AP_GPS: fixed antenna offset for blending #13802
Conversation
|
I agree with this change and can confirm that it resolves the issues I was having with Copter-4.0.4-dev when using GPS blending. |
libraries/AP_GPS/AP_GPS.cpp
Outdated
| @@ -795,8 +795,10 @@ void AP_GPS::update(void) | |||
|
|
|||
| #if defined(GPS_BLENDED_INSTANCE) | |||
| // copy the primary instance to the blended instance in case it is enabled later | |||
| state[GPS_BLENDED_INSTANCE] = state[primary_instance]; | |||
| _blended_antenna_offset = _antenna_offset[primary_instance]; | |||
| if (primary_instance != GPS_BLENDED_INSTANCE) { | |||
There was a problem hiding this comment.
JUST NUKE THIS ALREADY! Sorry to be a bear on it but I've pointed this out in review before: #12844 (comment) It's dead code that doesn't make sense. If we had removed it when first brought up we never would have had the problem...
There was a problem hiding this comment.
I may have pushed for the minimum change. If we're sure it can just be nuked then I'm OK with that and I'm happy to re-test afterwards..
| @@ -1654,14 +1656,6 @@ void AP_GPS::calc_blended_state(void) | |||
| _hgt_offset_cm[i] = (float)(state[GPS_BLENDED_INSTANCE].location.alt - state[i].location.alt) * alpha[i] + _hgt_offset_cm[i] * (1.0f - alpha[i]); | |||
| } | |||
|
|
|||
| // Calculate a corrected location for each GPS | |||
There was a problem hiding this comment.
If we are going to remove this then the calculation of _NE_pos_offset_m and _hgt_off_cm above should also be removed. Also worth observing they should have been locals here, rather then members of the GPS object. (Given that we track _blended_antenna_offset directly I think removing this is fine, but lets remove all the dead bits of code.
There was a problem hiding this comment.
done, note that this ended up removing quite a bit of code
we were accessing beyond the end of the antenna_offset array
|
I guess the next step for this is for me to test fly it right? I'll try and do that tomorrow. |
|
@priseborough, just FYI |
|
I flight tested this and it worked great, thanks! |
we were accessing beyond the end of the antenna_offset array. This could cause the EKF position and velocity data to be corrupted when blending
Bug introduced in
261465e AP_GPS: split out update_primary() from update()