px4-firmware/EKF/control.cpp

1263 lines
48 KiB
C++

/****************************************************************************
*
* Copyright (c) 2015 Estimation and Control Library (ECL). All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
* 3. Neither the name ECL nor the names of its contributors may be
* used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
* FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
* COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
* BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS
* OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
* AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
* ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*
****************************************************************************/
/**
* @file control.cpp
* Control functions for ekf attitude and position estimator.
*
* @author Paul Riseborough <p_riseborough@live.com.au>
*
*/
#include "../ecl.h"
#include "ekf.h"
#include "mathlib.h"
void Ekf::controlFusionModes()
{
// Store the status to enable change detection
_control_status_prev.value = _control_status.value;
// Get the magnetic declination
calcMagDeclination();
// monitor the tilt alignment
if (!_control_status.flags.tilt_align) {
// whilst we are aligning the tilt, monitor the variances
Vector3f angle_err_var_vec = calcRotVecVariances();
// Once the tilt variances have reduced to equivalent of 3deg uncertainty, re-set the yaw and magnetic field states
// and declare the tilt alignment complete
if ((angle_err_var_vec(0) + angle_err_var_vec(1)) < sq(0.05235f)) {
_control_status.flags.tilt_align = true;
_control_status.flags.yaw_align = resetMagHeading(_mag_sample_delayed.mag);
// send alignment status message to the console
if (_control_status.flags.baro_hgt) {
ECL_INFO("EKF aligned, (pressure height, IMU buf: %i, OBS buf: %i)",(int)_imu_buffer_length,(int)_obs_buffer_length);
} else if (_control_status.flags.ev_hgt) {
ECL_INFO("EKF aligned, (EV height, IMU buf: %i, OBS buf: %i)",(int)_imu_buffer_length,(int)_obs_buffer_length);
} else if (_control_status.flags.gps_hgt) {
ECL_INFO("EKF aligned, (GPS height, IMU buf: %i, OBS buf: %i)",(int)_imu_buffer_length,(int)_obs_buffer_length);
} else if (_control_status.flags.rng_hgt) {
ECL_INFO("EKF aligned, (range height, IMU buf: %i, OBS buf: %i)",(int)_imu_buffer_length,(int)_obs_buffer_length);
} else {
ECL_ERR("EKF aligned, (unknown height, IMU buf: %i, OBS buf: %i)",(int)_imu_buffer_length,(int)_obs_buffer_length);
}
}
}
// check faultiness (before pop_first_older_than) to see if we can change back to original height sensor
baroSample baro_init = _baro_buffer.get_newest();
_baro_hgt_faulty = !((_time_last_imu - baro_init.time_us) < 2 * BARO_MAX_INTERVAL);
gpsSample gps_init = _gps_buffer.get_newest();
_gps_hgt_faulty = !((_time_last_imu - gps_init.time_us) < 2 * GPS_MAX_INTERVAL);
rangeSample rng_init = _range_buffer.get_newest();
_rng_hgt_faulty = !((_time_last_imu - rng_init.time_us) < 2 * RNG_MAX_INTERVAL);
// check for arrival of new sensor data at the fusion time horizon
_gps_data_ready = _gps_buffer.pop_first_older_than(_imu_sample_delayed.time_us, &_gps_sample_delayed);
_mag_data_ready = _mag_buffer.pop_first_older_than(_imu_sample_delayed.time_us, &_mag_sample_delayed);
_delta_time_baro_us = _baro_sample_delayed.time_us;
_baro_data_ready = _baro_buffer.pop_first_older_than(_imu_sample_delayed.time_us, &_baro_sample_delayed);
// if we have a new baro sample save the delta time between this sample and the last sample which is
// used below for baro offset calculations
if (_baro_data_ready) {
_delta_time_baro_us = _baro_sample_delayed.time_us - _delta_time_baro_us;
}
// calculate 2,2 element of rotation matrix from sensor frame to earth frame
_R_rng_to_earth_2_2 = _R_to_earth(2, 0) * _sin_tilt_rng + _R_to_earth(2, 2) * _cos_tilt_rng;
_range_data_ready = _range_buffer.pop_first_older_than(_imu_sample_delayed.time_us, &_range_sample_delayed)
&& (_R_rng_to_earth_2_2 > 0.7071f);
checkForStuckRange();
_flow_data_ready = _flow_buffer.pop_first_older_than(_imu_sample_delayed.time_us, &_flow_sample_delayed)
&& (_R_to_earth(2, 2) > 0.7071f);
_ev_data_ready = _ext_vision_buffer.pop_first_older_than(_imu_sample_delayed.time_us, &_ev_sample_delayed);
_tas_data_ready = _airspeed_buffer.pop_first_older_than(_imu_sample_delayed.time_us, &_airspeed_sample_delayed);
// check for height sensor timeouts and reset and change sensor if necessary
controlHeightSensorTimeouts();
// control use of observations for aiding
controlMagFusion();
controlOpticalFlowFusion();
controlGpsFusion();
controlAirDataFusion();
controlBetaFusion();
controlDragFusion();
controlHeightFusion();
// For efficiency, fusion of direct state observations for position and velocity is performed sequentially
// in a single function using sensor data from multiple sources (GPS, baro, range finder, etc)
controlVelPosFusion();
// Additional data from an external vision sensor can also be fused.
controlExternalVisionFusion();
// report dead reckoning if we are no longer fusing measurements that constrain velocity drift
_is_dead_reckoning = (_time_last_imu - _time_last_pos_fuse > _params.no_aid_timeout_max)
&& (_time_last_imu - _time_last_vel_fuse > _params.no_aid_timeout_max)
&& (_time_last_imu - _time_last_of_fuse > _params.no_aid_timeout_max);
}
void Ekf::controlExternalVisionFusion()
{
// Check for new exernal vision data
if (_ev_data_ready) {
// external vision position aiding selection logic
if ((_params.fusion_mode & MASK_USE_EVPOS) && !_control_status.flags.ev_pos && _control_status.flags.tilt_align && _control_status.flags.yaw_align) {
// check for a exernal vision measurement that has fallen behind the fusion time horizon
if (_time_last_imu - _time_last_ext_vision < 2 * EV_MAX_INTERVAL) {
// turn on use of external vision measurements for position
_control_status.flags.ev_pos = true;
ECL_INFO("EKF commencing external vision position fusion");
// reset the position if we are not already aiding using GPS, else use a relative position
// method for fusing the position data
if (_control_status.flags.gps) {
_hpos_odometry = true;
} else {
resetPosition();
resetVelocity();
_hpos_odometry = false;
}
}
}
// external vision yaw aiding selection logic
if ((_params.fusion_mode & MASK_USE_EVYAW) && !_control_status.flags.ev_yaw && _control_status.flags.tilt_align) {
// don't start using EV data unless daa is arriving frequently
if (_time_last_imu - _time_last_ext_vision < 2 * EV_MAX_INTERVAL) {
// reset the yaw angle to the value from the observaton quaternion
// get the roll, pitch, yaw estimates from the quaternion states
Quatf q_init(_state.quat_nominal);
Eulerf euler_init(q_init);
// get initial yaw from the observation quaternion
extVisionSample ev_newest = _ext_vision_buffer.get_newest();
Quatf q_obs(ev_newest.quat);
Eulerf euler_obs(q_obs);
euler_init(2) = euler_obs(2);
// save a copy of the quaternion state for later use in calculating the amount of reset change
Quatf quat_before_reset = _state.quat_nominal;
// calculate initial quaternion states for the ekf
_state.quat_nominal = Quatf(euler_init);
// calculate the amount that the quaternion has changed by
_state_reset_status.quat_change = _state.quat_nominal * quat_before_reset.inversed();
// add the reset amount to the output observer buffered data
outputSample output_states;
unsigned output_length = _output_buffer.get_length();
for (unsigned i=0; i < output_length; i++) {
output_states = _output_buffer.get_from_index(i);
output_states.quat_nominal *= _state_reset_status.quat_change;
_output_buffer.push_to_index(i, output_states);
}
// apply the change in attitude quaternion to our newest quaternion estimate
// which was already taken out from the output buffer
_output_new.quat_nominal *= _state_reset_status.quat_change;
// capture the reset event
_state_reset_status.quat_counter++;
// flag the yaw as aligned
_control_status.flags.yaw_align = true;
// turn on fusion of external vision yaw measurements and disable all magnetoemter fusion
_control_status.flags.ev_yaw = true;
_control_status.flags.mag_hdg = false;
_control_status.flags.mag_3D = false;
_control_status.flags.mag_dec = false;
ECL_INFO("EKF commencing external vision yaw fusion");
}
}
// determine if we should start using the height observations
if (_params.vdist_sensor_type == VDIST_SENSOR_EV) {
// don't start using EV data unless daa is arriving frequently
if (!_control_status.flags.ev_hgt && (_time_last_imu - _time_last_ext_vision < 2 * EV_MAX_INTERVAL)) {
setControlEVHeight();
resetHeight();
}
}
// determine if we should use the vertical position observation
if (_control_status.flags.ev_hgt) {
_fuse_height = true;
}
// determine if we should use the horizontal position observations
if (_control_status.flags.ev_pos) {
_fuse_pos = true;
// correct position and height for offset relative to IMU
Vector3f pos_offset_body = _params.ev_pos_body - _params.imu_pos_body;
Vector3f pos_offset_earth = _R_to_earth * pos_offset_body;
_ev_sample_delayed.posNED(0) -= pos_offset_earth(0);
_ev_sample_delayed.posNED(1) -= pos_offset_earth(1);
_ev_sample_delayed.posNED(2) -= pos_offset_earth(2);
// if GPS data is being used, then use an incremental position fusion method for EV data
if (_control_status.flags.gps) {
_hpos_odometry = true;
}
}
// Fuse available NED position data into the main filter
if (_fuse_height || _fuse_pos) {
fuseVelPosHeight();
_fuse_pos = _fuse_height = false;
}
// determine if we should use the yaw observation
if (_control_status.flags.ev_yaw) {
fuseHeading();
}
}
// handle the case when we are relying on ev data and lose it
if (_control_status.flags.ev_pos && !_control_status.flags.gps && !_control_status.flags.opt_flow) {
// We are relying on ev aiding to constrain drift so after 5s without aiding we need to do something
if ((_time_last_imu - _time_last_pos_fuse > 5e6)) {
// Switch to the non-aiding mode, zero the velocity states
// and set the synthetic position to the current estimate
_control_status.flags.ev_pos = false;
_last_known_posNE(0) = _state.pos(0);
_last_known_posNE(1) = _state.pos(1);
_state.vel.setZero();
}
}
}
void Ekf::controlOpticalFlowFusion()
{
// Check for new optical flow data that has fallen behind the fusion time horizon
if (_flow_data_ready) {
// optical flow fusion mode selection logic
if ((_params.fusion_mode & MASK_USE_OF) // optical flow has been selected by the user
&& !_control_status.flags.opt_flow // we are not yet using flow data
&& _control_status.flags.tilt_align // we know our tilt attitude
&& (_time_last_imu - _time_last_hagl_fuse) < 5e5) // we have a valid distance to ground estimate
{
// If the heading is not aligned, reset the yaw and magnetic field states
if (!_control_status.flags.yaw_align) {
_control_status.flags.yaw_align = resetMagHeading(_mag_sample_delayed.mag);
}
// If the heading is valid, start using optical flow aiding
if (_control_status.flags.yaw_align) {
// set the flag and reset the fusion timeout
_control_status.flags.opt_flow = true;
_time_last_of_fuse = _time_last_imu;
// if we are not using GPS then the velocity and position states and covariances need to be set
if (!_control_status.flags.gps) {
// constrain height above ground to be above minimum possible
float heightAboveGndEst = fmaxf((_terrain_vpos - _state.pos(2)), _params.rng_gnd_clearance);
// calculate absolute distance from focal point to centre of frame assuming a flat earth
float range = heightAboveGndEst / _R_rng_to_earth_2_2;
if ((range - _params.rng_gnd_clearance) > 0.3f && _flow_sample_delayed.dt > 0.05f) {
// we should have reliable OF measurements so
// calculate X and Y body relative velocities from OF measurements
Vector3f vel_optflow_body;
vel_optflow_body(0) = - range * _flow_sample_delayed.flowRadXYcomp(1) / _flow_sample_delayed.dt;
vel_optflow_body(1) = range * _flow_sample_delayed.flowRadXYcomp(0) / _flow_sample_delayed.dt;
vel_optflow_body(2) = 0.0f;
// rotate from body to earth frame
Vector3f vel_optflow_earth;
vel_optflow_earth = _R_to_earth * vel_optflow_body;
// take x and Y components
_state.vel(0) = vel_optflow_earth(0);
_state.vel(1) = vel_optflow_earth(1);
} else {
_state.vel(0) = 0.0f;
_state.vel(1) = 0.0f;
}
// reset the velocity covariance terms
zeroRows(P,4,5);
zeroCols(P,4,5);
// reset the horizontal velocity variance using the optical flow noise variance
P[5][5] = P[4][4] = sq(range) * calcOptFlowMeasVar();
if (!_control_status.flags.in_air) {
// we are likely starting OF for the first time so reset the horizontal position and vertical velocity states
_state.pos(0) = 0.0f;
_state.pos(1) = 0.0f;
} else {
// set to the last known position
_state.pos(0) = _last_known_posNE(0);
_state.pos(1) = _last_known_posNE(1);
}
// reset the corresponding covariances
// we are by definition at the origin at commencement so variances are also zeroed
zeroRows(P,7,8);
zeroCols(P,7,8);
// align the output observer to the EKF states
alignOutputFilter();
}
}
} else if (!(_params.fusion_mode & MASK_USE_OF)) {
_control_status.flags.opt_flow = false;
}
// handle the case when we are relying on optical flow fusion and lose it
if (_control_status.flags.opt_flow && !_control_status.flags.gps && !_control_status.flags.ev_pos) {
// We are relying on flow aiding to constrain attitude drift so after 5s without aiding we need to do something
if ((_time_last_imu - _time_last_of_fuse > 5e6)) {
// Switch to the non-aiding mode, zero the velocity states
// and set the synthetic position to the current estimate
_control_status.flags.opt_flow = false;
_last_known_posNE(0) = _state.pos(0);
_last_known_posNE(1) = _state.pos(1);
_state.vel.setZero();
}
}
// fuse the data
if (_control_status.flags.opt_flow) {
// Update optical flow bias estimates
calcOptFlowBias();
// Fuse optical flow LOS rate observations into the main filter
fuseOptFlow();
_last_known_posNE(0) = _state.pos(0);
_last_known_posNE(1) = _state.pos(1);
}
}
}
void Ekf::controlGpsFusion()
{
// Check for new GPS data that has fallen behind the fusion time horizon
if (_gps_data_ready) {
// Determine if we should use GPS aiding for velocity and horizontal position
// To start using GPS we need angular alignment completed, the local NED origin set and GPS data that has not failed checks recently
if ((_params.fusion_mode & MASK_USE_GPS) && !_control_status.flags.gps) {
if (_control_status.flags.tilt_align && _NED_origin_initialised && (_time_last_imu - _last_gps_fail_us > 5e6)) {
// If the heading is not aligned, reset the yaw and magnetic field states
if (!_control_status.flags.yaw_align) {
_control_status.flags.yaw_align = resetMagHeading(_mag_sample_delayed.mag);
}
// If the heading is valid start using gps aiding
if (_control_status.flags.yaw_align) {
// if we are not already aiding with optical flow, then we need to reset the position and velocity
// otherwise we only need to reset the position
_control_status.flags.gps = true;
if (!_control_status.flags.opt_flow) {
if (!resetPosition() || !resetVelocity()) {
_control_status.flags.gps = false;
}
} else if (!resetPosition()) {
_control_status.flags.gps = false;
}
if (_control_status.flags.gps) {
ECL_INFO("EKF commencing GPS fusion");
_time_last_gps = _time_last_imu;
}
}
}
} else if (!(_params.fusion_mode & MASK_USE_GPS)) {
_control_status.flags.gps = false;
}
// handle the case when we now have GPS, but have not been using it for an extended period
if (_control_status.flags.gps && !_control_status.flags.opt_flow) {
// We are relying on GPS aiding to constrain attitude drift so after 7 seconds without aiding we need to do something
bool do_reset = (_time_last_imu - _time_last_pos_fuse > _params.no_gps_timeout_max) && (_time_last_imu - _time_last_vel_fuse > _params.no_gps_timeout_max);
// Our position measurments have been rejected for more than 14 seconds
do_reset |= _time_last_imu - _time_last_pos_fuse > 2 * _params.no_gps_timeout_max;
if (do_reset) {
// Reset states to the last GPS measurement
if (_control_status.flags.fixed_wing) {
// if flying a fixed wing aircraft, do a complete reset that includes yaw, velocity and position
realignYawGPS();
} else {
resetVelocity();
resetPosition();
}
ECL_WARN("EKF GPS fusion timeout - reset to GPS");
// Reset the timeout counters
_time_last_pos_fuse = _time_last_imu;
_time_last_vel_fuse = _time_last_imu;
}
}
// Only use GPS data for position and velocity aiding if enabled
if (_control_status.flags.gps) {
_fuse_pos = true;
_fuse_vert_vel = true;
_fuse_hor_vel = true;
// correct velocity for offset relative to IMU
Vector3f ang_rate = _imu_sample_delayed.delta_ang * (1.0f/_imu_sample_delayed.delta_ang_dt);
Vector3f pos_offset_body = _params.gps_pos_body - _params.imu_pos_body;
Vector3f vel_offset_body = cross_product(ang_rate, pos_offset_body);
Vector3f vel_offset_earth = _R_to_earth * vel_offset_body;
_gps_sample_delayed.vel -= vel_offset_earth;
// correct position and height for offset relative to IMU
Vector3f pos_offset_earth = _R_to_earth * pos_offset_body;
_gps_sample_delayed.pos(0) -= pos_offset_earth(0);
_gps_sample_delayed.pos(1) -= pos_offset_earth(1);
_gps_sample_delayed.hgt += pos_offset_earth(2);
}
} else {
// handle the case where we do not have GPS and have not been using it for an extended period, but are still relying on it
if ((_time_last_imu - _time_last_gps > 10e6) && (_time_last_imu - _time_last_airspeed > 1e6) && (_time_last_imu - _time_last_optflow > 1e6) && _control_status.flags.gps) {
// if we don't have a source of aiding to constrain attitude drift,
// then we need to switch to the non-aiding mode, zero the velocity states
// and set the synthetic GPS position to the current estimate
_control_status.flags.gps = false;
_last_known_posNE(0) = _state.pos(0);
_last_known_posNE(1) = _state.pos(1);
_state.vel.setZero();
ECL_WARN("EKF measurement timeout - stopping navigation");
}
}
}
void Ekf::controlHeightSensorTimeouts()
{
/*
* Handle the case where we have not fused height measurements recently and
* uncertainty exceeds the max allowable. Reset using the best available height
* measurement source, continue using it after the reset and declare the current
* source failed if we have switched.
*/
// Check for IMU accelerometer vibration induced clipping as evidenced by the vertical innovations being positive and not stale.
// Clipping causes the average accel reading to move towards zero which makes the INS think it is falling and produces positive vertical innovations
float var_product_lim = sq(_params.vert_innov_test_lim) * sq(_params.vert_innov_test_lim);
bool bad_vert_accel = (_control_status.flags.baro_hgt && // we can only run this check if vertical position and velocity observations are indepedant
(sq(_vel_pos_innov[5] * _vel_pos_innov[2]) > var_product_lim * (_vel_pos_innov_var[5] * _vel_pos_innov_var[2])) && // vertical position and velocity sensors are in agreement that we have a significant error
(_vel_pos_innov[2] > 0.0f) && // positive innovation indicates that the inertial nav thinks it is falling
((_imu_sample_delayed.time_us - _baro_sample_delayed.time_us) < 2 * BARO_MAX_INTERVAL) && // vertical position data is fresh
((_imu_sample_delayed.time_us - _gps_sample_delayed.time_us) < 2 * GPS_MAX_INTERVAL)); // vertical velocity data is fresh
// record time of last bad vert accel
if (bad_vert_accel) {
_time_bad_vert_accel = _time_last_imu;
} else {
_time_good_vert_accel = _time_last_imu;
}
// declare a bad vertical acceleration measurement and make the declaration persist
// for a minimum of 10 seconds
if (_bad_vert_accel_detected) {
_bad_vert_accel_detected = (_time_last_imu - _time_bad_vert_accel < BADACC_PROBATION);
} else {
_bad_vert_accel_detected = bad_vert_accel;
}
// check if height is continuously failing becasue of accel errors
bool continuous_bad_accel_hgt = ((_time_last_imu - _time_good_vert_accel) > (unsigned)_params.bad_acc_reset_delay_us);
// check if height has been inertial deadreckoning for too long
bool hgt_fusion_timeout = ((_time_last_imu - _time_last_hgt_fuse) > 5e6);
// reset the vertical position and velocity states
if ((P[9][9] > sq(_params.hgt_reset_lim)) && (hgt_fusion_timeout || continuous_bad_accel_hgt)) {
// boolean that indicates we will do a height reset
bool reset_height = false;
// handle the case where we are using baro for height
if (_control_status.flags.baro_hgt) {
// check if GPS height is available
gpsSample gps_init = _gps_buffer.get_newest();
bool gps_hgt_available = ((_time_last_imu - gps_init.time_us) < 2 * GPS_MAX_INTERVAL);
bool gps_hgt_accurate = (gps_init.vacc < _params.req_vacc);
baroSample baro_init = _baro_buffer.get_newest();
bool baro_hgt_available = ((_time_last_imu - baro_init.time_us) < 2 * BARO_MAX_INTERVAL);
// check for inertial sensing errors in the last 10 seconds
bool prev_bad_vert_accel = (_time_last_imu - _time_bad_vert_accel < BADACC_PROBATION);
// reset to GPS if adequate GPS data is available and the timeout cannot be blamed on IMU data
bool reset_to_gps = gps_hgt_available && gps_hgt_accurate && !_gps_hgt_faulty && !prev_bad_vert_accel;
// reset to GPS if GPS data is available and there is no Baro data
reset_to_gps = reset_to_gps || (gps_hgt_available && !baro_hgt_available);
// reset to Baro if we are not doing a GPS reset and baro data is available
bool reset_to_baro = !reset_to_gps && baro_hgt_available;
if (reset_to_gps) {
// set height sensor health
_baro_hgt_faulty = true;
// declare the GPS height healthy
_gps_hgt_faulty = false;
// reset the height mode
setControlGPSHeight();
// request a reset
reset_height = true;
ECL_WARN("EKF baro hgt timeout - reset to GPS");
} else if (reset_to_baro) {
// set height sensor health
_baro_hgt_faulty = false;
// reset the height mode
setControlBaroHeight();
// request a reset
reset_height = true;
ECL_WARN("EKF baro hgt timeout - reset to baro");
} else {
// we have nothing we can reset to
// deny a reset
reset_height = false;
}
}
// handle the case we are using GPS for height
if (_control_status.flags.gps_hgt) {
// check if GPS height is available
gpsSample gps_init = _gps_buffer.get_newest();
bool gps_hgt_available = ((_time_last_imu - gps_init.time_us) < 2 * GPS_MAX_INTERVAL);
bool gps_hgt_accurate = (gps_init.vacc < _params.req_vacc);
// check the baro height source for consistency and freshness
baroSample baro_init = _baro_buffer.get_newest();
bool baro_data_fresh = ((_time_last_imu - baro_init.time_us) < 2 * BARO_MAX_INTERVAL);
float baro_innov = _state.pos(2) - (_hgt_sensor_offset - baro_init.hgt + _baro_hgt_offset);
bool baro_data_consistent = fabsf(baro_innov) < (sq(_params.baro_noise) + P[8][8]) * sq(_params.baro_innov_gate);
// if baro data is acceptable and GPS data is inaccurate, reset height to baro
bool reset_to_baro = baro_data_consistent && baro_data_fresh && !_baro_hgt_faulty && !gps_hgt_accurate;
// if GPS height is unavailable and baro data is available, reset height to baro
reset_to_baro = reset_to_baro || (!gps_hgt_available && baro_data_fresh);
// if we cannot switch to baro and GPS data is available, reset height to GPS
bool reset_to_gps = !reset_to_baro && gps_hgt_available;
if (reset_to_baro) {
// set height sensor health
_gps_hgt_faulty = true;
_baro_hgt_faulty = false;
// reset the height mode
setControlBaroHeight();
// request a reset
reset_height = true;
ECL_WARN("EKF gps hgt timeout - reset to baro");
} else if (reset_to_gps) {
// set height sensor health
_gps_hgt_faulty = false;
// reset the height mode
setControlGPSHeight();
// request a reset
reset_height = true;
ECL_WARN("EKF gps hgt timeout - reset to GPS");
} else {
// we have nothing to reset to
reset_height = false;
}
}
// handle the case we are using range finder for height
if (_control_status.flags.rng_hgt) {
// check if range finder data is available
rangeSample rng_init = _range_buffer.get_newest();
bool rng_data_available = ((_time_last_imu - rng_init.time_us) < 2 * RNG_MAX_INTERVAL);
// check if baro data is available
baroSample baro_init = _baro_buffer.get_newest();
bool baro_data_available = ((_time_last_imu - baro_init.time_us) < 2 * BARO_MAX_INTERVAL);
// reset to baro if we have no range data and baro data is available
bool reset_to_baro = !rng_data_available && baro_data_available;
// reset to range data if it is available
bool reset_to_rng = rng_data_available;
if (reset_to_baro) {
// set height sensor health
_rng_hgt_faulty = true;
_baro_hgt_faulty = false;
// reset the height mode
setControlBaroHeight();
// request a reset
reset_height = true;
ECL_WARN("EKF rng hgt timeout - reset to baro");
} else if (reset_to_rng) {
// set height sensor health
_rng_hgt_faulty = false;
// reset the height mode
setControlRangeHeight();
// request a reset
reset_height = true;
ECL_WARN("EKF rng hgt timeout - reset to rng hgt");
} else {
// we have nothing to reset to
reset_height = false;
}
}
// handle the case where we are using external vision data for height
if (_control_status.flags.ev_hgt) {
// check if vision data is available
extVisionSample ev_init = _ext_vision_buffer.get_newest();
bool ev_data_available = ((_time_last_imu - ev_init.time_us) < 2 * EV_MAX_INTERVAL);
// check if baro data is available
baroSample baro_init = _baro_buffer.get_newest();
bool baro_data_available = ((_time_last_imu - baro_init.time_us) < 2 * BARO_MAX_INTERVAL);
// reset to baro if we have no vision data and baro data is available
bool reset_to_baro = !ev_data_available && baro_data_available;
// reset to ev data if it is available
bool reset_to_ev = ev_data_available;
if (reset_to_baro) {
// set height sensor health
_baro_hgt_faulty = false;
// reset the height mode
setControlBaroHeight();
// request a reset
reset_height = true;
ECL_WARN("EKF ev hgt timeout - reset to baro");
} else if (reset_to_ev) {
// reset the height mode
setControlEVHeight();
// request a reset
reset_height = true;
ECL_WARN("EKF ev hgt timeout - reset to ev hgt");
} else {
// we have nothing to reset to
reset_height = false;
}
}
// Reset vertical position and velocity states to the last measurement
if (reset_height) {
resetHeight();
// Reset the timout timer
_time_last_hgt_fuse = _time_last_imu;
}
}
}
void Ekf::controlHeightFusion()
{
// set control flags for the desired primary height source
if (_range_data_ready) {
// correct the range data for position offset relative to the IMU
Vector3f pos_offset_body = _params.rng_pos_body - _params.imu_pos_body;
Vector3f pos_offset_earth = _R_to_earth * pos_offset_body;
_range_sample_delayed.rng += pos_offset_earth(2) / _R_rng_to_earth_2_2;
}
if (_params.vdist_sensor_type == VDIST_SENSOR_BARO) {
_in_range_aid_mode = rangeAidConditionsMet(_in_range_aid_mode);
if (_in_range_aid_mode && _range_data_ready && !_rng_hgt_faulty) {
setControlRangeHeight();
_fuse_height = true;
// we have just switched to using range finder, calculate height sensor offset such that current
// measurment matches our current height estimate
if (_control_status_prev.flags.rng_hgt != _control_status.flags.rng_hgt) {
if (get_terrain_valid()) {
_hgt_sensor_offset = _terrain_vpos;
} else {
_hgt_sensor_offset = _R_rng_to_earth_2_2 * _range_sample_delayed.rng + _state.pos(2);
}
}
} else if (_baro_data_ready && !_baro_hgt_faulty &&
!(_in_range_aid_mode && !_range_data_ready && !_rng_hgt_faulty)) {
setControlBaroHeight();
_fuse_height = true;
_in_range_aid_mode = false;
// we have just switched to using baro height, we don't need to set a height sensor offset
// since we track a separate _baro_hgt_offset
if (_control_status_prev.flags.baro_hgt != _control_status.flags.baro_hgt) {
_hgt_sensor_offset = 0.0f;
}
} else if (_control_status.flags.gps_hgt && _gps_data_ready && !_gps_hgt_faulty) {
// switch to gps if there was a reset to gps
_fuse_height = true;
_in_range_aid_mode = false;
// we have just switched to using gps height, calculate height sensor offset such that current
// measurment matches our current height estimate
if (_control_status_prev.flags.gps_hgt != _control_status.flags.gps_hgt) {
_hgt_sensor_offset = _gps_sample_delayed.hgt - _gps_alt_ref + _state.pos(2);
}
}
}
// set the height data source to range if requested
if ((_params.vdist_sensor_type == VDIST_SENSOR_RANGE) && !_rng_hgt_faulty) {
setControlRangeHeight();
_fuse_height = _range_data_ready;
// we have just switched to using range finder, calculate height sensor offset such that current
// measurment matches our current height estimate
if (_control_status_prev.flags.rng_hgt != _control_status.flags.rng_hgt) {
// use the parameter rng_gnd_clearance if on ground to avoid a noisy offset initialization (e.g. sonar)
if (_control_status.flags.in_air && get_terrain_valid()) {
_hgt_sensor_offset = _terrain_vpos;
} else if (_control_status.flags.in_air) {
_hgt_sensor_offset = _R_rng_to_earth_2_2 * _range_sample_delayed.rng + _state.pos(2);
} else {
_hgt_sensor_offset = _params.rng_gnd_clearance;
}
}
} else if ((_params.vdist_sensor_type == VDIST_SENSOR_RANGE) && _baro_data_ready && !_baro_hgt_faulty) {
setControlBaroHeight();
_fuse_height = true;
// we have just switched to using baro height, we don't need to set a height sensor offset
// since we track a separate _baro_hgt_offset
if (_control_status_prev.flags.baro_hgt != _control_status.flags.baro_hgt) {
_hgt_sensor_offset = 0.0f;
}
}
// Determine if GPS should be used as the height source
if (_params.vdist_sensor_type == VDIST_SENSOR_GPS) {
_in_range_aid_mode = rangeAidConditionsMet(_in_range_aid_mode);
if (_in_range_aid_mode && _range_data_ready && !_rng_hgt_faulty) {
setControlRangeHeight();
_fuse_height = true;
// we have just switched to using range finder, calculate height sensor offset such that current
// measurment matches our current height estimate
if (_control_status_prev.flags.rng_hgt != _control_status.flags.rng_hgt) {
if (get_terrain_valid()) {
_hgt_sensor_offset = _terrain_vpos;
} else {
_hgt_sensor_offset = _R_rng_to_earth_2_2 * _range_sample_delayed.rng + _state.pos(2);
}
}
} else if (_gps_data_ready && !_gps_hgt_faulty &&
!(_in_range_aid_mode && !_range_data_ready && !_rng_hgt_faulty)) {
setControlGPSHeight();
_fuse_height = true;
_in_range_aid_mode = false;
// we have just switched to using gps height, calculate height sensor offset such that current
// measurment matches our current height estimate
if (_control_status_prev.flags.gps_hgt != _control_status.flags.gps_hgt) {
_hgt_sensor_offset = _gps_sample_delayed.hgt - _gps_alt_ref + _state.pos(2);
}
} else if (_control_status.flags.baro_hgt && _baro_data_ready && !_baro_hgt_faulty) {
// switch to baro if there was a reset to baro
_fuse_height = true;
_in_range_aid_mode = false;
// we have just switched to using baro height, we don't need to set a height sensor offset
// since we track a separate _baro_hgt_offset
if (_control_status_prev.flags.baro_hgt != _control_status.flags.baro_hgt) {
_hgt_sensor_offset = 0.0f;
}
}
}
// calculate a filtered offset between the baro origin and local NED origin if we are not using the baro as a height reference
if (!_control_status.flags.baro_hgt && _baro_data_ready) {
float local_time_step = 1e-6f * _delta_time_baro_us;
local_time_step = math::constrain(local_time_step, 0.0f, 1.0f);
// apply a 10 second first order low pass filter to baro offset
float offset_rate_correction = 0.1f * (_baro_sample_delayed.hgt + _state.pos(
2) - _baro_hgt_offset);
_baro_hgt_offset += local_time_step * math::constrain(offset_rate_correction, -0.1f, 0.1f);
}
if ((_time_last_imu - _time_last_hgt_fuse) > 2 * RNG_MAX_INTERVAL && _control_status.flags.rng_hgt && !_range_data_ready) {
// If we are supposed to be using range finder data as the primary height sensor, have missed or rejected measurements
// and are on the ground, then synthesise a measurement at the expected on ground value
if (!_control_status.flags.in_air) {
_range_sample_delayed.rng = _params.rng_gnd_clearance;
_range_sample_delayed.time_us = _imu_sample_delayed.time_us;
}
_fuse_height = true;
}
}
bool Ekf::rangeAidConditionsMet(bool in_range_aid_mode)
{
// if the parameter for range aid is enabled we allow to switch from using the primary height source to using range finder as height source
// under the following conditions
// 1) we are not further than max_range_for_dual_fusion away from the ground
// 2) our ground speed is not higher than max_vel_for_dual_fusion
// 3) Our terrain estimate is stable (needs better checks)
if (_params.range_aid) {
// check if we should use range finder measurements to estimate height, use hysteresis to avoid rapid switching
bool use_range_finder;
if (in_range_aid_mode) {
use_range_finder = (_terrain_vpos - _state.pos(2) < _params.max_hagl_for_range_aid) && get_terrain_valid();
} else {
// if we were not using range aid in the previous iteration then require the current height above terrain to be
// smaller than 70 % of the maximum allowed ground distance for range aid
use_range_finder = (_terrain_vpos - _state.pos(2) < 0.7f * _params.max_hagl_for_range_aid) && get_terrain_valid();
}
bool horz_vel_valid = (_control_status.flags.gps || _control_status.flags.ev_pos || _control_status.flags.opt_flow)
&& (_fault_status.value == 0);
if (horz_vel_valid) {
float ground_vel = sqrtf(_state.vel(0) * _state.vel(0) + _state.vel(1) * _state.vel(1));
if (in_range_aid_mode) {
use_range_finder &= ground_vel < _params.max_vel_for_range_aid;
} else {
// if we were not using range aid in the previous iteration then require the ground velocity to be
// smaller than 70 % of the maximum allowed ground velocity for range aid
use_range_finder &= ground_vel < 0.7f * _params.max_vel_for_range_aid;
}
} else {
use_range_finder = false;
}
// use hysteresis to check for hagl
if (in_range_aid_mode) {
use_range_finder &= ((_hagl_innov * _hagl_innov / (sq(_params.range_aid_innov_gate) * _hagl_innov_var)) < 1.0f);
} else {
// if we were not using range aid in the previous iteration then use a much lower (1/100) threshold to avoid
// switching to range finder too soon (wait for terrain to update).
use_range_finder &= ((_hagl_innov * _hagl_innov / (sq(_params.range_aid_innov_gate) * _hagl_innov_var)) < 0.01f);
}
return use_range_finder;
} else {
return false;
}
}
void Ekf::checkForStuckRange()
{
if (_range_data_ready && _range_sample_delayed.time_us - _time_last_rng_ready > 10e6 &&
_control_status.flags.in_air) {
_rng_stuck = true;
//require a variance of rangefinder values to check for "stuck" measurements
if (_rng_check_max_val - _rng_check_min_val > 1.0f) {
_time_last_rng_ready = _range_sample_delayed.time_us;
_rng_check_min_val = 0.0f;
_rng_check_max_val = 0.0f;
_rng_stuck = false;
} else {
if (_range_sample_delayed.rng > _rng_check_max_val) {
_rng_check_max_val = _range_sample_delayed.rng;
}
if (_rng_check_min_val < 0.1f || _range_sample_delayed.rng < _rng_check_min_val) {
_rng_check_min_val = _range_sample_delayed.rng;
}
_range_data_ready = false;
}
} else if (_range_data_ready) {
_time_last_rng_ready = _range_sample_delayed.time_us;
}
}
void Ekf::controlAirDataFusion()
{
// control activation and initialisation/reset of wind states required for airspeed fusion
// If both airspeed and sideslip fusion have timed out and we are not using a drag observation model then we no longer have valid wind estimates
bool airspeed_timed_out = _time_last_imu - _time_last_arsp_fuse > 10e6;
bool sideslip_timed_out = _time_last_imu - _time_last_beta_fuse > 10e6;
if (_control_status.flags.wind && airspeed_timed_out && sideslip_timed_out && !(_params.fusion_mode & MASK_USE_DRAG)) {
_control_status.flags.wind = false;
}
// Always try to fuse airspeed data if available and we are in flight and the filter is operating in a normal aiding mode
bool is_aiding = _control_status.flags.gps || _control_status.flags.opt_flow || _control_status.flags.ev_pos;
if (_tas_data_ready && _control_status.flags.in_air && is_aiding) {
// If starting wind state estimation, reset the wind states and covariances before fusing any data
if (!_control_status.flags.wind) {
// activate the wind states
_control_status.flags.wind = true;
// reset the timout timer to prevent repeated resets
_time_last_arsp_fuse = _time_last_imu;
_time_last_beta_fuse = _time_last_imu;
// reset the wind speed states and corresponding covariances
resetWindStates();
resetWindCovariance();
}
fuseAirspeed();
}
}
void Ekf::controlBetaFusion()
{
// control activation and initialisation/reset of wind states required for synthetic sideslip fusion fusion
// If both airspeed and sideslip fusion have timed out and we are not using a drag observation model then we no longer have valid wind estimates
bool sideslip_timed_out = _time_last_imu - _time_last_beta_fuse > 10e6;
bool airspeed_timed_out = _time_last_imu - _time_last_arsp_fuse > 10e6;
if(_control_status.flags.wind && airspeed_timed_out && sideslip_timed_out && !(_params.fusion_mode & MASK_USE_DRAG)) {
_control_status.flags.wind = false;
}
// Perform synthetic sideslip fusion when in-air and sideslip fuson had been enabled externally in addition to the following criteria:
// Suffient time has lapsed sice the last fusion
bool beta_fusion_time_triggered = _time_last_imu - _time_last_beta_fuse > _params.beta_avg_ft_us;
// The filter is operating in a mode where velocity states can be used
bool vel_states_active = _control_status.flags.gps || _control_status.flags.opt_flow || _control_status.flags.ev_pos;
if(beta_fusion_time_triggered && _control_status.flags.fuse_beta && _control_status.flags.in_air && vel_states_active) {
// If starting wind state estimation, reset the wind states and covariances before fusing any data
if (!_control_status.flags.wind) {
// activate the wind states
_control_status.flags.wind = true;
// reset the timeout timers to prevent repeated resets
_time_last_beta_fuse = _time_last_imu;
_time_last_arsp_fuse = _time_last_imu;
// reset the wind speed states and corresponding covariances
resetWindStates();
resetWindCovariance();
}
fuseSideslip();
}
}
void Ekf::controlDragFusion()
{
if (_params.fusion_mode & MASK_USE_DRAG) {
if (_control_status.flags.in_air) {
if (!_control_status.flags.wind) {
// reset the wind states and covariances when starting drag accel fusion
_control_status.flags.wind = true;
resetWindStates();
resetWindCovariance();
} else if (_drag_buffer.pop_first_older_than(_imu_sample_delayed.time_us, &_drag_sample_delayed)) {
fuseDrag();
}
} else {
_control_status.flags.wind = false;
}
}
}
void Ekf::controlMagFusion()
{
// If we are using external vision data for heading then no magnetometer fusion is used
if (_control_status.flags.ev_yaw) {
return;
}
// If we are on ground, store the local position and time to use as a reference
// Also reset the flight alignment flag so that the mag fields will be re-initialised next time we achieve flight altitude
if (!_control_status.flags.in_air) {
_last_on_ground_posD = _state.pos(2);
_flt_mag_align_complete = false;
_num_bad_flight_yaw_events = 0;
}
// check for new magnetometer data that has fallen behind the fusion time horizon
if (_mag_data_ready) {
// Determine if we should use simple magnetic heading fusion which works better when there are large external disturbances
// or the more accurate 3-axis fusion
if (_control_status.flags.mag_fault) {
// do no magnetometer fusion at all
_control_status.flags.mag_hdg = false;
_control_status.flags.mag_3D = false;
} else if (_params.mag_fusion_type == MAG_FUSE_TYPE_AUTO) {
// Check if height has increased sufficiently to be away from ground magnetic anomalies
bool height_achieved = (_last_on_ground_posD - _state.pos(2)) > 1.5f;
// Check if there has been enough change in horizontal velocity to make yaw observable
// Apply hysteresis to check to avoid rapid toggling
if (_yaw_angle_observable) {
_yaw_angle_observable = _accel_lpf_NE.norm() > _params.mag_acc_gate;
} else {
_yaw_angle_observable = _accel_lpf_NE.norm() > 2.0f * _params.mag_acc_gate;
}
_yaw_angle_observable = _yaw_angle_observable && (_control_status.flags.gps || _control_status.flags.ev_pos);
// check if there is enough yaw rotation to make the mag bias states observable
if (!_mag_bias_observable && (fabsf(_yaw_rate_lpf_ef) > _params.mag_yaw_rate_gate)) {
// initial yaw motion is detected
_mag_bias_observable = true;
_yaw_delta_ef = 0.0f;
_time_yaw_started = _imu_sample_delayed.time_us;
} else if (_mag_bias_observable) {
// monitor yaw rotation in 45 deg sections.
// a rotation of 45 deg is sufficient to make the mag bias observable
if (fabsf(_yaw_delta_ef) > 0.7854f) {
_time_yaw_started = _imu_sample_delayed.time_us;
_yaw_delta_ef = 0.0f;
}
// require sustained yaw motion of 50% the initial yaw rate threshold
float min_yaw_change_req = 0.5f * _params.mag_yaw_rate_gate * (1e-6f * (float)(_imu_sample_delayed.time_us - _time_yaw_started));
_mag_bias_observable = fabsf(_yaw_delta_ef) > min_yaw_change_req;
} else {
_mag_bias_observable = false;
}
// record the last time that movement was suitable for use of 3-axis magnetometer fusion
if (_mag_bias_observable || _yaw_angle_observable) {
_time_last_movement = _imu_sample_delayed.time_us;
}
// decide whether 3-axis magnetomer fusion can be used
bool use_3D_fusion = _control_status.flags.tilt_align && // Use of 3D fusion requires valid tilt estimates
_control_status.flags.in_air && // don't use when on the ground becasue of magnetic anomalies
(_flt_mag_align_complete || height_achieved) && // once in-flight field alignment has been performed, ignore relative height
((_imu_sample_delayed.time_us - _time_last_movement) < 2 * 1000 * 1000); // Using 3-axis fusion for a minimum period after to allow for false negatives
// perform switch-over
if (use_3D_fusion) {
if (!_control_status.flags.mag_3D) {
if (!_flt_mag_align_complete) {
// If we are flying a vehicle that flies forward, eg plane, then we can use the GPS course to check and correct the heading
if (_control_status.flags.fixed_wing && _control_status.flags.in_air) {
_control_status.flags.yaw_align = realignYawGPS();
_flt_mag_align_complete = _control_status.flags.yaw_align;
} else {
_control_status.flags.yaw_align = resetMagHeading(_mag_sample_delayed.mag);
_flt_mag_align_complete = _control_status.flags.yaw_align;
}
} else {
// reset the mag field covariances
zeroRows(P, 16, 21);
zeroCols(P, 16, 21);
// re-instate the last used variances
for (uint8_t index = 0; index <= 5; index ++) {
P[index+16][index+16] = _saved_mag_variance[index];
}
}
}
// only use one type of mag fusion at the same time
_control_status.flags.mag_3D = _flt_mag_align_complete;
_control_status.flags.mag_hdg = !_control_status.flags.mag_3D;
} else {
// save magnetic field state variances for next time
if (_control_status.flags.mag_3D) {
for (uint8_t index = 0; index <= 5; index ++) {
_saved_mag_variance[index] = P[index+16][index+16];
}
_control_status.flags.mag_3D = false;
}
_control_status.flags.mag_hdg = true;
}
// perform switch-over from only updating the mag states to updating all states
if (!_control_status.flags.update_mag_states_only && _control_status_prev.flags.update_mag_states_only) {
// When re-commencing use of magnetometer to correct vehicle states
// set the field state variance to the observation variance and zero
// the covariance terms to allow the field states re-learn rapidly
zeroRows(P, 16, 21);
zeroCols(P, 16, 21);
for (uint8_t index = 0; index <= 5; index ++) {
P[index+16][index+16] = sq(_params.mag_noise);
}
}
} else if (_params.mag_fusion_type == MAG_FUSE_TYPE_HEADING) {
// always use heading fusion
_control_status.flags.mag_hdg = true;
_control_status.flags.mag_3D = false;
} else if (_params.mag_fusion_type == MAG_FUSE_TYPE_3D) {
// if transitioning into 3-axis fusion mode, we need to initialise the yaw angle and field states
if (!_control_status.flags.mag_3D) {
_control_status.flags.yaw_align = resetMagHeading(_mag_sample_delayed.mag);
}
// always use 3-axis mag fusion
_control_status.flags.mag_hdg = false;
_control_status.flags.mag_3D = true;
} else {
// do no magnetometer fusion at all
_control_status.flags.mag_hdg = false;
_control_status.flags.mag_3D = false;
}
// if we are using 3-axis magnetometer fusion, but without external aiding, then the declination must be fused as an observation to prevent long term heading drift
// fusing declination when gps aiding is available is optional, but recommended to prevent problem if the vehicle is static for extended periods of time
if (_control_status.flags.mag_3D && (!_control_status.flags.gps || (_params.mag_declination_source & MASK_FUSE_DECL))) {
_control_status.flags.mag_dec = true;
} else {
_control_status.flags.mag_dec = false;
}
// fuse magnetometer data using the selected methods
if (_control_status.flags.mag_3D && _control_status.flags.yaw_align) {
fuseMag();
if (_control_status.flags.mag_dec) {
fuseDeclination();
}
} else if (_control_status.flags.mag_hdg && _control_status.flags.yaw_align) {
// fusion of an Euler yaw angle from either a 321 or 312 rotation sequence
fuseHeading();
} else {
// do no fusion at all
}
}
}
void Ekf::controlVelPosFusion()
{
// if we aren't doing any aiding, fake GPS measurements at the last known position to constrain drift
// Coincide fake measurements with baro data for efficiency with a minimum fusion rate of 5Hz
if (!_control_status.flags.gps && !_control_status.flags.opt_flow && !_control_status.flags.ev_pos
&& ((_time_last_imu - _time_last_fake_gps > 2e5) || _fuse_height)) {
_fuse_pos = true;
_time_last_fake_gps = _time_last_imu;
}
// Fuse available NED velocity and position data into the main filter
if (_fuse_height || _fuse_pos || _fuse_hor_vel || _fuse_vert_vel) {
fuseVelPosHeight();
_fuse_hor_vel = _fuse_vert_vel = _fuse_pos = _fuse_height = false;
}
}