File linearization.h

namespace sym

Typedefs

template<typename Scalar>
using SparseLinearization = Linearization<Eigen::SparseMatrix<Scalar>>

using SparseLinearizationd = SparseLinearization<double>

using SparseLinearizationf = SparseLinearization<float>

template<typename Scalar>
using DenseLinearization = Linearization<MatrixX<Scalar>>

using DenseLinearizationd = DenseLinearization<double>

using DenseLinearizationf = DenseLinearization<float>

Functions

template<typename Scalar>
sparse_matrix_structure_t GetSparseStructure(const Eigen::SparseMatrix<Scalar> &matrix)

Returns the sparse matrix structure of matrix.

template<typename Derived>
sparse_matrix_structure_t GetSparseStructure(const Eigen::EigenBase<Derived>&)

Return a default initialized sparse structure because arg is dense.

template<typename Scalar>
Eigen::Map<const VectorX<Scalar>> JacobianValues(const Eigen::SparseMatrix<Scalar> &matrix)

Returns coefficients of matrix. Overloads exist for both dense and sparse matrices to make writing generic code easier.

This version returns the non-zero values of an Eigen::SparseMatrix

Note: it returns a map, so be careful about mutating or disposing of matrix before you are finished with the output.

template<typename Scalar>
const MatrixX<Scalar> &JacobianValues(const MatrixX<Scalar> &matrix)

Returns coefficients of matrix. Overloads exist for both dense and sparse matrices to make writing generic code easier.

Returns a const-ref to the argument.

template<typename MatrixType>
struct Linearization

#include <linearization.h>

Class for storing a problem linearization evaluated at a Values (i.e. a residual, jacobian, hessian, and rhs)

MatrixType is expected to be an Eigen MatrixX or SparseMatrix.

Public Types

using Scalar = typename MatrixType::Scalar

using Vector = VectorX<Scalar>

using Matrix = MatrixType

Public Functions

inline void Reset()

Set to invalid

inline bool IsInitialized() const

Returns whether the linearization is currently valid for the corresponding values. Accessing any of the members when this is false could result in unexpected behavior

inline void SetInitialized(const bool initialized = true)

inline double Error() const

inline double LinearError(const Vector &x_update) const

Public Members

Vector residual

Matrix hessian_lower

Matrix jacobian

Vector rhs

Private Members

bool initialized_ = {false}