MRTKLIB Matrix Module — Matrix/vector operations and estimation. More...

#include <stdio.h>
#include "mrtklib/mrtk_foundation.h"

Include dependency graph for mrtk_mat.h:

This graph shows which files directly or indirectly include this file:

Typedefs
typedef void	fatalfunc_t(const char *)
	Fatal callback function type for allocation failures.

Functions
void	add_fatal (fatalfunc_t *func)
	Add fatal callback function for mat(), zeros(), imat().
double *	mat (int n, int m)
	Allocate memory for a double matrix.
int *	imat (int n, int m)
	Allocate memory for an integer matrix.
double *	zeros (int n, int m)
	Allocate a new zero matrix.
double *	eye (int n)
	Generate a new identity matrix.
double	dot (const double a, const double b, int n)
	Inner product of vectors.
double	norm (const double *a, int n)
	Euclidean norm of vector.
void	cross3 (const double a, const double b, double *c)
	Cross product of 3D vectors.
int	normv3 (const double a, double b)
	Normalize 3D vector.
void	matcpy (double A, const double B, int n, int m)
	Copy matrix.
void	matmul (const char tr, int n, int k, int m, double alpha, const double A, const double B, double beta, double C)
	Multiply matrix by matrix (C = alphaAB + beta*C).
int	matinv (double *A, int n)
	Inverse of matrix (A = A^-1).
int	solve (const char tr, const double A, const double Y, int n, int m, double X)
	Solve linear equation (X = A\Y or X = A'\Y).
int	lsq (const double A, const double y, int n, int m, double x, double Q)
	Least square estimation by solving normal equation.
int	filter (double x, double P, const double H, const double v, const double *R, int n, int m)
	Kalman filter state update.
int	smoother (const double xf, const double Qf, const double xb, const double Qb, int n, double xs, double Qs)
	Combine forward and backward filters by fixed-interval smoother.
void	matfprint (const double A, int n, int m, int p, int q, FILE fp)
	Print matrix to file pointer.
void	matprint (const double *A, int n, int m, int p, int q)
	Print matrix to stdout.

Detailed Description

MRTKLIB Matrix Module — Matrix/vector operations and estimation.

This header provides matrix and vector allocation, arithmetic, linear algebra (LU decomposition, matrix inversion, linear solve), and estimation algorithms (least squares, Kalman filter, smoother).

Note: All functions in this module are pure cut-and-paste extractions from rtkcmn.c with zero algorithmic changes.

Definition in file mrtk_mat.h.

Typedef Documentation

◆ fatalfunc_t

typedef void fatalfunc_t(const char *)

Fatal callback function type for allocation failures.

Definition at line 43 of file mrtk_mat.h.

Function Documentation

◆ add_fatal()

void add_fatal ( fatalfunc_t * func )

Add fatal callback function for mat(), zeros(), imat().

Parameters

[in] func Callback function

Definition at line 65 of file mrtk_mat.c.

References fatalfunc.

◆ cross3()

void cross3	(	const double *	a,
		const double *	b,
		double *	c )

Cross product of 3D vectors.

Parameters

[in]	a	Vector a (3 x 1)
[in]	b	Vector b (3 x 1)
[out]	c	Cross product a x b (3 x 1)

Definition at line 172 of file mrtk_mat.c.

Referenced by clas_osr_adjust_r_dts(), clas_osr_satcorr_update(), plane(), satantoff(), satpos_ssr(), and windupcorr().

Here is the caller graph for this function:

◆ dot()

double dot	(	const double *	a,
		const double *	b,
		int	n )

Inner product of vectors.

Parameters

[in]	a	Vector a (n x 1)
[in]	b	Vector b (n x 1)
[in]	n	Size of vectors

Returns: a' * b

Definition at line 151 of file mrtk_mat.c.

Referenced by antmodel(), clas_osr_satcorr_update(), clas_stec_grid_data(), clas_trop_grid_data(), deq(), ecef2pos(), find_grid_surround_pos(), norm(), ocean_tide_clasgrid(), peph2pos(), resdop(), satazel(), satpos_ssr(), tide_pl(), toml_rtod_ex(), valsol(), and windupcorr().

Here is the caller graph for this function:

◆ eye()

double * eye ( int n )

Generate a new identity matrix.

Parameters

[in] n Number of rows and columns

Returns: Matrix pointer (NULL if n<=0)

Definition at line 129 of file mrtk_mat.c.

References zeros().

Referenced by filter2_(), filter2_(), filter_(), lambda(), udpos(), udpos(), and udpos_ppp().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ filter()

int filter	(	double *	x,
		double *	P,
		const double *	H,
		const double *	v,
		const double *	R,
		int	n,
		int	m )

Kalman filter state update.

K=P*H*(H'*P*H+R)^-1, xp=x+K*v, Pp=(I-K*H')*P

Parameters

[in,out]	x	States vector (n x 1)
[in,out]	P	Covariance matrix of states (n x n)
[in]	H	Transpose of design matrix (n x m)
[in]	v	Innovation (measurement - model) (m x 1)
[in]	R	Covariance matrix of measurement error (m x m)
[in]	n	Number of states
[in]	m	Number of measurements

Returns: Status (0:ok, <0:error)

Note: Matrix stored by column-major order (Fortran convention). If state x[i]==0.0, does not update state x[i]/P[i+i*n].

Definition at line 524 of file mrtk_mat.c.

References filter_(), free, imat(), and mat().

Referenced by holdamb(), holdamb(), holdamb(), and relpos().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ imat()

int * imat	(	int	n,
		int	m )

Allocate memory for an integer matrix.

Parameters

[in]	n	Number of rows
[in]	m	Number of columns

Returns: Matrix pointer (NULL if n<=0 or m<=0)

Definition at line 92 of file mrtk_mat.c.

References fatalerr(), and malloc.

Referenced by clas_trop_grid_data(), filter(), filter2(), filter2(), filter2_(), filter2_(), matinv(), raim_fde(), resamb_LAMBDA(), and udpos().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ lsq()

int lsq	(	const double *	A,
		const double *	y,
		int	n,
		int	m,
		double *	x,
		double *	Q )

Least square estimation by solving normal equation.

Solves x = (A*A')^-1 * A * y.

Parameters

[in]	A	Transpose of (weighted) design matrix (n x m)
[in]	y	(Weighted) measurements (m x 1)
[in]	n	Number of parameters
[in]	m	Number of measurements (n <= m)
[out]	x	Estimated parameters (n x 1)
[out]	Q	Estimated parameters covariance matrix (n x n)

Returns: Status (0:ok, <0:error)

Note: For weighted least square, replace A and y by A*w and w*y (w=W^(1/2)). Matrix stored by column-major order (Fortran convention).

Definition at line 470 of file mrtk_mat.c.

References free, mat(), matinv(), and matmul().

Referenced by estpos(), and estvel().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ mat()

double * mat	(	int	n,
		int	m )

Allocate memory for a double matrix.

Parameters

[in]	n	Number of rows
[in]	m	Number of columns

Returns: Matrix pointer (NULL if n<=0 or m<=0)

Definition at line 76 of file mrtk_mat.c.

References fatalerr(), and malloc.

Referenced by calc_3points_weight(), clas_get_grid_index(), clas_ssr2osr(), clas_stec_grid_data(), clas_trop_grid_data(), ddres(), ddres(), ddres(), estpos(), estvel(), filter(), filter2(), filter2(), filter2_(), filter2_(), filter_(), holdamb(), holdamb(), holdamb(), lambda(), lambda_reduction(), lambda_search(), LD(), lsq(), ludcmp(), matinv(), pntpos(), ppp_rtk_pos(), raim_fde(), relpos(), relposvrs(), resamb_LAMBDA(), resamb_LAMBDA(), resamb_LAMBDA(), sddop(), sddop(), search(), smoother(), solve(), udpos(), udpos(), udpos_ppp(), zddop(), zddop(), and zeros().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ matcpy()

void matcpy	(	double *	A,
		const double *	B,
		int	n,
		int	m )

Copy matrix.

Parameters

[out]	A	Destination matrix (n x m)
[in]	B	Source matrix (n x m)
[in]	n	Number of rows
[in]	m	Number of columns

Definition at line 205 of file mrtk_mat.c.

Referenced by clas_get_grid_index(), decode_bdsraw(), decode_bnx_00(), decode_cnav(), decode_enav(), decode_galrawfnav(), decode_galrawinav(), decode_glorawca(), decode_gnav(), decode_ionutc(), decode_ionutc(), decode_L1ionutc(), decode_navicraw(), decode_qzssrawsubframeb(), estvel(), filter2_(), filter2_(), filter_(), matinv(), pntpos(), ppp_rtk_pos(), raim_fde(), raw2rtcm(), readpcv(), relpos(), relposvrs(), send_nmea(), setopt_apppos(), setopt_sta(), smoother(), solve(), strsvrthread(), udpos(), udpos_ppp(), and update_ionutc().

Here is the caller graph for this function:

◆ matfprint()

void matfprint	(	const double *	A,
		int	n,
		int	m,
		int	p,
		int	q,
		FILE *	fp )

Print matrix to file pointer.

Parameters

[in]	A	Matrix A (n x m)
[in]	n,m	Number of rows and columns
[in]	p,q	Total columns, columns under decimal point
[in]	fp	Output file pointer

Note: Matrix stored by column-major order (Fortran convention).

◆ matinv()

int matinv	(	double *	A,
		int	n )

Inverse of matrix (A = A^-1).

Parameters

[in,out]	A	Matrix (n x n)
[in]	n	Size of matrix

Returns: Status (0:ok, <0:error)

Definition at line 421 of file mrtk_mat.c.

References free, imat(), lubksb(), ludcmp(), mat(), and matcpy().

Referenced by calc_3points_weight(), clas_get_grid_index(), dops(), filter2_(), filter2_(), filter_(), lsq(), resamb_LAMBDA(), resamb_LAMBDA(), resamb_LAMBDA(), sddop(), sddop(), smoother(), and solve().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ matmul()

void matmul	(	const char *	tr,
		int	n,
		int	k,
		int	m,
		double	alpha,
		const double *	A,
		const double *	B,
		double	beta,
		double *	C )

Multiply matrix by matrix (C = alpha*A*B + beta*C).

Parameters

[in]	tr	Transpose flags ("N":normal, "T":transpose)
[in]	n,k,m	Size of (transposed) matrix A, B
[in]	alpha	Alpha coefficient
[in]	A,B	(Transposed) matrix A (n x m), B (m x k)
[in]	beta	Beta coefficient
[in,out]	C	Matrix C (n x k)

Definition at line 310 of file mrtk_mat.c.

Referenced by calc_3points_weight(), clas_stec_grid_data(), clas_trop_grid_data(), covecef(), covenu(), dops(), ecef2enu(), eci2ecef(), enu2ecef(), filter2_(), filter2_(), filter_(), lambda(), lsq(), ocean_tide_clasgrid(), resamb_LAMBDA(), resamb_LAMBDA(), resamb_LAMBDA(), resdop(), sddop(), sddop(), smoother(), solve(), sunmoonpos(), tidedisp(), udpos(), udpos(), and udpos_ppp().

Here is the caller graph for this function:

◆ matprint()

void matprint	(	const double *	A,
		int	n,
		int	m,
		int	p,
		int	q )

Print matrix to stdout.

Parameters

[in]	A	Matrix A (n x m)
[in]	n,m	Number of rows and columns
[in]	p,q	Total columns, columns under decimal point

◆ norm()

double norm	(	const double *	a,
		int	n )

Euclidean norm of vector.

Parameters

[in]	a	Vector a (n x 1)
[in]	n	Size of vector

Returns: || a ||

Definition at line 165 of file mrtk_mat.c.

References dot().

Referenced by antpos(), baseline(), baseline(), baseline_len(), clas_osr_satcorr_update(), clas_osr_zdres(), clas_ssr2osr(), combpeph(), constbl(), convrnx_s(), estpos(), estvel(), find_grid_surround_pos(), geodist(), ionmodel(), ionocorr(), miono_sel_area(), normv3(), out_iono_sys(), out_leaps(), out_time_sys(), outrnxevent(), outsols(), pcv_slot_has_data(), pephpos(), pickup_candidates_of_grid(), pntpos(), posdist(), ppp_rtk_pos(), readpcv(), relposvrs(), resdop(), satpos_ssr(), send_nmea(), setopt_sta(), setpcv(), shapiro(), tide_pl(), tidedisp(), udpos(), udpos(), udpos_ppp(), windupcorr(), zdres(), and zdres().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ normv3()

int normv3	(	const double *	a,
		double *	b )

Normalize 3D vector.

Parameters

[in]	a	Vector a (3 x 1)
[out]	b	Normalized vector (3 x 1), \|\| b \|\| = 1

Returns: Status (1:ok, 0:error)

Definition at line 183 of file mrtk_mat.c.

References norm().

Referenced by clas_osr_adjust_r_dts(), clas_osr_satcorr_update(), plane(), satantoff(), satpos_ssr(), and windupcorr().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ smoother()

int smoother	(	const double *	xf,
		const double *	Qf,
		const double *	xb,
		const double *	Qb,
		int	n,
		double *	xs,
		double *	Qs )

Combine forward and backward filters by fixed-interval smoother.

xs=Qs*(Qf^-1*xf+Qb^-1*xb), Qs=(Qf^-1+Qb^-1)^-1

Parameters

[in]	xf	Forward solutions (n x 1)
[in]	Qf	Forward solutions covariance matrix (n x n)
[in]	xb	Backward solutions (n x 1)
[in]	Qb	Backward solutions covariance matrix (n x n)
[in]	n	Number of solutions
[out]	xs	Smoothed solutions (n x 1)
[out]	Qs	Smoothed solutions covariance matrix (n x n)

Returns: Status (0:ok, <0:error)

Note: Matrix stored by column-major order (Fortran convention).

Definition at line 579 of file mrtk_mat.c.

References free, mat(), matcpy(), matinv(), and matmul().

Referenced by combres().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ solve()

int solve	(	const char *	tr,
		const double *	A,
		const double *	Y,
		int	n,
		int	m,
		double *	X )

Solve linear equation (X = A\Y or X = A'\Y).

Parameters

[in]	tr	Transpose flag ("N":normal, "T":transpose)
[in]	A	Input matrix A (n x n)
[in]	Y	Input matrix Y (n x m)
[in]	n,m	Size of matrix A, Y
[out]	X	Solution X = A\Y or X = A'\Y (n x m)

Returns: Status (0:ok, <0:error)

Note: Matrix stored by column-major order (Fortran convention). X can be same as Y.

Definition at line 443 of file mrtk_mat.c.

References free, mat(), matcpy(), matinv(), and matmul().

Referenced by lambda().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ zeros()

double * zeros	(	int	n,
		int	m )

Allocate a new zero matrix.

Parameters

[in]	n	Number of rows
[in]	m	Number of columns

Returns: Matrix pointer (NULL if n<=0 or m<=0)

Definition at line 108 of file mrtk_mat.c.

References calloc, fatalerr(), and mat().

Referenced by clas_ssr2osr(), eye(), filter2(), holdamb(), holdamb(), holdamb(), lambda(), lambda_reduction(), lambda_search(), pntpos(), ppp_rtk_pos(), raim_fde(), relpos(), relposvrs(), resamb_LAMBDA(), resamb_LAMBDA(), residual_test(), residual_test(), rtkinit(), search(), set_init_pb(), set_init_pb(), udbias(), udbias_ppp(), and udbias_ppp().

Here is the call graph for this function:

Here is the caller graph for this function:

Typedefs

Functions

Detailed Description

Typedef Documentation

◆ fatalfunc_t

Function Documentation

◆ add_fatal()

◆ cross3()

◆ dot()

◆ eye()

◆ filter()

◆ imat()

◆ lsq()

◆ mat()

◆ matcpy()

◆ matfprint()

◆ matinv()

◆ matmul()

◆ matprint()

◆ norm()

◆ normv3()

◆ smoother()

◆ solve()

◆ zeros()