gps_acc_fusion_filter Class Reference

gps_acc_fusion_filter - Kalman filter implementation for GPS and accelerometer fusion More...

#include <gps_acc_fusion_filter.h>

Inheritance diagram for gps_acc_fusion_filter:

Public Member Functions
	gps_acc_fusion_filter ()
void	reset (double x, double y, double ts, double x_vel, double y_vel, double acc_sigma_2, double pos_sigma_2)
void	predict (double x_acc, double y_acc, double ts_sec)
void	update (const kf_state &state, double pos_sigma_2, double vel_sigma_2=1e-6)
const kf_state	current_state () const

Private Member Functions
void	rebuild_F (double dt_sec)
	gps position sigma^2
void	rebuild_U (double xAcc, double yAcc)
void	rebuild_B (double dt_sec)
void	rebuild_Q (double dt_sec)
void	rebuild_R (double pos_sigma_2, double vel_sigma_2)

Private Attributes
double	m_last_predict_sec
double	m_acc_sigma_2
	timestamp (in seconds) of last predict() call
double	m_pos_sigma_2
	accelerometer sigma^2

Additional Inherited Members
Protected Member Functions inherited from kalman_filter< 4, 4, 2 >
	kalman_filter ()
void	estimate ()
void	correct ()
Protected Attributes inherited from kalman_filter< 4, 4, 2 >
mtx_t< double, state_d, state_d >	F
mtx_t< double, measure_d, state_d >	H
	state transition model
mtx_t< double, state_d, control_d >	B
	observation model
mtx_t< double, state_d, state_d >	Q
	control matrix
mtx_t< double, measure_d, measure_d >	R
	process noise covariance
mtx_t< double, control_d, 1 >	Uk
	observation noise covariance
mtx_t< double, measure_d, 1 >	Zk
	control vector
mtx_t< double, state_d, 1 >	Xk_km1
	actual measured values vector
mtx_t< double, state_d, state_d >	Pk_km1
	predicted state estimate
mtx_t< double, measure_d, 1 >	Yk
	predicted estimate covariance
mtx_t< double, measure_d, measure_d >	Sk
	measurement innovation
mtx_t< double, state_d, measure_d >	K
	innovation covariance
mtx_t< double, state_d, 1 >	Xk_k
	Kalman gain.
mtx_t< double, measure_d, 1 >	Yk_k
	updated (current) state
mtx_t< double, state_d, state_d >	Pk_k
	post fit residual
mtx_t< double, state_d, state_d >	I
	updated estimate covariance

Detailed Description

gps_acc_fusion_filter - Kalman filter implementation for GPS and accelerometer fusion

Template Parameters

state_d	Dimension of state vector (number of state variables)
measure_d	Dimension of measurement vector (number of measurements)
control_d	Dimension of control vector (number of control inputs)

Specialized Kalman filter for fusing GPS position data with accelerometer measurements to estimate 2D position and velocity. Uses kalman_filter<4, 4, 2> configuration:

State Vector (4D): [X, Y, X', Y']

• X, Y: Position coordinates (East, North) in meters
• X', Y': Velocity components (East, North) in m/s

Measurement Vector (4D): [GPS_X, GPS_Y, GPS_VX, GPS_VY]

• GPS position and velocity measurements

Control Vector (2D): [ACC_X, ACC_Y]

• Accelerometer measurements (East, North acceleration)

This filter provides improved position and velocity estimation by combining the absolute positioning from GPS with the high-frequency acceleration data, resulting in smoother tracking with reduced GPS noise and better dynamics.

Constructor & Destructor Documentation

gps_acc_fusion_filter()

gps_acc_fusion_filter::gps_acc_fusion_filter

(

)

Member Function Documentation

current_state()

const kf_state gps_acc_fusion_filter::current_state

(

)

const

predict()

void gps_acc_fusion_filter::predict	(	double	x_acc,
		double	y_acc,
		double	ts_sec
	)

rebuild_B()

void gps_acc_fusion_filter::rebuild_B ( double  dt_sec )

private

rebuild_F()

void gps_acc_fusion_filter::rebuild_F ( double  dt_sec )

private

gps position sigma^2

rebuild_Q()

void gps_acc_fusion_filter::rebuild_Q ( double  dt_sec )

private

rebuild_R()

void gps_acc_fusion_filter::rebuild_R ( double  pos_sigma_2, double  vel_sigma_2  )

private

rebuild_U()

void gps_acc_fusion_filter::rebuild_U ( double  xAcc, double  yAcc  )

private

reset()

void gps_acc_fusion_filter::reset	(	double	x,
		double	y,
		double	ts,
		double	x_vel,
		double	y_vel,
		double	acc_sigma_2,
		double	pos_sigma_2
	)

update()

void gps_acc_fusion_filter::update	(	const kf_state &	state,
		double	pos_sigma_2,
		double	vel_sigma_2 = 1e-6
	)

Member Data Documentation

m_acc_sigma_2

double gps_acc_fusion_filter::m_acc_sigma_2

private

timestamp (in seconds) of last predict() call

m_last_predict_sec

double gps_acc_fusion_filter::m_last_predict_sec

private

m_pos_sigma_2

double gps_acc_fusion_filter::m_pos_sigma_2

private

accelerometer sigma^2

---

The documentation for this class was generated from the following files:

• inc/gps_acc_fusion_filter.h
• src/gps_acc_fusion_filter.cpp