[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Paparazzi-devel] Simulation inconsistency
From: |
Pascal Brisset |
Subject: |
Re: [Paparazzi-devel] Simulation inconsistency |
Date: |
Thu, 01 May 2008 12:00:44 +0200 |
User-agent: |
Thunderbird 2.0.0.12 (X11/20080227) |
Hi,
Haiyang Chao wrote:
Hello, everyone.
I got a pretty strange inconsistent problem for the simulation part. I
changed "PITCH_NEUTRAL_DEFAULT" to the same value through two methods:
1 simulation gcs and 2 the MyAirframe.xml. The strange thing is that I
got different results at least for altitude tracking.
The simulator takes into account the PITCH_NEUTRAL_DEFAULT value from
the airframe file (flightModel.ml:142 and sim.ml:176 in sw/simulator) to
simulate the IR sensor. So it won't make a difference in a simulation if
you change this value in the airframe file: it will be added by the
sensor simulation and then subtracted by the autopilot code
(infrared.c:214). However, you will get something different if you
change it during the simulation (with the GCS) ...
To tune this value, you have to fly level in AUTO1, at the desired
throttle and speed and adjust the value until you don't need to touch
the pitch stick. The simulator won't help you for this ...
In the same time, you can set the NOMINAL_CRUISE_THROTTLE value
To tune the vertical loop, you will have to set THROTTLE_PGAIN,
AUTO_THROTTLE_IGAIN and PITCH_OF_VZ_PGAIN to be able to climb or descent
at a constant rate. To keep an altitude, you then have to tune the
ALTITUDE_PGAIN value.
We usually are able to get around a -5m+5m precision to fly level.
Good luck
--Pascal
We have a tiny 13 board working for real-flight waypoint tracking
recently but the altitude is not well tracked (10 or 20 meters below
the desired one in 1 minute auto2 flight time). So I used the same
configuration file for simulation and tried to locate the problem.
But I got increasing altitude in simulation all the time. I guess the
problem is from the "PITCH_NEUTRAL_DEFAULT" in IR section, which was
set as '5.72 degree' in real flight. Then I decreased this value to
0.005 (in radiance) through simulation gcs. The simulation runs ok,
the UAV stopped climbing at +30 m and maintained there. However, I got
different results while changing PITCH_NEUTRAL_DEFAULT directly in
MyAirframe.xml and recompile the code. The altitude climbing showed up
again.
Do u guys have any clue why this happened, or any good document for
parameter tuning?
I already started to dig in the simulation code, but just takes some
time to understand.
Thanks
Haiyang
------------------------------------------------------------------------
------------------------------------------------------------------------
_______________________________________________
Paparazzi-devel mailing list
address@hidden
http://lists.nongnu.org/mailman/listinfo/paparazzi-devel
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- Re: [Paparazzi-devel] Simulation inconsistency,
Pascal Brisset <=