Represents a combined configuration of the robot and the articulation objects. More...

#include <RobotSceneConfiguration.h>

Classes
struct	ArticulationCost
	Linear cost function for manipulating an articulated scene object. More...

Public Member Functions
	RobotSceneConfiguration ()
	Constructor.

	RobotSceneConfiguration (const RobotSceneConfiguration &other)
	Copy constructor. More...

virtual	~RobotSceneConfiguration ()
	Destructor.

float	getCost (RobotSceneConfiguration &previousConfig) const
	Computes the cost for transitioning from a previous configuration. More...

float	getGain (RobotSceneConfiguration &previousConfig) const
	Calculates the information gain achieved by moving to the new configuration. More...

float	getEvaluation (RobotSceneConfiguration &previousConfig) const
	Returns the evaluation of the current config. More...

void	set (const RobotSceneConfiguration &other)
	Copy the configuration over from another configuration. More...

void	setRandomArticulation (unsigned int &seed)
	Sets a random articulated object to a random configuration and the other articulated objects to the zero configuration. More...

void	setRandomCameraHeight (unsigned int &seed)
	Sets the camera height to a random value within the feasible range. More...

void	setRandomCameraPosition (unsigned int &seed, const std::vector< glm::vec3 > *const targetPoints)
	Sets the camera pose to a random pose with the camera facing a random point from targetPoints. More...

void	setCameraHeight (const float &value)
	Sets the camera height. More...

void	setCameraLocalTransform (const glm::mat4x4 &cameraTransform)
	Sets the camera to a given pose. More...

const glm::mat4x4 &	getCameraLocalTransform () const
	Returns the camera pose. More...

void	applyToScene (Scene *const scene) const
	Applies the articulation configuration to the scene and re-computes the scene graph transformations. More...

float	getArticulation (const size_t &handle) const
	Returns the current articulation pose for a given articulated object. More...

void	setArticulation (const size_t &handle, const float &value)
	Sets the current articulation pose for a given articulated object. More...

const GLuint &	getCount () const
	Returns the total number of observed pixels so far. More...

void	setCount (const GLuint &count)
	Sets the total number of observed pixels so far. More...

Static Public Member Functions
static void	loadCosts ()
	Load the cost coefficients and feasible ranges from the configuration file. More...

Static Public Attributes
static size_t	numArticulation
	The maximum number of articulated objects, hard-coded to 20.

Protected Attributes
glm::mat4x4	cameraLocalTransform
	Camera pose as homogeneous transformation matrix in world coordinates.

glm::vec3	cameraPosition
	Camera position in world coordinates.

float *	articulation
	Array current of articulation positions.

GLuint	count
	Total number of observed pixels so far.

Static Protected Attributes
static float	costCameraHeightChange
	Cost factor for changing the height of the camera, see loadCosts()

static float	costDistance
	Cost factor for moving the camera (Euclidean distance), see loadCosts()

static float	minCameraHeight
	Minimum feasible camera height above ground, see loadCosts()

static float	maxCameraHeight
	Maximum feasible camera height above ground, see loadCosts()

static float	gainFactor
	Coefficient for weighting the information gain relative to the costs, see loadCosts()

static ArticulationCost *	costArticulation
	Linear cost function for changing articulation, see loadCosts()

Friends
class	RandomSearchTask

class	HillclimbingTask

Detailed Description

Represents a combined configuration of the robot and the articulation objects.

Constructor & Destructor Documentation

gpu_coverage::RobotSceneConfiguration::RobotSceneConfiguration ( const RobotSceneConfiguration & other )

explicit

Copy constructor.

Parameters

[in] other RobotSceneConfiguration to copy data from.

Member Function Documentation

void gpu_coverage::RobotSceneConfiguration::applyToScene ( Scene *const scene ) const

Applies the articulation configuration to the scene and re-computes the scene graph transformations.

Parameters

[in] scene The scene to which to apply the articulation configuration.

float gpu_coverage::RobotSceneConfiguration::getArticulation ( const size_t & handle ) const

inline

Returns the current articulation pose for a given articulated object.

Parameters

[in] handle The index of the articulated object.

Returns: The articulation value between 0.0 and 1.0.

Exceptions

std::out_of_range The index of the articulated object is out of range.

const glm::mat4x4& gpu_coverage::RobotSceneConfiguration::getCameraLocalTransform ( ) const

inline

Returns the camera pose.

Returns: The camera pose as a 4x4 homogeneous transformation matrix in world coordinates.

float gpu_coverage::RobotSceneConfiguration::getCost ( RobotSceneConfiguration & previousConfig ) const

Computes the cost for transitioning from a previous configuration.

Parameters

[in] previousConfig The previous configuration.

Returns: Cost value.

The cost consists of costs for moving the camera from the old position to the new position and the costs for changing the articulated objects' configurations. See loadCosts() for loading the cost coefficients.

const GLuint& gpu_coverage::RobotSceneConfiguration::getCount ( ) const

inline

Returns the total number of observed pixels so far.

Returns: Number of observed pixels.

See also: setCount()

float gpu_coverage::RobotSceneConfiguration::getEvaluation ( RobotSceneConfiguration & previousConfig ) const

inline

Returns the evaluation of the current config.

Parameters

[in] previousConfig The previous config.

Returns: Evaluation value.

The evaluation value is the information gain minus the costs for getting from the previous configuration to the current configuration.

See getGain() and getCosts() for the individual components.

float gpu_coverage::RobotSceneConfiguration::getGain ( RobotSceneConfiguration & previousConfig ) const

Calculates the information gain achieved by moving to the new configuration.

Parameters

[in] previousConfig The previous configuration.

Returns: Information gain value.

Before calling this method, the setCount() method needs to be called with the number of observed pixels for both the previous and current configuration.

static void gpu_coverage::RobotSceneConfiguration::loadCosts ( )

static

Load the cost coefficients and feasible ranges from the configuration file.

The following parameters are used:

Parameter	Description	Comment
costCameraHeightChange	Cost factor for changing the height of the camera
costDistance	Cost factor for moving the camera (Euclidean distance)
costArticulation	Linear cost function for changing articulation (see ArticulationCost)	hard-coded, should be moved to config file
minCameraHeight	Minimum feasible camera height above ground
maxCameraHeight	Maximum feasible camera height above ground
gainFactor	Coefficient for weighting the information gain relative to the costs

void gpu_coverage::RobotSceneConfiguration::set ( const RobotSceneConfiguration & other )

inline

Copy the configuration over from another configuration.

Parameters

[in] other The other configuration to copy from.

void gpu_coverage::RobotSceneConfiguration::setArticulation	(	const size_t &	handle,
		const float &	value
	)

inline

Sets the current articulation pose for a given articulated object.

Parameters

[in]	handle	The index of the articulated object.
[in]	The	articulation value between 0.0 and 1.0.

Exceptions

std::out_of_range The index of the articulated object is out of range.

void gpu_coverage::RobotSceneConfiguration::setCameraHeight ( const float & value )

Sets the camera height.

Parameters

[in] value The new camera height above the ground.

void gpu_coverage::RobotSceneConfiguration::setCameraLocalTransform ( const glm::mat4x4 & cameraTransform )

Sets the camera to a given pose.

Parameters

[in] value The camera pose as a 4x4 homogeneous transformation matrix in world coordinates.

void gpu_coverage::RobotSceneConfiguration::setCount ( const GLuint & count )

inline

Sets the total number of observed pixels so far.

Returns: Number of observed pixels.

See also: getCount()

void gpu_coverage::RobotSceneConfiguration::setRandomArticulation ( unsigned int & seed )

Sets a random articulated object to a random configuration and the other articulated objects to the zero configuration.

Parameters

[in] seed Random seed.

void gpu_coverage::RobotSceneConfiguration::setRandomCameraHeight ( unsigned int & seed )

Sets the camera height to a random value within the feasible range.

Parameters

[in] seed Random seed.

void gpu_coverage::RobotSceneConfiguration::setRandomCameraPosition	(	unsigned int &	seed,
		const std::vector< glm::vec3 > *const	targetPoints
	)

Sets the camera pose to a random pose with the camera facing a random point from targetPoints.

Parameters

[in]	seed	Random seed.
[in]	targetPoints	A list of 3D points in world coordinates

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

include/gpu_coverage/RobotSceneConfiguration.h

Classes

Public Member Functions

Static Public Member Functions

Static Public Attributes

Protected Attributes

Static Protected Attributes

Friends

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation