JCBB Class Reference

Data association implementation that uses the Malhanobis Distance only as criterion to make observation matches. More...

#include <jcbb.hpp>

Inheritance diagram for JCBB:

Collaboration diagram for JCBB:

Public Member Functions
	JCBB (const DataAssociationParameters &params)
	~JCBB ()=default
Eigen::VectorXi	associate (const Eigen::VectorXd &landmarks, const Eigen::VectorXd &observations, const Eigen::MatrixXd &covariance, const Eigen::VectorXd &observation_confidences) const override
	Perform data association using the Joint Compatibility Branch & Bound algorithm.
Public Member Functions inherited from DataAssociationModel
virtual int	associate_n_filter (const std::vector< common_lib::structures::Cone > &perception_map, Eigen::VectorXf &_x_vector_, Eigen::MatrixXf &_p_matrix_, std::vector< int > &matched_ids, std::vector< Eigen::Vector2f > &matched_cone_positions, std::vector< Eigen::Vector2f > &new_features, ObservationModel *observation_model) const =0
	Associate the observed landmarks to the expected landmarks and update the state vector and the covariance matrix.
	DataAssociationModel (float max_landmark_distance)
virtual	~DataAssociationModel ()=default
	DataAssociationModel ()=default
	DataAssociationModel (DataAssociationParameters params)
virtual	~DataAssociationModel ()=default

Private Member Functions
void	search_branch_and_bound (int current_obs_idx, int num_observations, int num_landmarks, const Eigen::MatrixXd &distances, Eigen::VectorXi &current_hypothesis, Eigen::VectorXi &best_hypothesis, double &best_score, double current_score) const
	Recursive branch and bound search to find the best hypothesis.
bool	is_landmark_already_assigned (const Eigen::VectorXi &hypothesis, int landmark_idx) const
	Check if a landmark is already assigned in the current hypothesis.

Additional Inherited Members
Protected Member Functions inherited from DataAssociationModel
float	get_max_landmark_distance () const
Protected Attributes inherited from DataAssociationModel
DataAssociationParameters	_params_

Detailed Description

Data association implementation that uses the Malhanobis Distance only as criterion to make observation matches.

Definition at line 17 of file jcbb.hpp.

Constructor & Destructor Documentation

JCBB()

JCBB::JCBB ( const DataAssociationParameters &  params )

inline

Definition at line 45 of file jcbb.hpp.

~JCBB()

JCBB::~JCBB ( )

default

Member Function Documentation

associate()

Eigen::VectorXi JCBB::associate ( const Eigen::VectorXd &  landmarks, const Eigen::VectorXd &  observations, const Eigen::MatrixXd &  covariance, const Eigen::VectorXd &  observation_confidences  ) const

overridevirtual

Perform data association using the Joint Compatibility Branch & Bound algorithm.

This function finds the most consistent assignment of 2D observations to known 2D landmarks by maximizing the number of mutually compatible pairs. Uses Euclidean distance as the compatibility measure and a branch & bound search to efficiently explore the space of hypotheses.

Parameters

landmark_positions	Landmark positions in format [x1, y1, x2, y2, ...].
observation_positions	Observation positions in format [x1, y1, x2, y2, ...].
covariance	Covariance matrix of state estimate (unused in this version).
observation_confidences	Confidences for each observation (unused in this version).

Returns

Eigen::VectorXi of size num_observations:

• If element i >= 0 → index of associated landmark.
• If element i == -1 → observation is unassigned (new/clutter).

Implements DataAssociationModel.

Definition at line 3 of file jcbb.cpp.

Here is the call graph for this function:

is_landmark_already_assigned()

bool JCBB::is_landmark_already_assigned ( const Eigen::VectorXi &  hypothesis, int  landmark_idx  ) const

private

Check if a landmark is already assigned in the current hypothesis.

Parameters

hypothesis	Current hypothesis vector.
landmark_idx	Index of the landmark to check.

Returns

true if landmark is already assigned, false otherwise.

Definition at line 84 of file jcbb.cpp.

Here is the caller graph for this function:

search_branch_and_bound()

void JCBB::search_branch_and_bound ( int  current_obs_idx, int  num_observations, int  num_landmarks, const Eigen::MatrixXd &  distances, Eigen::VectorXi &  current_hypothesis, Eigen::VectorXi &  best_hypothesis, double &  best_score, double  current_score  ) const

private

Recursive branch and bound search to find the best hypothesis.

Parameters

current_obs_idx	Index of the observation being processed.
num_observations	Total number of observations.
num_landmarks	Total number of landmarks.
distances	Precomputed Euclidean distances between observations and landmarks.
current_hypothesis	Current hypothesis (partial).
best_hypothesis	Best hypothesis found so far (output).
best_score	Best score found so far (output).
current_score	Score of the current partial hypothesis.

Definition at line 33 of file jcbb.cpp.

Here is the call graph for this function:

Here is the caller graph for this function:

---

The documentation for this class was generated from the following files:

• src/perception_sensor_lib/include/perception_sensor_lib/data_association/jcbb.hpp
• src/perception_sensor_lib/src/data_association/jcbb.cpp