template class mrpt::graphslam::deciders::CICPCriteriaERD

ICP-based Edge Registration.

Register new edges in the graph with the last inserted node. Criterion for adding new edges should be the goodness of the candidate ICP edge. The nodes for ICP are picked based on the distance from the last inserted node.

Map type: 2D
MRPT rawlog format: #1, #2
Graph Type: CPosePDFGaussianInf
Observations: CObservation2DRangeScan, CObservation3DRangeScan
Edge Registration Strategy: ICP Goodnesss threshold
class_verbosity
- Section : EdgeRegistrationDeciderParameters
- default value : 1 (mrpt::system::LVL_INFO)
- Required : FALSE
LC_min_nodeid_diff
- Section : GeneralConfiguration
- Default value : 30
- Required : FALSE
- Description : Minimum NodeID difference for an edge to be considered a loop closure.
ICP_max_distance
- Section : EdgeRegistrationDeciderParameters
- Default value : 10 // meters
- Required : FALSE
- Description : Maximum distance for scan-matching. Decider tries to align the laser scans of the current node and each of the previous nodes that are found within the designated ICP_max_distance.
ICP_goodness_thresh
- Section : EdgeRegistrationDeciderParameters
- Default value : 0.75
- Required : FALSE
- Description : Threshold for accepting a scan-matching edge between the current and previous nodes
visualize_laser_scans
- Section : VisualizationParameters
- Default value : TRUE
- Required : FALSE
scans_img_external_dir
- Section : EdgeRegistrationDeciderParameters
- Default value : . (current directory)
- Required : FALSE
- Description : Only applicable in datasets with 3DRangeScans that are externally stored (not stored in the given .rawlog file).

Inherited Members

public:
    // typedefs

    typedef mrpt::graphslam::CRegistrationDeciderOrOptimizer<GRAPH_T> parent;
    typedef typename GRAPH_T::constraint_t constraint_t;
    typedef typename GRAPH_T::constraint_t::type_value pose_t;
    typedef mrpt::graphslam::deciders::CEdgeRegistrationDecider<GRAPH_T> parent_t;
    typedef mrpt::graphslam::deciders::CRangeScanOps<GRAPH_T> range_ops_t;
    typedef std::map<mrpt::graphs::TNodeID, mrpt::obs::CObservation2DRangeScan::Ptr> nodes_to_scans2D_t;

    // structs

    struct TMsg;
    struct TParams;

    //
methods

    virtual void setClassName(const std::string& name);
    bool isMultiRobotSlamClass();
    std::string getClassName() const;
    virtual void getEdgesStats(std::map<std::string, int>* edge_type_to_num) const;
    virtual bool justInsertedLoopClosure() const;
    virtual void getDescriptiveReport(std::string* report_str) const;
    virtual bool updateState(mrpt::obs::CActionCollection::Ptr action, mrpt::obs::CSensoryFrame::Ptr observations, mrpt::obs::CObservation::Ptr observation) = 0;
    virtual void setWindowManagerPtr(mrpt::graphslam::CWindowManager* win_manager);
    virtual void setCriticalSectionPtr(std::mutex* graph_section);
    virtual void initializeVisuals();
    virtual void updateVisuals();
    virtual void notifyOfWindowEvents(const std::map<std::string, bool>& events_occurred);
    virtual void loadParams(const std::string& source_fname);
    virtual void printParams() const;
    virtual void setGraphPtr(GRAPH_T* graph);
    virtual void initializeLoggers(const std::string& name);

Typedefs

typedef typename GRAPH_T::constraint_t constraint_t

Handy typedefs.

type of graph constraints

typedef typename GRAPH_T::constraint_t::type_value pose_t

type of underlying poses (2D/3D).

typedef CICPCriteriaERD<GRAPH_T> decider_t

self type - Handy type

Methods

virtual bool updateState(mrpt::obs::CActionCollection::Ptr action, mrpt::obs::CSensoryFrame::Ptr observations, mrpt::obs::CObservation::Ptr observation)

Generic method for fetching the incremental action-observations (or observation-only) measurements.

Returns:

True if operation was successful. Criteria for Success depend on the decider/optimizer implementing this method

virtual void notifyOfWindowEvents(const std::map<std::string, bool>& events_occurred)

Get a list of the window events that happened since the last call.

Method in derived classes is automatically called from the CGraphSlamEngine_t instance. After that, decider/optimizer should just fetch the parameters that it is interested in.

virtual void getEdgesStats(std::map<std::string, int>* edge_type_to_num) const

Fill the given map with the type of registered edges as well as the corresponding number of registration of each edge.

virtual void initializeVisuals()

Initialize visual objects in CDisplayWindow (e.g.

add an object to scene).

Parameters:

std::exception

If the method is called without having first provided a CDisplayWindow3D* to the class instance