dynamicgraph::sot::FilterDifferentiator Class Reference

#include <sot/core/filter-differentiator.hh>

Inheritance diagram for dynamicgraph::sot::FilterDifferentiator:

Public Member Functions
	DECLARE_SIGNAL_IN (x, dynamicgraph::Vector)
	Input signals. More...
	DECLARE_SIGNAL_OUT (x_filtered, dynamicgraph::Vector)
	Output signal x_filtered. More...
	DECLARE_SIGNAL_OUT (dx, dynamicgraph::Vector)
	DECLARE_SIGNAL_OUT (ddx, dynamicgraph::Vector)
	DECLARE_SIGNAL_INNER (x_dx_ddx, dynamicgraph::Vector)
	The following inner signals are used because this entity has some output signals whose related quantities are computed at the same time by the same algorithm To avoid the risk of recomputing the same things twice, we create an inner signal that groups together all the quantities that are computed together. More...
EIGEN_MAKE_ALIGNED_OPERATOR_NEW	FilterDifferentiator (const std::string &name)
	— CONSTRUCTOR —- More...
void	init (const double &timestep, const int &xSize, const Eigen::VectorXd &filter_numerator, const Eigen::VectorXd &filter_denominator)
	Initialize the FilterDifferentiator. More...
void	switch_filter (const Eigen::VectorXd &filter_numerator, const Eigen::VectorXd &filter_denominator)
virtual void	display (std::ostream &os) const

Protected Attributes
double	m_dt
int	m_x_size
	sampling timestep of the input signal More...
CausalFilter *	m_filter
	polynomial-fitting filters More...

Constructor & Destructor Documentation

FilterDifferentiator()

EIGEN_MAKE_ALIGNED_OPERATOR_NEW dynamicgraph::sot::FilterDifferentiator::FilterDifferentiator

(

const std::string &

name

)

— CONSTRUCTOR —-

Member Function Documentation

DECLARE_SIGNAL_IN()

dynamicgraph::sot::FilterDifferentiator::DECLARE_SIGNAL_IN	(	x	,
		dynamicgraph::Vector
	)

Input signals.

DECLARE_SIGNAL_INNER()

dynamicgraph::sot::FilterDifferentiator::DECLARE_SIGNAL_INNER	(	x_dx_ddx	,
		dynamicgraph::Vector
	)

The following inner signals are used because this entity has some output signals whose related quantities are computed at the same time by the same algorithm To avoid the risk of recomputing the same things twice, we create an inner signal that groups together all the quantities that are computed together.

Then the single output signals will depend on this inner signal, which is the one triggering the computations. Inner signals are not exposed, so that nobody can access them. This signal contains the estimated positions, velocities and accelerations.

DECLARE_SIGNAL_OUT() [1/3]

dynamicgraph::sot::FilterDifferentiator::DECLARE_SIGNAL_OUT	(	x_filtered	,
		dynamicgraph::Vector
	)

Output signal x_filtered.

DECLARE_SIGNAL_OUT() [2/3]

dynamicgraph::sot::FilterDifferentiator::DECLARE_SIGNAL_OUT	(	dx	,
		dynamicgraph::Vector
	)

DECLARE_SIGNAL_OUT() [3/3]

dynamicgraph::sot::FilterDifferentiator::DECLARE_SIGNAL_OUT	(	ddx	,
		dynamicgraph::Vector
	)

display()

virtual void dynamicgraph::sot::FilterDifferentiator::display ( std::ostream &  os ) const

virtual

init()

void dynamicgraph::sot::FilterDifferentiator::init	(	const double &	timestep,
		const int &	xSize,
		const Eigen::VectorXd &	filter_numerator,
		const Eigen::VectorXd &	filter_denominator
	)

Initialize the FilterDifferentiator.

Parameters

timestep	Period (in seconds) after which the sensors' data are updated.
sigSize	Size of the input signal.
delay	Delay (in seconds) introduced by the estimation. This should be a multiple of timestep.

Note

The estimationDelay is half of the length of the window used for the polynomial fitting. The larger the delay, the smoother the estimations.

switch_filter()

void dynamicgraph::sot::FilterDifferentiator::switch_filter	(	const Eigen::VectorXd &	filter_numerator,
		const Eigen::VectorXd &	filter_denominator
	)

Member Data Documentation

m_dt

double dynamicgraph::sot::FilterDifferentiator::m_dt

protected

m_filter

CausalFilter* dynamicgraph::sot::FilterDifferentiator::m_filter

protected

polynomial-fitting filters

m_x_size

int dynamicgraph::sot::FilterDifferentiator::m_x_size

protected

sampling timestep of the input signal