GraspTaskSimulator Class Reference

A class for simulating multiple grasping tasks. More...

#include <GraspTaskSimulator.hpp>

Classes
struct	GraspedObject
struct	SimState

Public Types
enum	StepState { GRASPING , LIFTING , NEW_GRASP , APPROACH }
typedef rw::core::Ptr< GraspTaskSimulator >	Ptr
	smart pointer type

Public Member Functions
	GraspTaskSimulator (rwsim::dynamics::DynamicWorkCell::Ptr dwc, int nrThreads=1)
	constructor More...
virtual	~GraspTaskSimulator ()
	destructor
void	load (const std::string &filename)
	load tasks
void	load (rwlibs::task::GraspTask::Ptr graspTasks)
rwlibs::task::GraspTask::Ptr	getTasks ()
rwlibs::task::GraspTask::Ptr	getResult ()
size_t	getNrTargets ()
	get the number of targets
ThreadSimulator::Ptr	getSimulator ()
std::vector< ThreadSimulator::Ptr >	getSimulators ()
void	init (rwsim::dynamics::DynamicWorkCell::Ptr dwc, const rw::kinematics::State &initState, std::string engineID="ODE")
void	startSimulation (const rw::kinematics::State &initState)
void	pauseSimulation ()
void	resumeSimulation ()
bool	isRunning ()
bool	isFinished ()
int	getStat (rwlibs::task::GraspResult::TestStatus status)
	get the statistics of the current simulations More...
std::vector< int >	getStat ()
	get the current statistics. The array is based on the enumeration rwlibs::task::GraspTask::TestStatus
std::string	getStatDescription ()
int	getNrTargetsDone ()
	get the number of targets that have been simulated until now
void	setAlwaysResting (bool alwaysResting)
void	setStepDelay (int delay)
	add a delay each time the simulation has stepped. usefull for debuging or visualization. More...
void	setWallTimeLimit (double limit)
	Sets the limit on wall time required for single grasp simulation. More...
void	setSimTimeLimit (double limit)
	Sets the limit on internal simulation time for single grasp simulation. More...
void	setStoreTimedStatePaths (bool enabled)
	Enable storing the simulations as TimedStatePaths. More...
void	forceSimulateAll (bool enabled)
	force the simulation of all tasks, even those that already have results in specified in the task More...
std::map< rwlibs::task::GraspSubTask *, std::map< rwlibs::task::GraspTarget *, rw::trajectory::TimedStatePath > >	getTimedStatePaths () const
	Get the current map of stored TimedStatePaths. More...

Protected Member Functions
virtual void	graspFinished (SimState &sstate)
	Callback called after each finished grasp.
virtual void	printGraspResult (SimState &sstate)
	Prints result after the grasp is finished.
virtual void	simulationFinished (SimState &sstate)
	Callback called after simulation is finished.

Protected Attributes
rwsim::dynamics::DynamicWorkCell::Ptr	_dwc
bool	_requestSimulationStop
int	_stepDelayMs
int	_autoSaveInterval
double	_maxObjectGripperDistanceThreshold
std::vector< int >	_stat
bool	_initialized
int	_nrOfThreads
int	_currentTargetIndex
bool	_alwaysResting
int	_gripperDim
int	_failed
int	_success
int	_slipped
int	_collision
int	_timeout
int	_simfailed
int	_skipped
int	_nrOfExperiments
int	_lastSaveTaskIndex
int	_totalNrOfExperiments
std::vector< rw::core::Ptr< rwsim::dynamics::RigidBody > >	_objects
rwsim::dynamics::DynamicDevice::Ptr	_dhand
rw::core::Ptr< rwsim::dynamics::RigidDevice >	_rhand
rw::models::Device::Ptr	_hand
rw::core::Ptr< rwsim::dynamics::KinematicBody >	_hbase
rw::kinematics::MovableFrame *	_mbase
rwlibs::simulation::SimulatedController::Ptr	_simGraspController
rw::kinematics::Frame *	_tcp
rw::kinematics::State	_homeState
std::map< rwsim::simulator::ThreadSimulator::Ptr, SimState >	_simStates
std::vector< rwsim::simulator::ThreadSimulator::Ptr >	_simulators
rwlibs::task::GraspSubTask *	_currentTask
rwlibs::task::GraspTarget *	_currentTarget
rwlibs::task::GraspTask::Ptr	_gtask
std::stack< std::pair< rwlibs::task::GraspSubTask *, rwlibs::task::GraspTarget * > >	_taskQueue
rw::core::Ptr< rw::proximity::CollisionDetector >	_collisionDetector
boost::mutex	_nextTargetLock
double	_wallTimeLimit
double	_simTimeLimit
bool	_storeTimedStatePaths
bool	_forceSimulateAll
std::map< rwlibs::task::GraspSubTask *, std::map< rwlibs::task::GraspTarget *, rw::trajectory::TimedStatePath > >	_timedStatePaths

Detailed Description

A class for simulating multiple grasping tasks.

A grasping task can be considered as a sequence of the following states

• initialize gripper joint values and pose
• approach the target "object"
• grasp target by closing fingers
• stop grasp when either fingers are still or there is a timeout
• calculate grasp success
• retract gripper
• calculate grasp success

The home and target poses are controlled through a task description file. Which is also used to write back all the results of the simulation.

The configuration of the simulation is setup through properties. These can be set from the command prompt, loaded by file, or edited in the gui. These properties include:

• Simulator
• TimeStepSize
• HandOpenConfig
• HandCloseConfig
• MinRestingTime

Constructor & Destructor Documentation

GraspTaskSimulator()

GraspTaskSimulator	(	rwsim::dynamics::DynamicWorkCell::Ptr	dwc,
		int	nrThreads = 1
	)

constructor

Parameters

dwc	[in] the dynamic workcell
nrThreads	[in] the number of parallel simulations to run

Member Function Documentation

forceSimulateAll()

void forceSimulateAll ( bool  enabled )

inline

force the simulation of all tasks, even those that already have results in specified in the task

Parameters

enabled

getStat()

int getStat ( rwlibs::task::GraspResult::TestStatus  status )

inline

get the statistics of the current simulations

Parameters

status

[in]

Returns

getTimedStatePaths()

std::map<rwlibs::task::GraspSubTask*, std::map<rwlibs::task::GraspTarget*, rw::trajectory::TimedStatePath> > getTimedStatePaths ( ) const

inline

Get the current map of stored TimedStatePaths.

Note that setStoreTimedStatePaths must be set to enabled before running the simulation or the map will be empty. Please call this function only after simulation is finished.

Returns

a map of maps to a TimedStatePath (indexed by GraspSubTask pointer and a GraspTarget pointer).

load()

void load

(

rwlibs::task::GraspTask::Ptr

graspTasks

)

Parameters

graspTasks

setSimTimeLimit()

void setSimTimeLimit ( double  limit )

inline

Sets the limit on internal simulation time for single grasp simulation.

Parameters

limit

[in] simulation time limit in seconds (default 30.0)

setStepDelay()

void setStepDelay ( int  delay )

inline

add a delay each time the simulation has stepped. usefull for debuging or visualization.

Parameters

delay

setStoreTimedStatePaths()

void setStoreTimedStatePaths ( bool  enabled )

inline

Enable storing the simulations as TimedStatePaths.

Parameters

enabled

[in] whether or not a TimeStatePath should be stored for each grasp simulation.

setWallTimeLimit()

void setWallTimeLimit ( double  limit )

inline

Sets the limit on wall time required for single grasp simulation.

Parameters

limit

[in] wall time limit in seconds (default 30.0)

---

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

• GraspTaskSimulator.hpp