AP_AHRS: choose the best accelerometer at each drift correction step
this greatly reduces the impact of aliasing on accelerometers by choosing the accelerometer that produces the smallest error term in DCM. The difference can be quite dramatic on the Pixhawk.
This commit is contained in:
Andrew Tridgell
parent
0b45d2bc06
commit
b53496d470
@ -47,7 +47,8 @@ public:
|
||||
_cos_yaw(1.0f),
|
||||
_sin_roll(0.0f),
|
||||
_sin_pitch(0.0f),
|
||||
_sin_yaw(0.0f)
|
||||
_sin_yaw(0.0f),
|
||||
_active_accel_instance(0)
|
||||
{
|
||||
// load default values from var_info table
|
||||
AP_Param::setup_object_defaults(this, var_info);
|
||||
@ -127,7 +128,7 @@ public:
|
||||
}
|
||||
|
||||
// accelerometer values in the earth frame in m/s/s
|
||||
const Vector3f &get_accel_ef(void) const { return _accel_ef; }
|
||||
const Vector3f &get_accel_ef(void) const { return _accel_ef[_active_accel_instance]; }
|
||||
|
||||
// Methods
|
||||
virtual void update(void) = 0;
|
||||
@ -301,6 +302,9 @@ public:
|
||||
// with very accurate position and velocity
|
||||
virtual bool have_inertial_nav(void) const { return false; }
|
||||
|
||||
// return the active accelerometer instance
|
||||
uint8_t get_active_accel_instance(void) const { return _active_accel_instance; }
|
||||
|
||||
protected:
|
||||
// settable parameters
|
||||
AP_Float beta;
|
||||
@ -348,7 +352,7 @@ protected:
|
||||
float _gyro_drift_limit;
|
||||
|
||||
// accelerometer values in the earth frame in m/s/s
|
||||
Vector3f _accel_ef;
|
||||
Vector3f _accel_ef[INS_MAX_INSTANCES];
|
||||
|
||||
// Declare filter states for HPF and LPF used by complementary
|
||||
// filter in AP_AHRS::groundspeed_vector
|
||||
@ -362,6 +366,9 @@ protected:
|
||||
// helper trig variables
|
||||
float _cos_roll, _cos_pitch, _cos_yaw;
|
||||
float _sin_roll, _sin_pitch, _sin_yaw;
|
||||
|
||||
// which accelerometer instance is active
|
||||
uint8_t _active_accel_instance;
|
||||
};
|
||||
|
||||
#include <AP_AHRS_DCM.h>
|
||||
|
||||
@ -57,7 +57,7 @@ AP_AHRS_DCM::update(void)
|
||||
// otherwise we may move too far. This happens when arming motors
|
||||
// in ArduCopter
|
||||
if (delta_t > 0.2f) {
|
||||
_ra_sum.zero();
|
||||
memset(&_ra_sum[0], 0, sizeof(_ra_sum));
|
||||
_ra_deltat = 0;
|
||||
return;
|
||||
}
|
||||
@ -90,8 +90,21 @@ AP_AHRS_DCM::matrix_update(float _G_Dt)
|
||||
// and including the P terms would give positive feedback into
|
||||
// the _P_gain() calculation, which can lead to a very large P
|
||||
// value
|
||||
_omega = _ins.get_gyro() + _omega_I;
|
||||
_omega.zero();
|
||||
|
||||
// average across all healthy gyros. This reduces noise on systems
|
||||
// with more than one gyro
|
||||
uint8_t healthy_count = 0;
|
||||
for (uint8_t i=0; i<_ins.get_gyro_count(); i++) {
|
||||
if (_ins.get_gyro_health(i)) {
|
||||
_omega += _ins.get_gyro(i);
|
||||
healthy_count++;
|
||||
}
|
||||
}
|
||||
if (healthy_count > 1) {
|
||||
_omega /= healthy_count;
|
||||
}
|
||||
_omega += _omega_I;
|
||||
_dcm_matrix.rotate((_omega + _omega_P + _omega_yaw_P) * _G_Dt);
|
||||
}
|
||||
|
||||
@ -287,9 +300,10 @@ AP_AHRS_DCM::_P_gain(float spin_rate)
|
||||
// always available. TODO check the necessity of adding adjustable acc threshold
|
||||
// and/or filtering accelerations before getting magnitude
|
||||
float
|
||||
AP_AHRS_DCM::_yaw_gain(Vector3f VdotEF)
|
||||
AP_AHRS_DCM::_yaw_gain(void) const
|
||||
{
|
||||
float VdotEFmag = sqrtf(sq(_accel_ef.x) + sq(_accel_ef.y));
|
||||
float VdotEFmag = pythagorous2(_accel_ef[_active_accel_instance].x,
|
||||
_accel_ef[_active_accel_instance].y);
|
||||
if (VdotEFmag <= 4.0f) {
|
||||
return 0.2f*(4.5f - VdotEFmag);
|
||||
}
|
||||
@ -439,7 +453,7 @@ AP_AHRS_DCM::drift_correction_yaw(void)
|
||||
// is proportional to how observable the heading is from the acceerations and GPS velocity
|
||||
// The accelration derived heading will be more reliable in turns than compass or GPS
|
||||
|
||||
_omega_yaw_P.z = error_z * _P_gain(spin_rate) * _kp_yaw * _yaw_gain(_accel_ef);
|
||||
_omega_yaw_P.z = error_z * _P_gain(spin_rate) * _kp_yaw * _yaw_gain();
|
||||
if (_flags.fast_ground_gains) {
|
||||
_omega_yaw_P.z *= 8;
|
||||
}
|
||||
@ -457,13 +471,14 @@ AP_AHRS_DCM::drift_correction_yaw(void)
|
||||
|
||||
|
||||
/**
|
||||
return an accel vector delayed by AHRS_ACCEL_DELAY samples
|
||||
return an accel vector delayed by AHRS_ACCEL_DELAY samples for a
|
||||
specific accelerometer instance
|
||||
*/
|
||||
Vector3f AP_AHRS_DCM::ra_delayed(const Vector3f &ra)
|
||||
Vector3f AP_AHRS_DCM::ra_delayed(uint8_t instance, const Vector3f &ra)
|
||||
{
|
||||
// get the old element, and then fill it with the new element
|
||||
Vector3f ret = _ra_delay_buffer;
|
||||
_ra_delay_buffer = ra;
|
||||
Vector3f ret = _ra_delay_buffer[instance];
|
||||
_ra_delay_buffer[instance] = ra;
|
||||
return ret;
|
||||
}
|
||||
|
||||
@ -487,10 +502,13 @@ AP_AHRS_DCM::drift_correction(float deltat)
|
||||
drift_correction_yaw();
|
||||
|
||||
// rotate accelerometer values into the earth frame
|
||||
_accel_ef = _dcm_matrix * _ins.get_accel();
|
||||
|
||||
// integrate the accel vector in the earth frame between GPS readings
|
||||
_ra_sum += _accel_ef * deltat;
|
||||
for (uint8_t i=0; i<_ins.get_accel_count(); i++) {
|
||||
if (_ins.get_accel_health(i)) {
|
||||
_accel_ef[i] = _dcm_matrix * _ins.get_accel(i);
|
||||
// integrate the accel vector in the earth frame between GPS readings
|
||||
_ra_sum[i] += _accel_ef[i] * deltat;
|
||||
}
|
||||
}
|
||||
|
||||
// keep a sum of the deltat values, so we know how much time
|
||||
// we have integrated over
|
||||
@ -575,8 +593,10 @@ AP_AHRS_DCM::drift_correction(float deltat)
|
||||
GA_e = Vector3f(0, 0, -1.0f);
|
||||
|
||||
bool using_gps_corrections = false;
|
||||
float ra_scale = 1.0f/(_ra_deltat*GRAVITY_MSS);
|
||||
|
||||
if (_flags.correct_centrifugal && (_have_gps_lock || _flags.fly_forward)) {
|
||||
float v_scale = gps_gain.get()/(_ra_deltat*GRAVITY_MSS);
|
||||
float v_scale = gps_gain.get() * ra_scale;
|
||||
Vector3f vdelta = (velocity - _last_velocity) * v_scale;
|
||||
GA_e += vdelta;
|
||||
GA_e.normalize();
|
||||
@ -588,22 +608,50 @@ AP_AHRS_DCM::drift_correction(float deltat)
|
||||
}
|
||||
|
||||
// calculate the error term in earth frame.
|
||||
_ra_sum /= (_ra_deltat * GRAVITY_MSS);
|
||||
// we do this for each available accelerometer then pick the
|
||||
// accelerometer that leads to the smallest error term. This takes
|
||||
// advantage of the different sample rates on different
|
||||
// accelerometers to dramatically reduce the impact of aliasing
|
||||
// due to harmonics of vibrations that match closely the sampling
|
||||
// rate of our accelerometers. On the Pixhawk we have the LSM303D
|
||||
// running at 800Hz and the MPU6000 running at 1kHz, by combining
|
||||
// the two the effects of aliasing are greatly reduced.
|
||||
Vector3f error[INS_MAX_INSTANCES];
|
||||
Vector3f GA_b[INS_MAX_INSTANCES];
|
||||
int8_t besti = -1;
|
||||
float best_error = 0;
|
||||
for (uint8_t i=0; i<_ins.get_accel_count(); i++) {
|
||||
if (!_ins.get_accel_health()) {
|
||||
// only use healthy sensors
|
||||
continue;
|
||||
}
|
||||
_ra_sum[i] *= ra_scale;
|
||||
|
||||
// get the delayed ra_sum to match the GPS lag
|
||||
Vector3f GA_b;
|
||||
if (using_gps_corrections) {
|
||||
GA_b = ra_delayed(_ra_sum);
|
||||
} else {
|
||||
GA_b = _ra_sum;
|
||||
// get the delayed ra_sum to match the GPS lag
|
||||
if (using_gps_corrections) {
|
||||
GA_b[i] = ra_delayed(i, _ra_sum[i]);
|
||||
} else {
|
||||
GA_b[i] = _ra_sum[i];
|
||||
}
|
||||
GA_b[i].normalize();
|
||||
if (GA_b[i].is_inf()) {
|
||||
// wait for some non-zero acceleration information
|
||||
continue;
|
||||
}
|
||||
error[i] = GA_b[i] % GA_e;
|
||||
float error_length = error[i].length();
|
||||
if (besti == -1 || error_length < best_error) {
|
||||
besti = i;
|
||||
best_error = error_length;
|
||||
}
|
||||
}
|
||||
GA_b.normalize();
|
||||
if (GA_b.is_inf()) {
|
||||
// wait for some non-zero acceleration information
|
||||
|
||||
if (besti == -1) {
|
||||
// no healthy accelerometers!
|
||||
return;
|
||||
}
|
||||
|
||||
Vector3f error = GA_b % GA_e;
|
||||
_active_accel_instance = besti;
|
||||
|
||||
#define YAW_INDEPENDENT_DRIFT_CORRECTION 0
|
||||
#if YAW_INDEPENDENT_DRIFT_CORRECTION
|
||||
@ -614,13 +662,13 @@ AP_AHRS_DCM::drift_correction(float deltat)
|
||||
float tilt = pythagorous2(GA_e.x, GA_e.y);
|
||||
|
||||
// equation 11
|
||||
float theta = atan2f(GA_b.y, GA_b.x);
|
||||
float theta = atan2f(GA_b[besti].y, GA_b[besti].x);
|
||||
|
||||
// equation 12
|
||||
Vector3f GA_e2 = Vector3f(cosf(theta)*tilt, sinf(theta)*tilt, GA_e.z);
|
||||
|
||||
// step 6
|
||||
error = GA_b % GA_e2;
|
||||
error = GA_b[besti] % GA_e2;
|
||||
error.z = earth_error_Z;
|
||||
#endif // YAW_INDEPENDENT_DRIFT_CORRECTION
|
||||
|
||||
@ -630,9 +678,9 @@ AP_AHRS_DCM::drift_correction(float deltat)
|
||||
// accelerometers at high roll angles as long as we have a GPS
|
||||
if (use_compass()) {
|
||||
if (have_gps() && gps_gain == 1.0f) {
|
||||
error.z *= sinf(fabsf(roll));
|
||||
error[besti].z *= sinf(fabsf(roll));
|
||||
} else {
|
||||
error.z = 0;
|
||||
error[besti].z = 0;
|
||||
}
|
||||
}
|
||||
|
||||
@ -640,19 +688,19 @@ AP_AHRS_DCM::drift_correction(float deltat)
|
||||
// hope the gyros are OK for a while. Just slowly reduce _omega_P
|
||||
// to prevent previous bad accels from throwing us off
|
||||
if (!_ins.healthy()) {
|
||||
error.zero();
|
||||
error[besti].zero();
|
||||
} else {
|
||||
// convert the error term to body frame
|
||||
error = _dcm_matrix.mul_transpose(error);
|
||||
error[besti] = _dcm_matrix.mul_transpose(error[besti]);
|
||||
}
|
||||
|
||||
if (error.is_nan() || error.is_inf()) {
|
||||
if (error[besti].is_nan() || error[besti].is_inf()) {
|
||||
// don't allow bad values
|
||||
check_matrix();
|
||||
return;
|
||||
}
|
||||
|
||||
_error_rp_sum += error.length();
|
||||
_error_rp_sum += best_error;
|
||||
_error_rp_count++;
|
||||
|
||||
// base the P gain on the spin rate
|
||||
@ -661,7 +709,7 @@ AP_AHRS_DCM::drift_correction(float deltat)
|
||||
// we now want to calculate _omega_P and _omega_I. The
|
||||
// _omega_P value is what drags us quickly to the
|
||||
// accelerometer reading.
|
||||
_omega_P = error * _P_gain(spin_rate) * _kp;
|
||||
_omega_P = error[besti] * _P_gain(spin_rate) * _kp;
|
||||
if (_flags.fast_ground_gains) {
|
||||
_omega_P *= 8;
|
||||
}
|
||||
@ -678,7 +726,7 @@ AP_AHRS_DCM::drift_correction(float deltat)
|
||||
|
||||
// accumulate some integrator error
|
||||
if (spin_rate < ToRad(SPIN_RATE_LIMIT)) {
|
||||
_omega_I_sum += error * _ki * _ra_deltat;
|
||||
_omega_I_sum += error[besti] * _ki * _ra_deltat;
|
||||
_omega_I_sum_time += _ra_deltat;
|
||||
}
|
||||
|
||||
@ -697,7 +745,7 @@ AP_AHRS_DCM::drift_correction(float deltat)
|
||||
}
|
||||
|
||||
// zero our accumulator ready for the next GPS step
|
||||
_ra_sum.zero();
|
||||
memset(&_ra_sum[0], 0, sizeof(_ra_sum));
|
||||
_ra_deltat = 0;
|
||||
_ra_sum_start = last_correction_time;
|
||||
|
||||
|
||||
@ -110,14 +110,14 @@ private:
|
||||
Vector3f _omega; // Corrected Gyro_Vector data
|
||||
|
||||
// variables to cope with delaying the GA sum to match GPS lag
|
||||
Vector3f ra_delayed(const Vector3f &ra);
|
||||
Vector3f _ra_delay_buffer;
|
||||
Vector3f ra_delayed(uint8_t instance, const Vector3f &ra);
|
||||
Vector3f _ra_delay_buffer[INS_MAX_INSTANCES];
|
||||
|
||||
// P term gain based on spin rate
|
||||
float _P_gain(float spin_rate);
|
||||
|
||||
// P term yaw gain based on rate of change of horiz velocity
|
||||
float _yaw_gain(Vector3f VdotEF);
|
||||
float _yaw_gain(void) const;
|
||||
|
||||
// state to support status reporting
|
||||
float _renorm_val_sum;
|
||||
@ -133,7 +133,7 @@ private:
|
||||
uint32_t _gps_last_update;
|
||||
|
||||
// state of accel drift correction
|
||||
Vector3f _ra_sum;
|
||||
Vector3f _ra_sum[INS_MAX_INSTANCES];
|
||||
Vector3f _last_velocity;
|
||||
float _ra_deltat;
|
||||
uint32_t _ra_sum_start;
|
||||
|
||||
Loading…
Reference in New Issue
Block a user