User-interface to formulate and solve model predictive control problems. More...

#include <real_time_algorithm.hpp>

Inheritance diagram for RealTimeAlgorithm:

Public Member Functions
virtual ControlLaw *	clone () const
virtual returnValue	feedbackStep (double currentTime, const DVector &_x, const DVector &_p=emptyConstVector, const VariablesGrid &_yRef=emptyConstVariablesGrid)
virtual double	getLengthControlHorizon () const
virtual double	getLengthPredictionHorizon () const
virtual uint	getNP () const
virtual uint	getNU () const
virtual uint	getNW () const
virtual uint	getNX () const
virtual uint	getNXA () const
virtual uint	getNY () const
virtual returnValue	init ()
virtual returnValue	init (double startTime, const DVector &_x=emptyConstVector, const DVector &_p=emptyConstVector, const VariablesGrid &_yRef=emptyConstVariablesGrid)
virtual returnValue	initializeAlgebraicStates (const VariablesGrid &_xa_init)
virtual returnValue	initializeAlgebraicStates (const char *fileName)
virtual returnValue	initializeControls (const VariablesGrid &_u_init)
virtual returnValue	initializeControls (const char *fileName)
virtual BooleanType	isDynamic () const
virtual BooleanType	isInRealTimeMode () const
virtual BooleanType	isStatic () const
RealTimeAlgorithm &	operator= (const RealTimeAlgorithm &rhs)
virtual returnValue	preparationStep (double nextTime=0.0, const VariablesGrid &_yRef=emptyConstVariablesGrid)
	RealTimeAlgorithm ()
	RealTimeAlgorithm (const OCP &ocp_, double _samplingTime=DEFAULT_SAMPLING_TIME)
	RealTimeAlgorithm (const RealTimeAlgorithm &rhs)
virtual returnValue	setReference (const VariablesGrid &ref)
virtual returnValue	shift (double timeShift=-1.0)
virtual returnValue	solve (double startTime, const DVector &_x, const DVector &_p=emptyConstVector, const VariablesGrid &_yRef=emptyConstVariablesGrid)
virtual returnValue	step (double currentTime, const DVector &_x, const DVector &_p=emptyConstVector, const VariablesGrid &_yRef=emptyConstVariablesGrid)
virtual	~RealTimeAlgorithm ()
Protected Member Functions
virtual returnValue	allocateNlpSolver (Objective F, DynamicDiscretization G, Constraint *H)
returnValue	clear ()
virtual returnValue	initializeNlpSolver (const OCPiterate &userInit)
virtual returnValue	initializeObjective (Objective *F)
returnValue	performFeedbackStep (double currentTime, const DVector &_x, const DVector &_p=emptyConstVector)
returnValue	performPreparationStep (const VariablesGrid &_yRef=emptyConstVariablesGrid, BooleanType isLastIteration=BT_TRUE)
virtual returnValue	setupLogging ()
virtual returnValue	setupOptions ()
Protected Attributes
DVector *	p0
VariablesGrid *	reference
DVector *	x0

Detailed Description

User-interface to formulate and solve model predictive control problems.

The class RealTimeAlgorithm serves as a user-interface to formulate and solve model predictive control problems.

Author:: Hans Joachim Ferreau, Boris Houska

Definition at line 56 of file real_time_algorithm.hpp.

Constructor & Destructor Documentation

BEGIN_NAMESPACE_ACADO RealTimeAlgorithm::RealTimeAlgorithm ( )

Default constructor.

Definition at line 50 of file real_time_algorithm.cpp.

RealTimeAlgorithm::RealTimeAlgorithm	(	const OCP &	ocp_,
		double	_samplingTime = `DEFAULT_SAMPLING_TIME`
	)

Constructor which takes the optimal control problem to be solved online together with the sampling time.

Parameters:

[in]	ocp_	Optimal control problem to be solved online.
[in]	_samplingTime	Sampling time.

Definition at line 69 of file real_time_algorithm.cpp.

RealTimeAlgorithm::RealTimeAlgorithm ( const RealTimeAlgorithm & rhs )

Copy constructor (deep copy).

Parameters:

[in] rhs Right-hand side object.

Definition at line 91 of file real_time_algorithm.cpp.

RealTimeAlgorithm::~RealTimeAlgorithm ( ) [virtual]

Destructor.

Definition at line 104 of file real_time_algorithm.cpp.

Member Function Documentation

returnValue RealTimeAlgorithm::allocateNlpSolver	(	Objective *	F,
		DynamicDiscretization *	G,
		Constraint *	H
	)		`[protected, virtual]`

(not yet documented).

Parameters:

[in] . .

Returns:: SUCCESSFUL_RETURN

Implements OptimizationAlgorithmBase.

Definition at line 576 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::clear ( ) [protected]

Frees memory of all members of the real time parameters.

Returns:: SUCCESSFUL_RETURN

Reimplemented from OptimizationAlgorithmBase.

Definition at line 551 of file real_time_algorithm.cpp.

ControlLaw * RealTimeAlgorithm::clone ( ) const [virtual]

Clone constructor (deep copy).

Returns:: Pointer to deep copy of base class type

Implements ControlLaw.

Definition at line 133 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::feedbackStep	(	double	currentTime,
		const DVector &	_x,
		const DVector &	_p = `emptyConstVector`,
		const VariablesGrid &	_yRef = `emptyConstVariablesGrid`
	)		`[virtual]`

Performs next feedback step of the control law based on given inputs.

Parameters:

[in]	currentTime	Current time.
[in]	_x	Most recent value for differential states.
[in]	_p	Most recent value for parameters.
[in]	_yRef	Current piece of reference trajectory (if not specified during previous preparationStep).

Note:: If a non-empty reference trajectory is provided, this one is used instead of the possibly set-up build-in one.

Returns:: SUCCESSFUL_RETURN

Definition at line 251 of file real_time_algorithm.cpp.

double RealTimeAlgorithm::getLengthControlHorizon ( ) const [virtual]

Returns length of the control horizon (for the case a predictive control law is used).

Returns:: Length of the control horizon

Reimplemented from ControlLaw.

Definition at line 442 of file real_time_algorithm.cpp.

double RealTimeAlgorithm::getLengthPredictionHorizon ( ) const [virtual]

Returns length of the prediction horizon (for the case a predictive control law is used).

Returns:: Length of the prediction horizon

Reimplemented from ControlLaw.

Definition at line 436 of file real_time_algorithm.cpp.

uint RealTimeAlgorithm::getNP ( ) const [virtual]

Returns number of parameters.

Returns:: Number of parameters

Reimplemented from ControlLaw.

Definition at line 418 of file real_time_algorithm.cpp.

uint RealTimeAlgorithm::getNU ( ) const [virtual]

Returns number of controls.

Returns:: Number of controls

Reimplemented from ControlLaw.

Definition at line 413 of file real_time_algorithm.cpp.

uint RealTimeAlgorithm::getNW ( ) const [virtual]

Returns number of (estimated) disturbances.

Returns:: Number of (estimated) disturbances

Reimplemented from ControlLaw.

Definition at line 423 of file real_time_algorithm.cpp.

uint RealTimeAlgorithm::getNX ( ) const [virtual]

Returns number of (estimated) differential states.

Returns:: Number of (estimated) differential states

Reimplemented from ControlLaw.

Definition at line 403 of file real_time_algorithm.cpp.

uint RealTimeAlgorithm::getNXA ( ) const [virtual]

Returns number of (estimated) algebraic states.

Returns:: Number of (estimated) algebraic states

Reimplemented from ControlLaw.

Definition at line 408 of file real_time_algorithm.cpp.

uint RealTimeAlgorithm::getNY ( ) const [virtual]

Returns number of process outputs.

Returns:: Number of process outputs

Reimplemented from ControlLaw.

Definition at line 429 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::init ( ) [virtual]

Initializes the (internal) optimization algorithm part of the RealTimeAlgorithm.

Returns:: SUCCESSFUL_RETURN,
RET_OPTALG_INIT_FAILED

Definition at line 169 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::init	(	double	startTime,
		const DVector &	_x = `emptyConstVector`,
		const DVector &	_p = `emptyConstVector`,
		const VariablesGrid &	_yRef = `emptyConstVariablesGrid`
	)		`[virtual]`

Initializes the control law with given start values and performs a number of consistency checks.

Parameters:

[in]	_startTime	Start time.
[in]	_x	Initial value for differential states.
[in]	_p	Initial value for parameters.
[in]	_yRef	Initial value for reference trajectory.

Returns:: SUCCESSFUL_RETURN

Definition at line 183 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::initializeAlgebraicStates ( const VariablesGrid & _xa_init ) [virtual]

Initializes algebraic states of the control law.

Parameters:

[in] _xa_init Initial value for algebraic states.

Returns:: SUCCESSFUL_RETURN

Reimplemented from ControlLaw.

Definition at line 140 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::initializeAlgebraicStates ( const char * fileName ) [virtual]

Initializes algebraic states of the control law from data file.

Parameters:

[in] fileName Name of file containing initial value for algebraic states.

Returns:: SUCCESSFUL_RETURN,
RET_FILE_CAN_NOT_BE_OPENED

Reimplemented from ControlLaw.

Definition at line 147 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::initializeControls ( const VariablesGrid & _u_init ) [virtual]

Initializes controls of the control law.

Parameters:

[in] _u_init Initial value for controls.

Returns:: SUCCESSFUL_RETURN

Reimplemented from ControlLaw.

Definition at line 154 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::initializeControls ( const char * fileName ) [virtual]

Initializes controls of the control law from data file.

Parameters:

[in] fileName Name of file containing initial value for controls.

Returns:: SUCCESSFUL_RETURN,
RET_FILE_CAN_NOT_BE_OPENED

Reimplemented from ControlLaw.

Definition at line 161 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::initializeNlpSolver ( const OCPiterate & userInit ) [protected, virtual]

(not yet documented).

Parameters:

[in] . .

Returns:: SUCCESSFUL_RETURN

Implements OptimizationAlgorithmBase.

Definition at line 587 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::initializeObjective ( Objective * F ) [protected, virtual]

(not yet documented).

Parameters:

[in] . .

Returns:: SUCCESSFUL_RETURN

Implements OptimizationAlgorithmBase.

Definition at line 603 of file real_time_algorithm.cpp.

BooleanType RealTimeAlgorithm::isDynamic ( ) const [virtual]

Returns whether the control law is based on dynamic optimization or a static one.

Returns:: BT_TRUE iff control law is based on dynamic optimization,
BT_FALSE otherwise

Implements ControlLaw.

Definition at line 449 of file real_time_algorithm.cpp.

BooleanType RealTimeAlgorithm::isInRealTimeMode ( ) const [virtual]

Returns whether the control law is working in real-time mode.

Returns:: BT_TRUE iff control law is working in real-time mode,
BT_FALSE otherwise

Reimplemented from ControlLaw.

Definition at line 464 of file real_time_algorithm.cpp.

BooleanType RealTimeAlgorithm::isStatic ( ) const [virtual]

Returns whether the control law is a static one or based on dynamic optimization.

Returns:: BT_TRUE iff control law is a static one,
BT_FALSE otherwise

Implements ControlLaw.

Definition at line 455 of file real_time_algorithm.cpp.

RealTimeAlgorithm & RealTimeAlgorithm::operator= ( const RealTimeAlgorithm & rhs )

Assignment operator (deep copy).

Parameters:

[in] rhs Right-hand side object.

Definition at line 111 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::performFeedbackStep	(	double	currentTime,
		const DVector &	_x,
		const DVector &	_p = `emptyConstVector`
	)		`[protected]`

(not yet documented).

Parameters:

[in] . .

Returns:: SUCCESSFUL_RETURN

Definition at line 618 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::performPreparationStep	(	const VariablesGrid &	_yRef = `emptyConstVariablesGrid`,
		BooleanType	isLastIteration = `BT_TRUE`
	)		`[protected]`

(not yet documented).

Parameters:

[in] . .

Returns:: SUCCESSFUL_RETURN

Definition at line 648 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::preparationStep	(	double	nextTime = `0.0`,
		const VariablesGrid &	_yRef = `emptyConstVariablesGrid`
	)		`[virtual]`

Performs next preparation step of the control law based on given inputs.

Parameters:

[in]	nextTime	Time at next step.
[in]	_yRef	Piece of reference trajectory for next step (required for hotstarting).

Returns:: SUCCESSFUL_RETURN

Reimplemented from ControlLaw.

Definition at line 318 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::setReference ( const VariablesGrid & ref ) [virtual]

Assigns new reference trajectory for the next real-time step.

Parameters:

[in] ref Current piece of new reference trajectory.

Returns:: SUCCESSFUL_RETURN,
RET_REFERENCE_SHIFTING_WORKS_FOR_LSQ_TERMS_ONLY,
RET_MEMBER_NOT_INITIALISED

Definition at line 391 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::setupLogging ( ) [protected, virtual]

Sets-up default logging information.

Returns:: SUCCESSFUL_RETURN

Reimplemented from Logging.

Definition at line 539 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::setupOptions ( ) [protected, virtual]

Sets-up default options.

Returns:: SUCCESSFUL_RETURN

Reimplemented from Options.

Definition at line 484 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::shift ( double timeShift = -1.0 ) [virtual]

Shifts the data for preparating the next real-time step.

Returns:: RET_NOT_YET_IMPLEMENTED

Reimplemented from ControlLaw.

Definition at line 379 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::solve	(	double	startTime,
		const DVector &	_x,
		const DVector &	_p = `emptyConstVector`,
		const VariablesGrid &	_yRef = `emptyConstVariablesGrid`
	)		`[virtual]`

(not yet documented).

Parameters:

[in] . .

Returns:: SUCCESSFUL_RETURN

Definition at line 331 of file real_time_algorithm.cpp.

returnValue RealTimeAlgorithm::step	(	double	currentTime,
		const DVector &	_x,
		const DVector &	_p = `emptyConstVector`,
		const VariablesGrid &	_yRef = `emptyConstVariablesGrid`
	)		`[virtual]`

Performs next step of the control law based on given inputs.

Parameters:

[in]	currentTime	Current time.
[in]	_x	Most recent value for differential states.
[in]	_p	Most recent value for parameters.
[in]	_yRef	Current piece of reference trajectory or piece of reference trajectory for next step (required for hotstarting).