MLM Class Reference

Motion Location Model - GPS/Accelerometer sensor fusion estimator. More...

#include <mlm.h>

Public Member Functions
	MLM (void)
	Default constructor with standard noise parameters.
	MLM (double acc_sigma_2, double loc_sigma_2, double vel_sigma_2)
	Constructor with custom noise parameters.
	~MLM (void)
	Destructor.
bool	process_acc_data (const enu_accelerometer &acc, double time_sec)
	Process accelerometer measurement (prediction step)
void	process_gps_data (const gps_coordinate &gps, double time_sec)
	Process GPS measurement (update step)
gps_coordinate	predicted_coordinate () const
	Get current estimated position.

Private Attributes
gps_acc_fusion_filter	m_fk
	Kalman filter for GPS-accelerometer fusion.
GeographicLib::LocalCartesian	m_lc
	Local cartesian coordinate converter.
bool	m_got_start_point
	Flag indicating if initial reference point is set.
double	m_acc_sigma_2
	Accelerometer noise variance [m²/s⁴].
double	m_loc_sigma_2
	Location measurement noise variance [m²].
double	m_vel_sigma_2
	Velocity process noise variance [m²/s²].

Detailed Description

Motion Location Model - GPS/Accelerometer sensor fusion estimator.

This class implements a Kalman filter-based sensor fusion system that combines GPS position measurements with accelerometer data to provide improved position and velocity estimates. The system uses ENU (East-North-Up) coordinate frame for local calculations and converts to/from geographic coordinates.

The filter operates in predict-update cycles:

• Prediction: Uses accelerometer data to propagate state forward in time
• Update: Corrects predictions using GPS position measurements

Noise characteristics are configurable via sigma-squared parameters for accelerometer, location, and velocity uncertainties.

Constructor & Destructor Documentation

MLM() [1/2]

MLM::MLM

(

void

)

Default constructor with standard noise parameters.

MLM() [2/2]

MLM::MLM	(	double	acc_sigma_2,
		double	loc_sigma_2,
		double	vel_sigma_2
	)

Constructor with custom noise parameters.

Parameters

acc_sigma_2	Accelerometer noise variance [m²/s⁴]
loc_sigma_2	Location measurement noise variance [m²]
vel_sigma_2	Velocity process noise variance [m²/s²]

~MLM()

MLM::~MLM

(

void

)

Destructor.

Member Function Documentation

predicted_coordinate()

gps_coordinate MLM::predicted_coordinate

(

)

const

Get current estimated position.

Returns

Predicted GPS coordinate based on current filter state

process_acc_data()

bool MLM::process_acc_data	(	const enu_accelerometer &	acc,
		double	time_sec
	)

Process accelerometer measurement (prediction step)

Propagates the filter state forward using accelerometer data as control input. This performs the prediction step of the Kalman filter.

Parameters

acc	ENU-frame accelerometer measurement
time_sec	Timestamp in seconds

Returns

true if processing successful, false otherwise

process_gps_data()

void MLM::process_gps_data	(	const gps_coordinate &	gps,
		double	time_sec
	)

Process GPS measurement (update step)

Corrects the predicted state using GPS position measurement. This performs the update/correction step of the Kalman filter.

Parameters

gps	GPS coordinate measurement
time_sec	Timestamp in seconds

Member Data Documentation

m_acc_sigma_2

double MLM::m_acc_sigma_2

private

Accelerometer noise variance [m²/s⁴].

m_fk

gps_acc_fusion_filter MLM::m_fk

private

Kalman filter for GPS-accelerometer fusion.

m_got_start_point

bool MLM::m_got_start_point

private

Flag indicating if initial reference point is set.

m_lc

GeographicLib::LocalCartesian MLM::m_lc

private

Local cartesian coordinate converter.

m_loc_sigma_2

double MLM::m_loc_sigma_2

private

Location measurement noise variance [m²].

m_vel_sigma_2

double MLM::m_vel_sigma_2

private

Velocity process noise variance [m²/s²].

---

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

• inc/mlm.h
• src/mlm.cpp