IcePandel Class Reference

IceRecoBase derived TTask processor to perform Convoluted Pandel fitting. More...

#include "IcePandel.h"

Inheritance diagram for IcePandel:

Detailed Description

IceRecoBase derived TTask processor to perform Convoluted Pandel fitting.

Copyright(c) 2003, IceCube Experiment at the South Pole. All rights reserved.
 
Author: The IceCube NCFS-based Offline Project.
Contributors are mentioned in the code where appropriate.
 
Permission to use, copy, modify and distribute this software and its
documentation strictly for non-commercial purposes is hereby granted
without fee, provided that the above copyright notice appears in all
copies and that both the copyright notice and this permission notice
appear in the supporting documentation.
The authors make no claims about the suitability of this software for
any purpose. It is provided "as is" without express or implied warranty.

// Class IcePandel
// IceRecoBase derived TTask processor to perform track fitting via minimisation
// of a Gauss convoluted Pandel pdf.
//
// In case an event has been rejected by an NcEventSelector (based) processor,
// this task (and its sub-tasks) is not executed.
//
// The code in this processor is based on the algorithms as developed
// by Oladipo Fadiran, George Japaridze (Clark Atlanta University, USA)
// and Nick van Eijndhoven (Utrecht University, The Netherlands).
// For details please refer to Astroparticle Physics 28 (2007) 456-462.
//
// For the minimisation process the TFitter facility, which is basically Minuit,
// is used. Minimisation is performed by invokation of the SIMPLEX method,
// followed by an invokation of HESSE to determine the uncertainties on the results. 
// The statistics of the TFitter result are stored as an NcSignal object
// in the track, which can be obtained via the GetFitDetails memberfunction.
// In the minimisation procedure an overall plausibility for the fitted track
// is determined based on a Gauss convoluted Pandel pdf value for each used hit.
// This track plausibility is expressed in terms of a Bayesian psi value
// w.r.t. a Gauss Convoluted Pandel PDF.
// The Baysian psi value is defined as -loglikelihood in a decibel scale.
// This implies psi=-10*log10(L) where L=p(D|HI) being the likelihood of
// the data D under the hypothesis H and prior information I.
// Since all (associated) hits contribute independently to the Bayesian psi
// value, this psi value is built up by summation of the various hit contributions.
// As such, the FitDetails entries contain the statistics of all the different
// hit contributions, like PsiMedian, PsiMean, and PsiSigma.
// The Bayesian psi value is available in the fit details under the name "PsiSum".
// In addition the standard Minuit results like IERFIT, FCN, EDM etc... are
// also available from the FitDetails.
//
// The Gauss convoluted Pandel value is evaluated in various areas in the distance-time
// space as described in Astropart. Phys. 28 (2007) 456-462 by N. van Eijndhoven,
// O. Fadiran and G. Japaridze
// In case the distance-time point of a certain hit falls outside the
// validity rectangle, the point is moved onto the corresponding side location
// of the rectangle. For this new location the Pandel value is evaluated for
// this hit and an extra penalty is added to the corresponding psi value
// for this hit. 
// The penalty value can be specified via the memberfunction SetPenalty().
// This allows investigation/tuning of the sensitivity to hits with
// extreme distance and/or time residual values.
//
// As can be seen in the publication mentioned above, the pdf depends on
// the PMT time jitter, the photon scattering length and the photon absorption length.
// These parameters can be specified via the memberfunctions SetTimeJitter(),
// SetScatteringLength() and SetAbsorptionLength(), respectively. 
//
// A separate treatment of the phase and group velocities is introduced
// which will provide more accurate time residuals due to the different
// velocities of the Cerenkov wave front (v_phase) and the actually detected
// photons (v_group).
// This distinction between v_phase and v_group can be (de)activated via the
// memberfunction SetVgroupUsage().
//
// Use the UseTracks memberfunction to specify the first guess tracks
// to be processed by the minimiser (e.g. IceDwalkIC, IceDwalkI, IceDwalkDC).
//
// Use the SelectHits() memberfunction to specify the hits to be used.
//
// Information about the actual parameter settings can be found in the event
// structure itself via the device named "IcePandel".
//
// The fit processor printlevel can be selected via the memberfunction SetPrintLevel().
// By default all printout is suppressed (i.e. level=-2).
// 
// An example of how to invoke this processor after Xtalk, hit cleaning
// and a direct walk first guess estimate can be found in the ROOT example
// macro icepandel.cc which resides in the /macros subdirectory.
// 
// The minimisation results are stored in the IceEvent structure as
// tracks with as default the basename "IcePandel".
// This track basename identifier can be modified by the user via the
// SetTrackName() memberfunction. This will allow unique identification
// of tracks which are produced when re-processing existing data with
// different criteria.
// A suffix "4..." will be added to the tracks basename automatically according
// to the input track class that was used for the fit.
// For instance, the produced track name is IcePandel4DwalkIC when the input track
// was an IceDwalkIC track.
//
// The mass and charge of the track are left 0, since no distinction can
// be made between positive or negative tracks and also no mass can be determined.
// However, the user can define the track charge by invokation of the
// memberfunction SetCharge().
// This facility may be used to distinguish tracks produced by the
// various reconstruction algorithms in a (3D) colour display
// (see the class NcHelix for further details).  
//
// The produced track will be linked as a Track Hypothesis to the
// (first guess) track which was used as input, the so called parent track.
// As such, a pointer to the parent track is available via the GetParentTrack facility
// of the produced "IcePandel" tracks.
// Furthermore, all the hits that were used in the minisation are available
// via the GetSignal facility of the produced "IcePandel" tracks.
//
// An example of how the various data can be accessed is given below,
// where "evt" indicates the pointer to the IceEvent structure.
//
// Example for accessing data :
// ----------------------------
// TObjArray* tracks=evt->GetTracks("IcePandel4DwalkIC");
// if (!tracks) return;
// NcPosition* r0=0;
// Float_t fcn=0;
// for (Int_t jtk=0; jtk<tracks->GetEntries(); jtk++)
// {
//  NcTrack* tx=(NcTrack*)tracks->At(jtk);
//  if (!tx) continue;
//  tx->Data("sph");
//  r0=tx->GetReferencePoint();
//  if (r0) r0->Data();
//  sx=(NcSignal*)tx->GetFitDetails();
//  if (sx) fcn=sx->GetSignal("FCN");
//  NcTrack* parent=tx->GetParentTrack();
//  if (!parent) continue;
//  parent->Data("sph");
//  r0=parent->GetReferencePoint();
//  if (r0) r0->Data();
// }
//
// Notes :
// -------
// 1) This processor only works properly on data which are calibrated
//    and contain tracks from first guess algorithms like e.g. IceDwalk.
// 2) In view of the usage of TFitter/Minuit minimisation, a global pointer
//    to the instance of this class (gIcePandel) and a global static
//    wrapper function (IcePandelFCN) have been introduced, to allow the
//    actual minimisation to be performed via the memberfunction FitFCN.
//    This implies that in a certain processing job only 1 instance of
//    this IcePandel class may occur.
//
//--- Author: Nick van Eijndhoven 09-feb-2006 Utrecht University
//- Modified: NvE $Date: 2013-02-27 23:43:36 +0100 (Wed, 27 Feb 2013) $ IIHE-VUB, Brussels

Definition at line 17 of file IcePandel.h.

Public Member Functions
	IcePandel (const char *name="IcePandel", const char *title="Gauss convoluted Pandel fitting")
virtual	~IcePandel ()
virtual void	Exec (Option_t *opt)
void	FitFCN (Int_t &, Double_t *, Double_t &, Double_t *, Int_t)
void	SelectHits (Int_t mode=1)
void	SetPenalty (Float_t val)
void	SetPrintLevel (Int_t level)
Public Member Functions inherited from IceRecoBase
	IceRecoBase (const char *name="IceRecoBase", const char *title="Base class for IceCube reconstruction tasks")
virtual	~IceRecoBase ()
void	SetAbsorptionLength (Float_t lambda, TString s)
void	SetCharge (Float_t charge)
void	SetCleaned (Int_t flag, TString s)
void	SetFlipAngles (Float_t thetatrk, Float_t thetahits)
void	SetMaxHits (Int_t nmax, TString s)
void	SetMaxMod (Int_t nmax, TString s)
void	SetMinAhits (Int_t nmin, TString s)
void	SetMinAmods (Int_t nmin, TString s)
void	SetMinMod (Int_t nmin, TString s)
void	SetScatteringLength (Float_t lambda, TString s)
void	SetSingleHit (Int_t ndoms, TString s, Int_t ndoms1=0)
void	SetSLChitUsage (Int_t flag, TString s)
void	SetTimeJitter (Float_t sigma, TString s)
void	SetTrackName (TString s)
void	SetVgroupUsage (Int_t flag, TString s)
void	UseTracks (TString classname, Int_t n=-1)

Protected Attributes
NcSignal *	fFitstats
TFitter *	fFitter
TObjArray *	fHits
Float_t	fPenalty
Int_t	fPrint
NcSample	fPsistats
Int_t	fSelhits
NcTrack *	fTkfit
Protected Attributes inherited from IceRecoBase
Float_t	fCharge
Int_t	fCleanA
Int_t	fCleanDC
Int_t	fCleanI
Int_t	fCleanIC
IceEvent *	fEvt
Int_t	fFirst
Float_t	fLabsA
Float_t	fLabsDL
Float_t	fLabsLD
Float_t	fLabsUD
Float_t	fLambdaA
Float_t	fLambdaDL
Float_t	fLambdaLD
Float_t	fLambdaUD
Int_t	fMaxhitsA
Int_t	fMaxhitsDC
Int_t	fMaxhitsI
Int_t	fMaxhitsIC
Int_t	fMaxmodA
Int_t	fMaxmodDC
Int_t	fMaxmodI
Int_t	fMaxmodIC
Int_t	fMinahitsA
Int_t	fMinahitsDC
Int_t	fMinahitsI
Int_t	fMinahitsIC
Int_t	fMinamodsA
Int_t	fMinamodsDC
Int_t	fMinamodsI
Int_t	fMinamodsIC
Int_t	fMinmodA
Int_t	fMinmodDC
Int_t	fMinmodI
Int_t	fMinmodIC
NcDevice	fParams
Int_t	fSingle1A
Int_t	fSingle1DC
Int_t	fSingle1I
Int_t	fSingle1IC
Int_t	fSingleA
Int_t	fSingleDC
Int_t	fSingleI
Int_t	fSingleIC
Int_t	fSlcDC
Int_t	fSlcI
Int_t	fSlcIC
Float_t	fThetahits
Float_t	fThetatrk
TString	fTrackname
Float_t	fTsigmaA
Float_t	fTsigmaDC
Float_t	fTsigmaIC
TObjArray *	fUseNames
TArrayI *	fUseNtk
Int_t	fVgroupA
Int_t	fVgroupDC
Int_t	fVgroupI
Int_t	fVgroupIC

Additional Inherited Members
Protected Member Functions inherited from IceRecoBase
virtual void	FlipTrack (NcTrack *t) const

Constructor & Destructor Documentation

IcePandel()

IcePandel::IcePandel	(	const char *	name = "IcePandel",
		const char *	title = "Gauss convoluted Pandel fitting" )

// Default constructor.

Definition at line 175 of file IcePandel.cxx.

~IcePandel()

IcePandel::~IcePandel ( )

virtual

// Default destructor.

Definition at line 230 of file IcePandel.cxx.

Member Function Documentation

Exec()

void IcePandel::Exec ( Option_t * opt )

virtual

// Implementation of the Convoluted Pandel hit fitting procedure.

Reimplemented from IceRecoBase.

Definition at line 319 of file IcePandel.cxx.

FitFCN()

void IcePandel::FitFCN	(	Int_t &	,
		Double_t *	,
		Double_t &	f,
		Double_t *	x,
		Int_t	)

// Minimisation of the Bayesian psi value for a track w.r.t. a Convoluted Pandel PDF.
// The Baysian psi value is defined as -loglikelihood in a decibel scale.
// This implies psi=-10*log10(L) where L=p(D|HI) being the likelihood of
// the data D under the hypothesis H and prior information I.

Definition at line 724 of file IcePandel.cxx.

SelectHits()

void IcePandel::SelectHits

(

Int_t

mode = 1

)

// Specification of the hits to be used in the minimisation.
//
// mode = 0 : All hits (except IceTop) of the complete event are used.
//        1 : Only the associated hits are used for each first guess track.
//        2 : All hits are used of those detector systems (e.g. Amanda, InIce)
//            that were used to construct the first guess track.
//
// Note : The user can select to use only hits which survived a hit cleaning processor
//        via the SetClean() memberfunction.
//
// For the default value please refer to the default constructor of this class.

Definition at line 276 of file IcePandel.cxx.

SetPenalty()

void IcePandel::SetPenalty

(

Float_t

val

)

// Set user defined psi penalty value (in dB) in the minimiser for
// distance-time points that fall outside the validity rectangle.
// This allows investigation/tuning of the sensitivity to hits with
// extreme distance and/or time residual values.
// For the default value please refer to the default constructor of this class.

Definition at line 302 of file IcePandel.cxx.

SetPrintLevel()

void IcePandel::SetPrintLevel

(

Int_t

level

)

// Set the fitter (Minuit) print level.
// See the TFitter and TMinuit docs for details.
//
// Note : level=-2 suppresses also all fit processor warnings.
//        
// The default in the constructor is level=-2. 

Definition at line 260 of file IcePandel.cxx.

Member Data Documentation

fFitstats

NcSignal* IcePandel::fFitstats

protected

Definition at line 35 of file IcePandel.h.

fFitter

TFitter* IcePandel::fFitter

protected

Definition at line 32 of file IcePandel.h.

fHits

TObjArray* IcePandel::fHits

protected

Definition at line 31 of file IcePandel.h.

fPenalty

Float_t IcePandel::fPenalty

protected

Definition at line 33 of file IcePandel.h.

fPrint

Int_t IcePandel::fPrint

protected

Definition at line 29 of file IcePandel.h.

fPsistats

NcSample IcePandel::fPsistats

protected

Definition at line 36 of file IcePandel.h.

fSelhits

Int_t IcePandel::fSelhits

protected

Definition at line 30 of file IcePandel.h.

fTkfit

NcTrack* IcePandel::fTkfit

protected

Definition at line 34 of file IcePandel.h.

---

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

• icepack/source/IcePandel.h
• icepack/source/IcePandel.cxx