Class EKF

Defined in File ekf.hpp

Class Documentation

class EKF

Public Functions

EKF(): Constructor

void set_parameters(const Motion_Noise &motion_noise, const Measurement_Noise &measurement_noise, const Controller &_controller): Sets the parameters required by the EKF.

Matrix3d covariance_to_eigen(const double covariance[16]) const: Converts a covariance given by a double array to an Eigen matrix.

void eigen_to_covariance(const Matrix3d &covariance_eigen, double covariance[16]) const: Converts a covariance given by an Eigen matrix to a double array.

void get_motion_jacobi(const State &state, const Control &control, double integration_step, Matrix3d &F_x, Matrix32d &F_u) const: Computes the Jacobians of the motion equations with respect to the state and control.

Matrix3d get_observation_jacobi() const: Computes the Jacobian of the observation equations with respect to the state.

Matrix2d get_motion_covariance(const State &state, const Control &control, double integration_step) const: Returns the motion covariance in control space.

Matrix3d get_observation_covariance() const: Returns the observation covariance.

Matrix23d get_controller_gain(const Control &control) const: Returns the gain of the controller.

Matrix3d get_rotation_matrix(double angle) const: Returns the rotation matrix from a global frame to a local frame.

void predict(const State_With_Covariance &state, const Control &control, double integration_step, State_With_Covariance &state_pred) const: Predicts the covariances based on the paper: Rapidly-exploring random belief trees for motion planning under uncertainty, A. Bry and N. Roy, IEEE ICRA 2011.

void update(const State_With_Covariance &state_pred, State_With_Covariance &state_corr) const: Predicts the covariances.

Public Members

EIGEN_MAKE_ALIGNED_OPERATOR_NEW: Overload operator new for fixed-size vectorizable Eigen member variable.