A controller that implements a simple PID control loop. More...

Public Member Functions
Construction
	PIDFeedback (Robot *const _r, const double _p, const double _i, const double _d)

	PIDFeedback (Robot *const _r, XMLNode &_node)

	PIDFeedback (Robot *const _r, const PIDFeedback &_c)

virtual std::unique_ptr< ControllerMethod >	Clone (Robot *const _r) const override

virtual	~PIDFeedback ()

Interface
virtual Control	operator() (const Cfg &_current, const Cfg &_target, const double _dt) override

Interface
virtual Control	GetRandomControl (const Cfg &_current, const double _dt) const noexcept

ControlSet *	GetControlSet () noexcept

void	SetControlSet (ControlSet *const _c) noexcept

Static Public Member Functions
Construction
static std::unique_ptr< ControllerMethod >	Factory (Robot *const _r, XMLNode &_node)

Protected Member Functions
Control Selection Overrides
virtual std::vector< double >	ComputeDesiredForce (const Cfg &_current, const Cfg &_target, const double _dt) override

Helpers
void	Initialize (const Cfg &_target)

Control Selection
virtual Control	ComputeNearestControl (const Cfg &_current, std::vector< double > &&_force)

virtual Control	ComputeNearestContinuousControl (const Cfg &_current, std::vector< double > &&_force)

virtual Control	ComputeNearestDiscreteControl (const Cfg &_current, std::vector< double > &&_force)

Protected Attributes
Internal State
double	m_p
	The gain for proportional error.

double	m_i
	The gain for integral error.

double	m_d
	The gain for derivative error.

Cfg	m_previousError
	The last step's error value.

Cfg	m_integral
	The integrated error history.

Cfg	m_target
	The previous target.

Internal State
Robot *const	m_robot
	The owning robot object.

ControlSet *	m_controls {nullptr}
	The discrete controls, if any.

bool	m_debug {false}
	Show debug messages?

Detailed Description

A controller that implements a simple PID control loop.

Constructor & Destructor Documentation

◆ PIDFeedback() [1/3]

PIDFeedback::PIDFeedback	(	Robot *const	_r,
		const double	_p,
		const double	_i,
		const double	_d
	)

Construct a PID feedback controller.

Parameters

[in]	_r	The robot to control.
[in]	_p	The proportional gain.
[in]	_i	The integral gain.
[in]	_d	The derivative gain.

◆ PIDFeedback() [2/3]

PIDFeedback::PIDFeedback	(	Robot *const	_r,
		XMLNode &	_node
	)

Construct a PID feedback controller from an XML node.

Parameters

[in]	_r	The robot to control.
[in]	_node	The XML node to parse.

◆ PIDFeedback() [3/3]

PIDFeedback::PIDFeedback	(	Robot *const	_r,
		const PIDFeedback &	_c
	)

Copy a controller for another robot.

Parameters

_r	The destination robot.
_c	The controller to copy.

Member Function Documentation

◆ Clone()

std::unique_ptr< ControllerMethod > PIDFeedback::Clone ( Robot *const _r ) const

overridevirtual

Create a copy of this controller for another robot. This is provided so that we can copy controllers without knowing their derived type.

Parameters

_r	The robot receiving the cloned controller.

Returns: The cloned controller.

Implements ControllerMethod.

◆ operator()()

Control PIDFeedback::operator()	(	const Cfg &	_current,
		const Cfg &	_target,
		const double	_dt
	)

overridevirtual

Find the best available control to steer a robot from a starting configuration to a target configuration.

Parameters

_current	The current configuration.
_target	The target configuration.
_dt	The timestep length.

Returns: The best available control for steering from _current to _target.

Reimplemented from ControllerMethod.

◆ ComputeDesiredForce()

std::vector< double > PIDFeedback::ComputeDesiredForce	(	const Cfg &	_current,
		const Cfg &	_target,
		const double	_dt
	)

overrideprotectedvirtual

Compute the desired generalized force (or velocity) in the robot's local frame to move from the current position to the target.

Parameters

_current	The current configuration.
_target	The target configuration.
_dt	The timestep length.

Returns: The ideal generalized force.

Implements ControllerMethod.

◆ Initialize()

void PIDFeedback::Initialize ( const Cfg & _target )

protected

PID controllers re-initialize every time they change targets.

Parameters

[in] _target The target configuration.

◆ Factory()

std::unique_ptr< ControllerMethod > ControllerMethod::Factory	(	Robot *const	_r,
		XMLNode &	_node
	)

staticinherited

Create a dynamically-allocated controller from an XML node.

Parameters

_r	The robot to control.
_node	The XML node to read parameters from.

Returns: A controller of the type described by _node.

◆ GetRandomControl()

Control ControllerMethod::GetRandomControl	(	const Cfg &	_current,
		const double	_dt
	)		const

virtualnoexceptinherited

Find a random control.

Parameters

_current	The current configuration.
_dt	The timestep length.

Returns: A random control that can be executed from _current for _dt steps.

Reimplemented in CarlikeNeedleController, and MatlabNeedleController.

◆ GetControlSet()

ControlSet * ControllerMethod::GetControlSet ( )

noexceptinherited

Get the discrete set of controls that this controller can use, if any.

Returns: The discrete set of controls for this controller, or null if it uses a continuous space(s) of controls.

◆ SetControlSet()

void ControllerMethod::SetControlSet ( ControlSet *const _c )

noexceptinherited

Set a discrete set of controls for this controller to use. This limits the controller to a subset of all possible controls accepted by a robot's actuators.

Parameters

_c	The control set to use.

◆ ComputeNearestControl()

Control ControllerMethod::ComputeNearestControl	(	const Cfg &	_current,
		std::vector< double > &&	_force
	)

protectedvirtualinherited

Compute the control that produces the closest force to the ideal.

Parameters

_current	The current configuration.
_force	The desired force in the robot's local frame.

Returns: The control whose result is nearest to _force.

◆ ComputeNearestContinuousControl()

Control ControllerMethod::ComputeNearestContinuousControl	(	const Cfg &	_current,
		std::vector< double > &&	_force
	)

protectedvirtualinherited

Compute the continuous control that produces the closest force to the ideal.

Parameters

_current	The current configuration.
_force	The desired force in the robot's local frame.

Returns: The control whose result is nearest to _force.

Reimplemented in CarlikeNeedleController.

◆ ComputeNearestDiscreteControl()

Control ControllerMethod::ComputeNearestDiscreteControl	(	const Cfg &	_current,
		std::vector< double > &&	_force
	)

protectedvirtualinherited

Compute the discrete control that produces the closest force to the ideal.

Parameters

_current	The current configuration.
_force	The desired force in the robot's local frame.

Returns: The control whose result is nearest to _force.

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

PIDFeedback.h
PIDFeedback.cpp

Public Member Functions

Static Public Member Functions

Protected Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

◆ PIDFeedback() [1/3]

◆ PIDFeedback() [2/3]

◆ PIDFeedback() [3/3]

Member Function Documentation

◆ Clone()

◆ operator()()

◆ ComputeDesiredForce()

◆ Initialize()

◆ Factory()

◆ GetRandomControl()

◆ GetControlSet()

◆ SetControlSet()

◆ ComputeNearestControl()

◆ ComputeNearestContinuousControl()

◆ ComputeNearestDiscreteControl()