Merge pull request #116 from PX4/pr-covReset

after a height reset zero the cross covariances for height and vertical
This commit is contained in:
Paul Riseborough
2016-05-01 21:44:38 +10:00
+37 -2
View File
@@ -103,9 +103,17 @@ void Ekf::resetHeight()
if (_time_last_imu - range_newest.time_us < 2 * RNG_MAX_INTERVAL) {
// calculate the new vertical position using range sensor
float new_pos_down = _hgt_sensor_offset - range_newest.rng;
// update the state and assoicated variance
_state.pos(2) = new_pos_down;
// reset the associated covariance values
zeroRows(P, 8, 8);
zeroCols(P, 8, 8);
// the state variance is the same as the observation
P[8][8] = sq(_params.range_noise);
vert_pos_reset = true;
} else {
@@ -122,7 +130,14 @@ void Ekf::resetHeight()
if (_time_last_imu - baro_newest.time_us < 2 * BARO_MAX_INTERVAL) {
_state.pos(2) = _hgt_sensor_offset - baro_newest.hgt + _baro_hgt_offset;
// reset the associated covariance values
zeroRows(P, 8, 8);
zeroCols(P, 8, 8);
// the state variance is th esame as the observation
P[8][8] = sq(_params.baro_noise);
vert_pos_reset = true;
} else {
@@ -133,7 +148,14 @@ void Ekf::resetHeight()
// initialize vertical position and velocity with newest gps measurement
if (_time_last_imu - gps_newest.time_us < 2 * GPS_MAX_INTERVAL) {
_state.pos(2) = _hgt_sensor_offset - gps_newest.hgt + _gps_alt_ref;
// reset the associated covarince values
zeroRows(P, 8, 8);
zeroCols(P, 8, 8);
// the state variance is the same as the observation
P[8][8] = sq(gps_newest.hacc);
vert_pos_reset = true;
} else {
@@ -145,13 +167,26 @@ void Ekf::resetHeight()
_baro_hgt_offset = baro_newest.hgt + _state.pos(2);
}
// If we are using GPS, then use it to reset the vertical velocity
// reset the vertical velocity covariance values
zeroRows(P, 5, 5);
zeroCols(P, 5, 5);
// reset the vertical velocity state
if (_control_status.flags.gps && (_time_last_imu - gps_newest.time_us < 2 * GPS_MAX_INTERVAL)) {
// If we are using GPS, then use it to reset the vertical velocity
_state.vel(2) = gps_newest.vel(2);
// the state variance is the same as the observation
P[5][5] = sq(1.5f * gps_newest.sacc);
} else {
P[5][5] = fminf(sq(_state.vel(2)),1000.0f);
// we don't know what the vertical velocity is, so set it to zero
_state.vel(2) = 0.0f;
// Set the variance to a value large enough to allow the state to converge quickly
// that does not destabilise the filter
P[5][5] = fminf(sq(_state.vel(2)),100.0f);
}
vert_vel_reset = true;