Re: [Paparazzi-devel] heading with AHRS_USE_GPS_HEADING but no GPS connected

[Felix Ruess]

Well, you said that the attitude/heading seems to be rather good when you have mag disabled (and no gps fix). So the integration of the gyro seems to work fine.

Also you cannot disable the gyro with this AHRS filter, as this is the "foundation" of it. The mag (and accelerometer for roll/pitch) are only used to stabilize it in the long term, as only simple integration of gyros accumulates errors and hence drifts.

You say that the heading "drifts" towards some other value... a stable value? If you rotate the imu/ap and then it first changes correctly, but then slowly moves back to the old value, it indicates a wrong "correction" of the mag.

Did you check if the mag is working correctly? E.g. plot the MAG_RAW and then the SCALED values?

See https://wiki.paparazziuav.org/wiki/ImuCalibration#Magnetometer:

If that doesn't look ok, recheck your calibration.

Also make sure that the Local_Magnetic_Field is set correctly.

On Wed, Feb 12, 2014 at 7:58 PM, alonso acuña <address@hidden> wrote:

Ok good so far. Could it be possible that my problem is related to the fusion of the mag and the gyro? What I see is that the mag gives a stable value but the heading drifts towards some other value, usually back to the value it had before.  What would be the expected behaviour if the gyro is not used for heading and only the mag is used?

On Wed, Feb 12, 2014 at 12:12 PM, Felix Ruess <address@hidden> wrote:

Alonso,

please also note that at initialization (without mag) the heading of the IMU is set to zero. The whole estimation happens in the IMU frame.

Only upon "exporting" the result via set_body_state_from_quat() is the rotation into body frame done and hence _body_ heading is -BODY_TO_IMU_PSI after startup.

Cheers, Felix

On Wed, Feb 12, 2014 at 10:56 AM, Kadir ÇİMENCİ <address@hidden> wrote:

Hi Felix,

I wasn't aware of that fusion, i thougt the raw mag/gps sensor values are determining the heading directly,thanks :) 

Kadir

2014-02-12 11:41 GMT+02:00 Felix Ruess <address@hidden>:

Hi,

Of course the gyro data is integrated to give you attitude (including heading).... that's the whole point of these AHRS filters...

If you have neither magnetometer nor can use the course over ground of the GPS for heading, you don't have any absolute heading information (wrt. gegraphic north).

While in the short term the heading would be correct relative to the initial attitude, it will drift away over time.

Cheers, Felix

On Wed, Feb 12, 2014 at 8:10 AM, Kadir ÇİMENCİ <address@hidden> wrote:

very interesting case :)  A possible source for heading, came to my mind, is the integration of vertical gyro data but i don't think paparazzi project have implemented a method like this.. i am very curious about the solution ..

Kadir

2014-02-12 6:00 GMT+02:00 alonso acuña <address@hidden>:

Yes I am using ahrs int_cmpl_quat

On Tue, Feb 11, 2014 at 4:27 PM, Felix Ruess <address@hidden> wrote:

Hi Alonso,

what IMU driver and AHRS algorithm are you using?

Assuming you are e.g. using ahrs int_cmpl_quat, did you disable the magnetometer to use GPS heading instead (via <configure name="USE_MAGNETOMETER" value="FALSE"/>)?

If not, you should get a warning that you should probably disable the mag to use gps heading, as using both could have a bad effect if they don't give you the same heading....

Cheers, Felix

On Tue, Feb 11, 2014 at 8:36 AM, alonso acuña <address@hidden> wrote:

Hello. 

I have seen several interesting things:

 - somehow the heading is accurately calculated even when the GPS is not connected. I cannot understand where the heading value is coming from. 

- when the chip is powered up the heading is set to -BODY_TO_IMU_PSI instead of 0 as documented in the file

 

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel