WorldBuilder Class Reference

Inheritance diagram for WorldBuilder:

Classes
struct	CompareCluster

Public Types
enum	CLUSTERING { BOX =0 , KMEANS =1 , PARTITION =2 }

Public Member Functions
b2Body *	makeBody (b2World &w, const BodyFeatures &features)
	Creates a body in the box2d world, and if the features represent a disturbance to which the attention window needs to be assigned, a flag is assigned to the body user data. More...
std::vector< BodyFeatures >	processData (const CoordinateContainer &, const b2Transform &)
	returns a bounding box encompassing all points provided More...
std::vector< BodyFeatures >	cluster_data (const CoordinateContainer &pts, const b2Transform &start, CLUSTERING clustering=PARTITION)
	Cluster point cloud data using a custom algorithm. More...
bool	checkDisturbance (Pointf, bool &, Task *curr=NULL, float range=0.025)
virtual std::vector< BodyFeatures >	getFeatures (const CoordinateContainer &current, b2Transform start, CLUSTERING clustering=PARTITION)
	clusters point clouds into objects and returns a vector of body features More...
virtual void	buildWorld (b2World &, b2Transform, Direction, Disturbance disturbance=Disturbance(), float halfWindowWidth=0.15, CLUSTERING clustering=CLUSTERING::PARTITION, Task *task=NULL)
	Creates bodies (objects) in the box2d world. More...
b2PolygonShape	object_filtering_box (float halfWindowWidth, float boxLength, b2Transform start, Direction d)
	Makes a box. More...
std::pair< bool, BodyFeatures >	bounding_approx_poly (std::vector< cv::Point2f >nb)
std::vector< std::vector< cv::Point2f > >	kmeans_clusters (std::vector< cv::Point2f >, std::vector< cv::Point2f > &)
	Clusters points using k-means algorithm. More...
std::vector< std::vector< cv::Point2f > >	partition_clusters (std::vector< cv::Point2f >)
	Clusters points using the partition algorithm. More...
b2Vec2	averagePoint (const CoordinateContainer &, Disturbance &, float rad=0.025)
int	getBodies ()
void	add_body_count ()
void	resetBodies ()
void	add_iteration (int i=1)
b2Body *	get_robot (b2World *)
b2Fixture *	get_chassis (b2Body *)
b2AABB	makeRobotSensor (b2Body *const robotBody, const Disturbance &focus) const
	Makes the robot attention window, i.e. a distal sensor which is comprised between the extremes of the robot body and the disturbance. More...
std::vector< BodyFeatures > &	get_world_objects ()
void	set_world_objects (const std::vector< BodyFeatures > &wo)
void	setSimulationStep (float f)

Static Public Member Functions
static std::pair< CoordinateContainer, bool >	salientPoints (b2Transform, const CoordinateContainer &, std::pair< Pointf, Pointf >)
static std::pair< Pointf, Pointf >	bounds (Direction, b2Transform t, float boxLength, float halfWindowWidth, std::vector< Pointf > *_bounds=NULL)

Protected Attributes
int	iteration =0
char	bodyFile [100]
float	simulationStep =BOX2DRANGE
int	bodies =0
std::vector< BodyFeatures >	world_objects

Friends
class	Configurator

Member Function Documentation

buildWorld()

void WorldBuilder::buildWorld ( b2World &  world, b2Transform  start, Direction  d, Disturbance  disturbance = Disturbance(), float  halfWindowWidth = 0.15, CLUSTERING  clustering = CLUSTERING::PARTITION, Task *  task = NULL  )

virtual

Creates bodies (objects) in the box2d world.

Parameters

disturbance
halfWindowWidth
clustering
task

Reimplemented in LaserFocus, EveryOtherPointBuilder, and EverythingBuilder.

cluster_data()

std::vector< BodyFeatures > WorldBuilder::cluster_data	(	const CoordinateContainer &	pts,
		const b2Transform &	start,
		CLUSTERING	clustering = PARTITION
	)

Cluster point cloud data using a custom algorithm.

Parameters

pts	point cloud
start	start robot transform
clustering	the clustering algorithm return a vector of bodyfeatures

getFeatures()

std::vector< BodyFeatures > WorldBuilder::getFeatures ( const CoordinateContainer &  current, b2Transform  start, CLUSTERING  clustering = PARTITION  )

virtual

clusters point clouds into objects and returns a vector of body features

Parameters

current	point cloud
start	robot position
partition	algorithm used for partition

Reimplemented in LaserFocus, EveryOtherPointBuilder, EveryOtherFeatureBuilder, and WorldPointBuilder.

kmeans_clusters()

std::vector< std::vector< cv::Point2f > > WorldBuilder::kmeans_clusters	(	std::vector< cv::Point2f >	points,
		std::vector< cv::Point2f > &	centers
	)

Clusters points using k-means algorithm.

Returns

std::vector <std::vector<cv::Point2f>>

makeBody()

b2Body * WorldBuilder::makeBody	(	b2World &	w,
		const BodyFeatures &	features
	)

Creates a body in the box2d world, and if the features represent a disturbance to which the attention window needs to be assigned, a flag is assigned to the body user data.

Parameters

w	the box2d world
features	features of the body to be created

Returns

* b2Body*

makeRobotSensor()

b2AABB WorldBuilder::makeRobotSensor	(	b2Body *const	robotBody,
		const Disturbance &	focus
	)		const

Makes the robot attention window, i.e. a distal sensor which is comprised between the extremes of the robot body and the disturbance.

Parameters

robotBody
focus	a disturbance representing the focus of the attention window

Returns

b2AABB

object_filtering_box()

b2PolygonShape WorldBuilder::object_filtering_box	(	float	halfWindowWidth,
		float	boxLength,
		b2Transform	start,
		Direction	d
	)

Makes a box.

Parameters

halfWindowWidth
boxLength
start
d

Returns

b2PolygonShape

partition_clusters()

std::vector< std::vector< cv::Point2f > > WorldBuilder::partition_clusters

(

std::vector< cv::Point2f >

points

)

Clusters points using the partition algorithm.

Returns

std::vector <std::vector<cv::Point2f>>

processData()

std::vector< BodyFeatures > WorldBuilder::processData	(	const CoordinateContainer &	points,
		const b2Transform &	start
	)

returns a bounding box encompassing all points provided

Returns

std::vector <BodyFeatures>

---

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

• src/worldbuilder.h
• src/worldbuilder.cpp