FGFilter Class Reference

Detailed Description

Encapsulates a filter for the flight control system.

The filter component can simulate any first or second order filter. The Tustin substitution is used to take filter definitions from LaPlace space to the time domain. The general format for a filter specification is:

<typename name="name">
  <input> property </input>
  <c1> value|property </c1>
  [<c2> value|property </c2>]
  [<c3> value|property </c3>]
  [<c4> value|property </c4>]
  [<c5> value|property </c5>]
  [<c6> value|property </c6>]
  [<clipto>
    <min> {[-]property name | value} </min>
    <max> {[-]property name | value} </max>
  </clipto>]
  [<output> property </output>]
</typename>

The numerical integration of filters is made by a Runge-Kutta scheme of order 2 except for the second order filter which uses an RK scheme of order 3.

For a lag filter of the form \(\frac{C_1}{s+C_1}\), the corresponding filter definition is:

<lag_filter name="name">
  <input> property </input>
  <c1> value|property </c1>
  [<clipto>
    <min> {[-]property name | value} </min>
    <max> {[-]property name | value} </max>
  </clipto>]
  [<output> property <output>]
</lag_filter>

As an example, for the specific filter \(\frac{600}{s+600}\) the corresponding filter definition could be:

<lag_filter name="Heading Roll Error Lag">
  <input> fcs/heading-command </input>
  <c1> 600 </c1>
</lag_filter>

For a lead-lag filter of the form \(\frac{C_1s+C_2}{C_3s+C_4}\), the corresponding filter definition is:

<lead_lag_filter name="name">
  <input> property </input>
  <c1> value|property <c/1>
  <c2> value|property <c/2>
  <c3> value|property <c/3>
  <c4> value|property <c/4>
  [<clipto>
    <min> {[-]property name | value} </min>
    <max> {[-]property name | value} </max>
  </clipto>]
  [<output> property </output>]
</lead_lag_filter>

For a washout filter of the form \(\frac{s}{s+C_1}\), the corresponding filter definition is:

<washout_filter name="name">
  <input> property </input>
  <c1> value </c1>
  [<clipto>
    <min> {[-]property name | value} </min>
    <max> {[-]property name | value} </max>
  </clipto>]
  [<output> property </output>]
</washout_filter>

For a second order filter of the form \(\frac{C_1s^2+C_2s+C_3}{C_4s^2+C_5s+C_6}\), the corresponding filter definition is:

<second_order_filter name="name">
  <input> property </input>
  <c1> value|property </c1>
  <c2> value|property </c2>
  <c3> value|property </c3>
  <c4> value|property </c4>
  <c5> value|property </c5>
  <c6> value|property </c6>
  [<clipto>
    <min> {[-]property name | value} </min>
    <max> {[-]property name | value} </max>
  </clipto>]
  [<output> property </output>]
</second_order_filter>

For an integrator of the form \(\frac{C_1}{s}\), the corresponding filter definition is:

<integrator name="{string}">
  <input> {property} </input>
  <c1 type="rect|trap|ab2|ab3"> {[-]property | number} </c1>
  [<trigger> {property} </trigger>]
  [<clipto>
    <min> {[-]property | number} </min>
    <max> {[-]property | number} </max>
  </clipto>]
  [<output> {property} </output>]
</integrator>

For the integrator, the trigger features the following behavior. If the trigger property value is:

• 0: no action is taken - the output is calculated normally
• not 0: (or simply greater than zero), all current and previous inputs will be set to 0.0

By default, the integration scheme is the trapezoidal scheme.

An integrator is equivalent to a PID with the following parameters:

<pid name="{string}">
  <input> {[-]property} </input>
  <kp> 0.0 </kp>
  <ki type="rect|trap|ab2|ab3"> {number|[-]property} </ki>
  <kd> 0.0 </kd>
  <trigger> {property} </trigger>
  [<clipto>
  <min> {[-]property | value} </min>
  <max> {[-]property | value} </max>
  </clipto>]
  [<output> {property} </output>]
</pid>

As a consequence, JSBSim internally uses PID controllers to simulate INTEGRATOR filters.

In all the filter specifications above, an <output> element is also seen. This is so that the last component in a "string" can copy its value to the appropriate output, such as the elevator, or speedbrake, etc.

Author

Jon S. Berndt

Definition at line 215 of file FGFilter.h.

#include <FGFilter.h>

Inheritance diagram for FGFilter:

Collaboration diagram for FGFilter:

Public Member Functions
	FGFilter (FGFCS *fcs, Element *element)
void	ResetPastStates (void) override
bool	Run (void) override
Public Member Functions inherited from FGFCSComponent
	FGFCSComponent (FGFCS *fcs, Element *el)
	Constructor.
virtual	~FGFCSComponent ()
	Destructor.
std::string	GetName (void) const
double	GetOutput (void) const
virtual double	GetOutputPct (void) const
std::string	GetType (void) const
virtual void	SetOutput (void)
Public Member Functions inherited from FGJSBBase
	FGJSBBase ()
	Constructor for FGJSBBase.
virtual	~FGJSBBase ()
	Destructor for FGJSBBase.
void	disableHighLighting (void)
	Disables highlighting in the console output.

Additional Inherited Members
Public Types inherited from FGJSBBase
enum	{ eL = 1 , eM , eN }
	Moments L, M, N. More...
enum	{ eP = 1 , eQ , eR }
	Rates P, Q, R. More...
enum	{ eU = 1 , eV , eW }
	Velocities U, V, W. More...
enum	{ eX = 1 , eY , eZ }
	Positions X, Y, Z. More...
enum	{ ePhi = 1 , eTht , ePsi }
	Euler angles Phi, Theta, Psi. More...
enum	{ eDrag = 1 , eSide , eLift }
	Stability axis forces, Drag, Side force, Lift. More...
enum	{ eRoll = 1 , ePitch , eYaw }
	Local frame orientation Roll, Pitch, Yaw. More...
enum	{ eNorth = 1 , eEast , eDown }
	Local frame position North, East, Down. More...
enum	{ eLat = 1 , eLong , eRad }
	Locations Radius, Latitude, Longitude. More...
enum	{   inNone = 0 , inDegrees , inRadians , inMeters ,   inFeet }
	Conversion specifiers. More...
Static Public Member Functions inherited from FGJSBBase
static const std::string &	GetVersion (void)
	Returns the version number of JSBSim.
static constexpr double	KelvinToFahrenheit (double kelvin)
	Converts from degrees Kelvin to degrees Fahrenheit.
static constexpr double	CelsiusToRankine (double celsius)
	Converts from degrees Celsius to degrees Rankine.
static constexpr double	RankineToCelsius (double rankine)
	Converts from degrees Rankine to degrees Celsius.
static constexpr double	KelvinToRankine (double kelvin)
	Converts from degrees Kelvin to degrees Rankine.
static constexpr double	RankineToKelvin (double rankine)
	Converts from degrees Rankine to degrees Kelvin.
static constexpr double	FahrenheitToCelsius (double fahrenheit)
	Converts from degrees Fahrenheit to degrees Celsius.
static constexpr double	CelsiusToFahrenheit (double celsius)
	Converts from degrees Celsius to degrees Fahrenheit.
static constexpr double	CelsiusToKelvin (double celsius)
	Converts from degrees Celsius to degrees Kelvin.
static constexpr double	KelvinToCelsius (double kelvin)
	Converts from degrees Kelvin to degrees Celsius.
static constexpr double	FeetToMeters (double measure)
	Converts from feet to meters.
static bool	EqualToRoundoff (double a, double b)
	Finite precision comparison.
static bool	EqualToRoundoff (float a, float b)
	Finite precision comparison.
static bool	EqualToRoundoff (float a, double b)
	Finite precision comparison.
static bool	EqualToRoundoff (double a, float b)
	Finite precision comparison.
static constexpr double	Constrain (double min, double value, double max)
	Constrain a value between a minimum and a maximum value.
static constexpr double	sign (double num)
Static Public Attributes inherited from FGJSBBase
static char	highint [5] = {27, '[', '1', 'm', '\0' }
	highlights text
static char	halfint [5] = {27, '[', '2', 'm', '\0' }
	low intensity text
static char	normint [6] = {27, '[', '2', '2', 'm', '\0' }
	normal intensity text
static char	reset [5] = {27, '[', '0', 'm', '\0' }
	resets text properties
static char	underon [5] = {27, '[', '4', 'm', '\0' }
	underlines text
static char	underoff [6] = {27, '[', '2', '4', 'm', '\0' }
	underline off
static char	fgblue [6] = {27, '[', '3', '4', 'm', '\0' }
	blue text
static char	fgcyan [6] = {27, '[', '3', '6', 'm', '\0' }
	cyan text
static char	fgred [6] = {27, '[', '3', '1', 'm', '\0' }
	red text
static char	fggreen [6] = {27, '[', '3', '2', 'm', '\0' }
	green text
static char	fgdef [6] = {27, '[', '3', '9', 'm', '\0' }
	default text
static short	debug_lvl = 1
Protected Member Functions inherited from FGFCSComponent
virtual void	bind (Element *el, FGPropertyManager *pm)
void	CheckInputNodes (size_t MinNodes, size_t MaxNodes, Element *el)
void	Clip (void)
void	Delay (void)
Static Protected Member Functions inherited from FGJSBBase
static std::string	CreateIndexedPropertyName (const std::string &Property, int index)
Protected Attributes inherited from FGFCSComponent
bool	clip
FGParameter_ptr	ClipMax
FGParameter_ptr	ClipMin
bool	cyclic_clip
unsigned int	delay
double	delay_time
double	dt
FGFCS *	fcs
int	index
std::vector< FGPropertyValue_ptr >	InitNodes
double	Input
std::vector< FGPropertyValue_ptr >	InputNodes
std::string	Name
double	Output
std::vector< double >	output_array
std::vector< FGPropertyNode_ptr >	OutputNodes
std::string	Type
Static Protected Attributes inherited from FGJSBBase
static constexpr double	radtodeg = 180. / M_PI
static constexpr double	degtorad = M_PI / 180.
static constexpr double	hptoftlbssec = 550.0
static constexpr double	psftoinhg = 0.014138
static constexpr double	psftopa = 47.88
static constexpr double	fttom = 0.3048
static constexpr double	ktstofps = 1852./(3600*fttom)
static constexpr double	fpstokts = 1.0 / ktstofps
static constexpr double	inchtoft = 1.0/12.0
static constexpr double	m3toft3 = 1.0/(fttom*fttom*fttom)
static constexpr double	in3tom3 = inchtoft*inchtoft*inchtoft/m3toft3
static constexpr double	inhgtopa = 3386.38
static constexpr double	slugtolb = 32.174049
	Note that definition of lbtoslug by the inverse of slugtolb and not to a different constant you can also get from some tables will make lbtoslug*slugtolb == 1 up to the magnitude of roundoff.
static constexpr double	lbtoslug = 1.0/slugtolb
static constexpr double	kgtolb = 2.20462
static constexpr double	kgtoslug = 0.06852168
static const std::string	needed_cfg_version = "2.0"
static const std::string	JSBSim_version = JSBSIM_VERSION " " __DATE__ " " __TIME__

Constructor & Destructor Documentation

FGFilter()

FGFilter	(	FGFCS *	fcs,
		Element *	element
	)

Definition at line 52 of file FGFilter.cpp.

  : FGFCSComponent(fcs, element), DynamicFilter(false), Initialize(true)
{
  C[1] = C[2] = C[3] = C[4] = C[5] = C[6] = nullptr;
 
  CheckInputNodes(1, 1, element);
 
  auto PropertyManager = fcs->GetPropertyManager();
  for (int i=1; i<7; i++)
    ReadFilterCoefficients(element, i, PropertyManager);
 
  if      (Type == "LAG_FILTER")          FilterType = eLag        ;
  else if (Type == "LEAD_LAG_FILTER")     FilterType = eLeadLag    ;
  else if (Type == "SECOND_ORDER_FILTER") FilterType = eOrder2     ;
  else if (Type == "WASHOUT_FILTER")      FilterType = eWashout    ;
  else                                    FilterType = eUnknown    ;
 
  CalculateDynamicFilters();
 
  bind(element, PropertyManager.get());
 
  Debug(0);
}

~FGFilter()

~FGFilter

(

)

Definition at line 78 of file FGFilter.cpp.

{
  Debug(1);
}

Member Function Documentation

ResetPastStates()

void ResetPastStates ( void  )

overridevirtual

Reimplemented from FGFCSComponent.

Definition at line 85 of file FGFilter.cpp.

{
  FGFCSComponent::ResetPastStates();
 
  Input = 0.0; Initialize = true;
}

Run()

bool Run ( void  )

overridevirtual

Reimplemented from FGFCSComponent.

Definition at line 150 of file FGFilter.cpp.

{
  if (Initialize) {
 
    PreviousOutput2 = PreviousInput2 = PreviousOutput1 = PreviousInput1 = Output = Input;
    Initialize = false;
 
  } else {
 
    Input = InputNodes[0]->getDoubleValue();
 
    if (DynamicFilter) CalculateDynamicFilters();
 
    switch (FilterType) {
      case eLag:
        Output = (Input + PreviousInput1) * ca + PreviousOutput1 * cb;
        break;
      case eLeadLag:
        Output = Input * ca + PreviousInput1 * cb + PreviousOutput1 * cc;
        break;
      case eOrder2:
        Output = Input * ca + PreviousInput1 * cb + PreviousInput2 * cc
                            - PreviousOutput1 * cd - PreviousOutput2 * ce;
        break;
      case eWashout:
        Output = Input * ca - PreviousInput1 * ca + PreviousOutput1 * cb;
        break;
      case eUnknown:
        break;
    }
 
  }
 
  PreviousOutput2 = PreviousOutput1;
  PreviousOutput1 = Output;
  PreviousInput2  = PreviousInput1;
  PreviousInput1  = Input;
 
  Clip();
  SetOutput();
 
  return true;
}

---

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

• FGFilter.h
• FGFilter.cpp