Class nHQP::TaskData

Defined in File nHQP.h

Nested Relationships

This class is a nested type of Class nHQP.

Class Documentation

class TaskData

The TaskData class is a helper class containing all that is needed to setup, regularize, and solve the i-th QP problem.

Public Functions

TaskData(int num_free_vars, TaskPtr task, ConstraintPtr constraint, BackEnd::Ptr back_end)

void set_min_sv_ratio(double sv)

int compute_nullspace_dimension(double threshold)

void set_nullspace_dimension(int ns_dim)

void compute_cost(const Eigen::MatrixXd *AN_nullspace, const Eigen::VectorXd &q0)

void compute_contraints(const Eigen::MatrixXd *AN_nullspace, const Eigen::VectorXd &q0)

bool update_and_solve()

bool compute_nullspace()

const Eigen::MatrixXd &get_nullspace() const

const Eigen::VectorXd &get_solution() const

bool enable_logger(XBot::MatLogger2::Ptr logger, std::string log_prefix)

void set_perform_A_b_regularization(bool perform_A_b_regularization_)

skip_A_b_regularization skip the regularization of the A matrix and b vector based on the method void regularize_A_b(double threshold).

This is particularly suggested when a single layer is used together with a joint-space task. Regularization in this case is not needed and usually perform badly.

Parameters: perform_A_b_regularization – true or false, default true

void set_perform_selective_null_space_regularization(bool perform_selective_null_space_regularization_)

set_perform_selective_null_space_regularization regularization of the Hessian matrix using max singular value and null_space

Parameters: perform_selective_null_space_regularization_ – true or false, default true