AP_NavEKF3: move badMagYaw onto stack

This commit is contained in:
Peter Barker authored and Peter Barker committed 2020-12-01 10:40:09 +11:00
1 parent addd57dcab
commit 0529ddcd67
3 files changed
+1 -3

No files matched your search

@@ -145,7 +145,7 @@ void NavEKF3_core::realignYawGPS()
float yawErr = MAX(fabsf(wrap_PI(gpsYaw - velYaw)),fabsf(wrap_PI(gpsYaw - eulerAngles.z)));
// If the angles disagree by more than 45 degrees and GPS innovations are large or no previous yaw alignment, we declare the magnetic yaw as bad
badMagYaw = ((yawErr > 0.7854f) && (velTestRatio > 1.0f) && (PV_AidingMode == AID_ABSOLUTE)) || !yawAlignComplete;
bool badMagYaw = ((yawErr > 0.7854f) && (velTestRatio > 1.0f) && (PV_AidingMode == AID_ABSOLUTE)) || !yawAlignComplete;
// correct yaw angle using GPS ground course if compass yaw bad
if (badMagYaw) {
-1
View File
@@ -429,7 +429,6 @@ void NavEKF3_core::InitialiseVariablesMag()
magYawResetTimer_ms = imuSampleTime_ms;
magTimeout = false;
allMagSensorsFailed = false;
badMagYaw = false;
finalInflightMagInit = false;
mag_state.q0 = 1;
mag_state.DCM.identity();
-1
View File
@@ -943,7 +943,6 @@ private:
bool hgtTimeout; // boolean true if height measurements have failed innovation consistency check and timed out
bool magTimeout; // boolean true if magnetometer measurements have failed for too long and have timed out
bool tasTimeout; // boolean true if true airspeed measurements have failed for too long and have timed out
bool badMagYaw; // boolean true if the magnetometer is declared to be producing bad data
bool badIMUdata; // boolean true if the bad IMU data is detected
float gpsNoiseScaler; // Used to scale the GPS measurement noise and consistency gates to compensate for operation with small satellite counts