Template Class ResidualModelNumDiffTpl

Defined in File residual.hpp

Inheritance Relationships

Base Type

public crocoddyl::ResidualModelAbstractTpl< _Scalar > (Template Class ResidualModelAbstractTpl)

Class Documentation

template<typename _Scalar> class ResidualModelNumDiffTpl : public crocoddyl::ResidualModelAbstractTpl<_Scalar>

This class computes the numerical differentiation of a residual model.

It computes the Jacobian of the residual model via numerical differentiation, i.e., \(\mathbf{R_x}\) and \(\mathbf{R_u}\) which denote the Jacobians of the residual function \(\mathbf{r}(\mathbf{x},\mathbf{u})\).

See also

ResidualModelAbstractTpl(), calcDiff()

Public Types

typedef ResidualDataAbstractTpl<Scalar> ResidualDataAbstract

typedef ResidualModelAbstractTpl<Scalar> Base

typedef ResidualDataNumDiffTpl<Scalar> Data

typedef DataCollectorAbstractTpl<Scalar> DataCollectorAbstract

typedef MathBaseTpl<Scalar> MathBase

typedef MathBaseTpl<Scalar>::VectorXs VectorXs

typedef MathBaseTpl<Scalar>::MatrixXs MatrixXs

typedef boost::function<void(const VectorXs&, const VectorXs&)> ReevaluationFunction

Public Functions

explicit ResidualModelNumDiffTpl(const std::shared_ptr<Base> &model)

Initialize the numdiff residual model.

Parameters:: model – Residual model that we want to apply the numerical differentiation

virtual ~ResidualModelNumDiffTpl() = default: Initialize the numdiff residual model.

virtual void calc(const std::shared_ptr<ResidualDataAbstract> &data, const Eigen::Ref<const VectorXs> &x, const Eigen::Ref<const VectorXs> &u) override

Compute the residual vector.

Parameters:

data – [in] Residual data
x – [in] State point \(\mathbf{x}\in\mathbb{R}^{ndx}\)
u – [in] Control input \(\mathbf{u}\in\mathbb{R}^{nu}\)

virtual void calc(const std::shared_ptr<ResidualDataAbstract> &data, const Eigen::Ref<const VectorXs> &x) override

virtual void calcDiff(const std::shared_ptr<ResidualDataAbstract> &data, const Eigen::Ref<const VectorXs> &x, const Eigen::Ref<const VectorXs> &u) override

Compute the Jacobian of the residual vector.

It computes the Jacobian the residual function. It assumes that calc() has been run first.

Parameters:

data – [in] Residual data
x – [in] State point \(\mathbf{x}\in\mathbb{R}^{ndx}\)
u – [in] Control input \(\mathbf{u}\in\mathbb{R}^{nu}\)

virtual void calcDiff(const std::shared_ptr<ResidualDataAbstract> &data, const Eigen::Ref<const VectorXs> &x) override

virtual std::shared_ptr<ResidualDataAbstract> createData(DataCollectorAbstract *const data) override

Create the residual data.

The default data contains objects to store the values of the residual vector and their Jacobians. However, it is possible to specialize this function if we need to create additional data, for instance, to avoid dynamic memory allocation.

Parameters:: data – Data collector
Returns:: the residual data

template<typename NewScalar> ResidualModelNumDiffTpl<NewScalar> cast() const

const std::shared_ptr<Base> &get_model() const: Return the original residual model.

const Scalar get_disturbance() const: Return the disturbance constant used by the numerical differentiation routine.

void set_disturbance(const Scalar disturbance): Modify the disturbance constant used by the numerical differentiation routine.

void set_reevals(const std::vector<ReevaluationFunction> &reevals)

Register functions that updates the shared data computed for a system rollout The updated data is used to evaluate of the gradient and Hessian.

Parameters:: reevals – are the registered functions.

Public Members

EIGEN_MAKE_ALIGNED_OPERATOR_NEW typedef _Scalar Scalar

Protected Attributes

std::size_t nu_: Control dimension.

std::shared_ptr<StateAbstract> state_: State description.

VectorXs unone_: No control vector.