue as #23705, if I use the latest version of px4 and gz sim
what should I do? should I wait until this bug fixed then apply my modification into the latest version of px4 and gz?
@TigerWuu Thanks for looking into the problem.
You would have to reapply your modification to the latest 'main' of the PX4 repository to apply your changes. Otherwise your changes would not get applied in future versions
Describe the bug
PX4 software-in-the-loop calculates the airspeed using differential pressure in SensorAirspeed.cpp in the simulation module. However, SensorAirspeed.cpp generates the fake differential pressure by body velocity, it will induce some airspeed calculating and wind estimation problems when the wind comes from the x-axis of the body frame
I think a more feasible modification is using the wind triangle concept, obtaining the ground speed from uorb topic and true wind value from gazebo plugin, and calculating the true airspeed to generate the differential pressure. What do you guys think?
To Reproduce
Expected behavior
the windspeed_east and windspeed_north value should converge to real value (0, -2)
Screenshot / Media
No response
Flight Log
No response
Software Version
Flight controller
No response
Vehicle type
Hybrid VTOL
How are the different components wired up (including port information)
No response
Additional context
No response