#include <gndoep.h>

Inheritance diagram for gcop::GnDoep< T, _nx, _nu, _np, _ng, Tz, _nz, T1, _nx1 >:

Collaboration diagram for gcop::GnDoep< T, _nx, _nu, _np, _ng, Tz, _nz, T1, _nx1 >:

Public Member Functions
	GnDoep (System< T, _nx, _nu, _np > &sys, Sensor< T1, _nx1, _nu, _np, Tz, _nz > &sensor, LqSensorCost< T, _nx, _nu, _np, _ng, Tz, _nz > &cost, vector< double > &ts, vector< T > &xs, vector< Vectorcd > &us, Vectormd &p, vector< double > &ts1, Func_type _project=NULL, bool update=true)
virtual	~GnDoep ()
void	Iterate ()
Public Attributes
int	info
double	fnorm
double	covfac
int	inputs
int	values
VectorXd	s
	optimization vector
GnCost< T, _nx, _nu, _np, _ng, Tz, _nz, T1, _nx1 > *	functor
NumericalDiff< GnCost< T, _nx, _nu, _np, _ng, Tz, _nz, T1, _nx1 > > *	numDiff
LevenbergMarquardt < NumericalDiff< GnCost< T, _nx, _nu, _np, _ng, Tz, _nz, T1, _nx1 > > > *	lm

template<typename T = VectorXd, int _nx = Dynamic, int _nu = Dynamic, int _np = Dynamic, int _ng = Dynamic, typename Tz = VectorXd, int _nz = Dynamic, typename T1 = T, int _nx1 = _nx>
class gcop::GnDoep< T, _nx, _nu, _np, _ng, Tz, _nz, T1, _nx1 >

Constructor & Destructor Documentation

template<typename T , int _nx, int _nu, int _np, int _ng, typename Tz , int _nz, typename T1 , int _nx1>

gcop::GnDoep< T, _nx, _nu, _np, _ng, Tz, _nz, T1, _nx1 >::GnDoep	(	System< T, _nx, _nu, _np > &	sys,
		Sensor< T1, _nx1, _nu, _np, Tz, _nz > &	sensor,
		LqSensorCost< T, _nx, _nu, _np, _ng, Tz, _nz > &	cost,
		vector< double > &	ts,
		vector< T > &	xs,
		vector< Vectorcd > &	us,
		Vectormd &	p,
		vector< double > &	ts1,
		Func_type	_project = `NULL`,
		bool	update = `true`
	)

Create an optimal estimation problem using a system, sensor, a cost, and sensor measurements and parameters of the system.

The optimal estimation is done over selecting the right parameters for the system and also process noise to ensure the estimation cost on sensor measurements and noise is minimized Parameters can also have a prior which is added into the cost function. The parameters passed in here are used as initial guess to the system

You should set the ts[1:N+1], xs[0]; Known controls us

After initialization, every call to Iterate() will optimize the parameters of the system, process noise and modify them accordingly.

Parameters:

sys	system
cost	cost
tparam	trajectory parametrization
ts	(N+1) sequence of discrete times
xs	(N+1) sequence of discrete states
us	(N) sequence of control inputs
p	static parameter vector initial guess of it
ng
_project	Optional function to project T into T1. Not needed if T == T1
update	whether to update trajectory xs using initial state xs[0] and inputs us. This is necessary only if xs was not already generated from us.

References gcop::GnDoep< T, _nx, _nu, _np, _ng, Tz, _nz, T1, _nx1 >::inputs, and gcop::GnDoep< T, _nx, _nu, _np, _ng, Tz, _nz, T1, _nx1 >::values.

template<typename T , int _nx, int _nu, int _np, int _ng, typename Tz , int _nz, typename T1 , int _nx1>

gcop::GnDoep< T, _nx, _nu, _np, _ng, Tz, _nz, T1, _nx1 >::~GnDoep ( ) [virtual]

Member Function Documentation

template<typename T , int _nx, int _nu, int _np, int _ng, typename Tz , int _nz, typename T1 , int _nx1>

void gcop::GnDoep< T, _nx, _nu, _np, _ng, Tz, _nz, T1, _nx1 >::Iterate ( ) [virtual]

Perform one DOCP iteration. Internally calls: are updated.

Reimplemented from gcop::Doep< T, _nx, _nu, _np, Tz, _nz, T1, _nx1 >.

References gcop::GnCost< T, _nx, _nu, _np, _ng, _ntp >::doep, and np.