GENFIT Rev: NoNumberAvailable
Loading...
Searching...
No Matches
MplTrackRep.cc
Go to the documentation of this file.
1/* Copyright 2019, Belle II Collaboration
2 Authors: Dmitrii Neverov
3
4 This file is part of GENFIT.
5
6 GENFIT is free software: you can redistribute it and/or modify
7 it under the terms of the GNU Lesser General Public License as published
8 by the Free Software Foundation, either version 3 of the License, or
9 (at your option) any later version.
10
11 GENFIT is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU Lesser General Public License for more details.
15
16 You should have received a copy of the GNU Lesser General Public License
17 along with GENFIT. If not, see <http://www.gnu.org/licenses/>.
18*/
19
20#include "MplTrackRep.h"
21
22#include <FieldManager.h>
23#include <TDatabasePDG.h>
24#include <MeasurementOnPlane.h>
25#include <Exception.h>
26
27#include <math.h>
28
29#include <TBuffer.h>
30
31using namespace genfit;
32
33MplTrackRep::MplTrackRep(int pdgCode, float magCharge, char propDir) :
34 RKTrackRep(pdgCode, propDir),
35 m_magCharge(magCharge),
36 m_mass(TDatabasePDG::Instance()->GetParticle(pdgCode)->Mass()) // We could ofc use AbsTrackRep::getMass(state) but we have no state here to call on
37{
38}
39
43
44double MplTrackRep::getCharge(const StateOnPlane& state) const {
45
46 if (dynamic_cast<const MeasurementOnPlane*>(&state) != nullptr) {
47 Exception exc("RKTrackRep::getCharge - cannot get charge from MeasurementOnPlane",__LINE__,__FILE__);
48 exc.setFatal();
49 throw exc;
50 }
51
52 double pdgCharge( m_magCharge * (this->getPDGCharge() > 0 ? 1.0 : -1.0));
53
54 // return pdgCharge with sign of q/p
55 if (state.getState()(0) * pdgCharge < 0)
56 return -pdgCharge;
57 else
58 return pdgCharge;
59}
60
62 M7x7* jacobianT,
63 M1x3& SA,
64 double S,
65 bool varField,
66 bool /*calcOnlyLastRowOfJ*/) const
67{
68 // The algorithm is
69 // E Lund et al 2009 JINST 4 P04001 doi:10.1088/1748-0221/4/04/P04001
70 // "Track parameter propagation through the application of a new adaptive Runge-Kutta-Nyström method in the ATLAS experiment"
71 // http://inspirehep.net/search?ln=en&ln=en&p=10.1088/1748-0221/4/04/P04001&of=hb&action_search=Search&sf=earliestdate&so=d&rm=&rg=25&sc=0
72 // where the transport of the Jacobian is described in
73 // L. Bugge, J. Myrheim Nucl.Instrum.Meth. 160 (1979) 43-48
74 // "A Fast Runge-kutta Method For Fitting Tracks In A Magnetic Field"
75 // http://inspirehep.net/record/145692
76 // and
77 // L. Bugge, J. Myrheim Nucl.Instrum.Meth. 179 (1981) 365-381
78 // "Tracking And Track Fitting"
79 // http://inspirehep.net/record/160548
80
81 // important fixed numbers
82 static const double EC ( 0.000149896229 ); // c/(2*10^12) resp. c/2Tera FIXME this 1/2 here is super sneaky
83 static const double P3 ( 1./3. ); // 1/3
84 static const double DLT ( .0002 ); // max. deviation for approximation-quality test
85 double sign = state7[6] > 0 ? 1.0 : -1.0;
86 // Aux parameters
87 M1x3& R = *((M1x3*) &state7[0]); // Start coordinates [cm] (x, y, z)
88 M1x3& A = *((M1x3*) &state7[3]); // Start directions (ax, ay, az); ax^2+ay^2+az^2=1
89 double S3(0), S4(0), PS2(0);
90 M1x3 H0 = {{0.,0.,0.}}, H1 = {{0.,0.,0.}}, H2 = {{0.,0.,0.}};
91 M1x3 r = {{0.,0.,0.}};
92 // Variables for Runge Kutta solver
93 double A0(0), A1(0), A2(0), A3(0), A4(0), A5(0), A6(0);
94 double B0(0), B1(0), B2(0), B3(0), B4(0), B5(0), B6(0);
95 double C0(0), C1(0), C2(0), C3(0), C4(0), C5(0), C6(0);
96 // Additional variables for momentum evolution FIXME these are all cryptic in accordance with the rest of the code around
97 double D0(0), D1(0), D2(0), D3(0), D4(0), D5(0);
98 double F0(0), F1(0), F2(0), F3(0);
99 double AH0(0), AH1(0), AH2(0), AH3(0);
100
101 //
102 // Runge Kutta Extrapolation
103 //
104 S3 = P3*S;
105 S4 = 0.25*S;
106 PS2 = m_magCharge * EC * S * sign;
107
108 // First point
109 r[0] = R[0]; r[1] = R[1]; r[2]=R[2];
110 FieldManager::getInstance()->getFieldVal(r[0], r[1], r[2], H0[0], H0[1], H0[2]); // magnetic field in 10^-1 T = kGauss
111 H0[0] *= PS2; H0[1] *= PS2; H0[2] *= PS2; // H0 is PS2*(Hx, Hy, Hz) @ R0; effectively this is h/2 * Force
112 D0 = fabs(m_magCharge/state7[6]); // p_n
113 F0 = std::sqrt(m_mass * m_mass + D0 * D0) / (D0 * D0); // E / p^2
114 AH0 = A[0]*H0[0] + A[1]*H0[1] + A[2]*H0[2]; // A dot Force
115
116 A0 = F0 * (H0[0] - A[0] * AH0); B0 = F0 * (H0[1] - A[1] * AH0); C0 = F0 * (H0[2] - A[2] * AH0); // h/2 * k_1
117 A2 = A[0]+A0 ; B2 = A[1]+B0 ; C2 = A[2]+C0 ; // r'_n + h/2 * k_1
118 A1 = A2+A[0] ; B1 = B2+A[1] ; C1 = C2+A[2] ; // 2*r'_n + h/2 * k_1
119
120 // Second point
121 if (varField) {
122 r[0] += A1*S4; r[1] += B1*S4; r[2] += C1*S4;
123 FieldManager::getInstance()->getFieldVal(r[0], r[1], r[2], H1[0], H1[1], H1[2]);
124 H1[0] *= PS2; H1[1] *= PS2; H1[2] *= PS2; // H1 is PS2*(Hx, Hy, Hz) @ (x, y, z) + 0.25*S * [(A0, B0, C0) + 2*(ax, ay, az)]
125 }
126 else { H1 = H0; };
127 D1 = D0 + F0 * D0 * AH0; // p_n + h/2 * l_1
128 F1 = std::sqrt(m_mass * m_mass + D1 * D1) / (D1 * D1); // E / p^2
129 AH1 = A2*H1[0] + B2*H1[1] + C2*H1[2]; // A dot Force
130
131 A3 = A[0] + F1*(H1[0] - A2*AH1); B3 = A[1] + F1*(H1[1] - B2*AH1); C3 = A[2] + F1*(H1[2] - C2*AH1); // r'_n + h/2 * k_2
132 D2 = D0 + F1 * D1 * AH1; // p_n + h/2 * l_2
133 F2 = std::sqrt(m_mass * m_mass + D2 * D2) / (D2 * D2); // E / p^2
134 AH2 = A3*H1[0] + B3*H1[1] + C3*H1[2]; // A dot Force
135
136 A4 = A[0] + F2*(H1[0] - A3*AH2); B4 = A[1] + F2*(H1[1] - B3*AH2); C4 = A[2] + F2*(H1[2] - C3*AH2); // r'_n + h/2 * k_3
137 A5 = A4-A[0]+A4 ; B5 = B4-A[1]+B4 ; C5 = C4-A[2]+C4 ; // r'_n + h * k_3
138 D3 = D0 + 2.0 * F2 * D2 * AH2; // p_n + h * l_3
139 F3 = std::sqrt(m_mass * m_mass + D3 * D3) / (D3 * D3); // E / p^2
140 AH3 = A4*H1[0] + B4*H1[1] + C4*H1[2]; // A dot Force
141
142 // Last point
143 if (varField) {
144 r[0]=R[0]+S*A4; r[1]=R[1]+S*B4; r[2]=R[2]+S*C4; //setup.Field(r,H2);
145 FieldManager::getInstance()->getFieldVal(r[0], r[1], r[2], H2[0], H2[1], H2[2]);
146 H2[0] *= PS2; H2[1] *= PS2; H2[2] *= PS2; // H2 is PS2*(Hx, Hy, Hz) @ (x, y, z) + 0.25*S * (A4, B4, C4)
147 }
148 else { H2 = H0; };
149 A6 = F3 * (H2[0] - A5*AH3); B6 = F3 * (H2[1] - B5*AH3); C6 = F3 * (H2[2] - C5*AH3); // h/2 * k_4
150 D4 = F3 * D3 * AH3 - D0; // h/2 * l_4 - p_n
151 D5 = P3*(D1 + 2*D2 + D3 + D4); //p_n+1
152
153 //
154 // Derivatives of track parameters
155 //
156 if(jacobianT != nullptr){
157
158 // jacobianT //FIXME seems in magnetic case there are no shortcuts?
159 // 1 0 0 0 0 0 0 x
160 // 0 1 0 0 0 0 0 y
161 // 0 0 1 0 0 0 0 z
162 // x x x x x x 0 a_x
163 // x x x x x x 0 a_y
164 // x x x x x x 0 a_z
165 // x x x x x x 1 q/p
166 M7x7& J = *jacobianT;
167
168 // cppcheck-suppress unreadVariable
169 double dA0(0), dA2(0), dA3(0), dA4(0), dA5(0), dA6(0);
170 // cppcheck-suppress unreadVariable
171 double dB0(0), dB2(0), dB3(0), dB4(0), dB5(0), dB6(0);
172 // cppcheck-suppress unreadVariable
173 double dC0(0), dC2(0), dC3(0), dC4(0), dC5(0), dC6(0);
174 // cppcheck-suppress unreadVariable
175 double dD0(0), dD1(0), dD2(0), dD3(0), dD4(0);
176
177 int start(0);
178
179// if (!calcOnlyLastRowOfJ) {
180
181// if (!varField) { // FIXME let's be honest and calculate everything everytime
182// // d(x, y, z)/d(x, y, z) submatrix is unit matrix
183// J(0, 0) = 1; J(1, 1) = 1; J(2, 2) = 1;
184// // d(ax, ay, az)/d(ax, ay, az) submatrix is 0
185// // start with d(x, y, z)/d(ax, ay, az)
186// start = 3;
187// }
188
189 for(int i=start; i<7; ++i) {
190
191 //first point
192 dD0 = -D0*D0/m_magCharge/sign*J(i,6);
193 dA0 = (1/(F0*F0*D0*D0*D0) - 2/D0)*A0*dD0 - (D1-D0)/D0*J(i,3) - F0*A[0]*(J(i,3)*H0[0] + J(i,4)*H0[1] + J(i,5)*H0[2]); // FIXME A true marvel of clarity
194 dB0 = (1/(F0*F0*D0*D0*D0) - 2/D0)*B0*dD0 - (D1-D0)/D0*J(i,4) - F0*A[1]*(J(i,3)*H0[0] + J(i,4)*H0[1] + J(i,5)*H0[2]);
195 dC0 = (1/(F0*F0*D0*D0*D0) - 2/D0)*C0*dD0 - (D1-D0)/D0*J(i,5) - F0*A[2]*(J(i,3)*H0[0] + J(i,4)*H0[1] + J(i,5)*H0[2]);
196
197 dD1 = dD0 + (1/(F0*F0*D0*D0*D0) - 1/D0)*(D1-D0)*dD0 + F0*D0*(J(i,3)*H0[0] + J(i,4)*H0[1] + J(i,5)*H0[2]);
198 dA2 = dA0+J(i, 3);
199 dB2 = dB0+J(i, 4);
200 dC2 = dC0+J(i, 5);
201
202 //second point
203 dD2 = dD0 + (1/(F1*F1*D1*D1*D1) - 1/D1)*(D2-D0)*dD1 + F1*D1*(dA2*H1[0] + dB2*H1[1] + dC2*H1[2]);
204 dA3 = J(i,3)+(1/(F1*F1*D1*D1*D1) - 2/D1)*(A2-A[0])*dD1 - (D2-D0)/D1*dA2 - F1*A2*(dA2*H1[0] + dB2*H1[1] + dC2*H1[2]); // FIXME it's only getting better
205 dB3 = J(i,4)+(1/(F1*F1*D1*D1*D1) - 2/D1)*(B2-A[1])*dD1 - (D2-D0)/D1*dB2 - F1*B2*(dA2*H1[0] + dB2*H1[1] + dC2*H1[2]);
206 dC3 = J(i,5)+(1/(F1*F1*D1*D1*D1) - 2/D1)*(C2-A[2])*dD1 - (D2-D0)/D1*dC2 - F1*C2*(dA2*H1[0] + dB2*H1[1] + dC2*H1[2]);
207
208 dD3 = dD0 + 2*(1/(F2*F2*D2*D2*D2) - 1/D2)*(D3-D0)*dD2 + 2*F2*D2*(dA3*H1[0] + dB3*H1[1] + dC3*H1[2]);
209 dA4 = J(i, 3)+(1/(F2*F2*D2*D2*D2) - 2/D2)*(A3-A[0])*dD2 - (D3-D0)/D2*dA3 - F2*A3*(dA3*H1[0] + dB3*H1[1] + dC3*H1[2]);
210 dB4 = J(i, 4)+(1/(F2*F2*D2*D2*D2) - 2/D2)*(B3-A[1])*dD2 - (D3-D0)/D2*dB3 - F2*B3*(dA3*H1[0] + dB3*H1[1] + dC3*H1[2]);
211 dC4 = J(i, 5)+(1/(F2*F2*D2*D2*D2) - 2/D2)*(C3-A[2])*dD2 - (D3-D0)/D2*dC3 - F2*C3*(dA3*H1[0] + dB3*H1[1] + dC3*H1[2]);
212
213 //last point
214 dA5 = dA4+dA4-J(i, 3); // }
215 dB5 = dB4+dB4-J(i, 4); // } = 2*(dA4, dB4, dC4) - dA
216 dC5 = dC4+dC4-J(i, 5); // }
217
218 dD4 = -dD0+(1/(F3*F3*D3*D3*D3) - 1/D3)*(D4+D0)*dD3 + F3*D3*(dA5*H2[0] + dB5*H2[1] + dC5*H2[2]);
219 dA6 = (1/(F3*F3*D3*D3*D3) - 2/D3)*(A4-A[0])*dD3 - (D4+D0)/D3*dA5 - F3*A5*(dA5*H2[0] + dB5*H2[1] + dC5*H2[2]);
220 dB6 = (1/(F3*F3*D3*D3*D3) - 2/D3)*(B4-A[1])*dD3 - (D4+D0)/D3*dB5 - F3*B5*(dA5*H2[0] + dB5*H2[1] + dC5*H2[2]);
221 dC6 = (1/(F3*F3*D3*D3*D3) - 2/D3)*(C4-A[2])*dD3 - (D4+D0)/D3*dC5 - F3*C5*(dA5*H2[0] + dB5*H2[1] + dC5*H2[2]);
222
223 // this gives the same results as multiplying the old with the new Jacobian
224 J(i, 0) += (dA2+dA3+dA4)*S3; J(i, 3) = ((dA0+2.*dA3)+(dA5+dA6))*P3; // dR := dR + S3*[(dA2, dB2, dC2) + (dA3, dB3, dC3) + (dA4, dB4, dC4)]
225 J(i, 1) += (dB2+dB3+dB4)*S3; J(i, 4) = ((dB0+2.*dB3)+(dB5+dB6))*P3; // dA := 1/3*[(dA0, dB0, dC0) + 2*(dA3, dB3, dC3) + (dA5, dB5, dC5) + (dA6, dB6, dC6)]
226 J(i, 2) += (dC2+dC3+dC4)*S3; J(i, 5) = ((dC0+2.*dC3)+(dC5+dC6))*P3;
227 J(i,6) = -m_magCharge*sign/D5/D5*P3*(dD1 + 2*dD2 + dD3 + dD4);
228 }
229
230// } // end if (!calcOnlyLastRowOfJ)
231
232 }
233 //
234 // Track parameters in last point
235 //
236 R[0] += (A2+A3+A4)*S3; A[0] += (SA[0]=((A0+2.*A3)+(A5+A6))*P3-A[0]); // R = R0 + S3*[(A2, B2, C2) + (A3, B3, C3) + (A4, B4, C4)]
237 R[1] += (B2+B3+B4)*S3; A[1] += (SA[1]=((B0+2.*B3)+(B5+B6))*P3-A[1]); // A = 1/3*[(A0, B0, C0) + 2*(A3, B3, C3) + (A5, B5, C5) + (A6, B6, C6)]
238 R[2] += (C2+C3+C4)*S3; A[2] += (SA[2]=((C0+2.*C3)+(C5+C6))*P3-A[2]); // SA = A_new - A_old
239 state7[6] = m_magCharge * sign / D5; // g / p_n+1
240
241 // normalize A
242 double CBA ( 1./sqrt(A[0]*A[0]+A[1]*A[1]+A[2]*A[2]) ); // 1/|A|
243 A[0] *= CBA; A[1] *= CBA; A[2] *= CBA;
244
245
246 // Test approximation quality on given step
247 double EST ( fabs((A1+A6)-(A3+A4)) +
248 fabs((B1+B6)-(B3+B4)) +
249 fabs((C1+C6)-(C3+C4)) ); // EST = ||(ABC1+ABC6)-(ABC3+ABC4)||_1 = ||(axzy x H0 + ABC5 x H2) - (ABC2 x H1 + ABC3 x H1)||_1
250 if (debugLvl_ > 0) {
251 debugOut << " RKTrackRep::RKPropagate. Step = "<< S << "; quality EST = " << EST << " \n";
252 }
253
254 // Prevent the step length increase from getting too large, this is
255 // just the point where it becomes 10.
256 if (EST < DLT*1e-5)
257 return 10;
258
259 // Step length increase for a fifth order Runge-Kutta, see e.g. 17.2
260 // in Numerical Recipes. FIXME: move to caller.
261 return pow(DLT/EST, 1./5.);
262}
263
264void MplTrackRep::Streamer(TBuffer &R__b)
265{
266 // I guess I have to reimplement this since it can not be taken from RKTrackRep?
267 typedef ::genfit::MplTrackRep thisClass;
268 UInt_t R__s, R__c;
269 if (R__b.IsReading()) {
270 ::genfit::AbsTrackRep::Streamer(R__b);
271 Version_t R__v = R__b.ReadVersion(&R__s, &R__c); if (R__v) { }
272 R__b.CheckByteCount(R__s, R__c, thisClass::IsA());
274 lastEndState_.setRep(this);
275 } else {
276 ::genfit::AbsTrackRep::Streamer(R__b);
277 R__c = R__b.WriteVersion(thisClass::IsA(), kTRUE);
278 R__b.SetByteCount(R__c, kTRUE);
279 }
280}
double getPDGCharge() const
Get the charge of the particle of the pdg code.
unsigned int debugLvl_
Exception class for error handling in GENFIT (provides storage for diagnostic information)
Definition Exception.h:48
void setFatal(bool b=true)
Set fatal flag.
Definition Exception.h:61
TVector3 getFieldVal(const TVector3 &position)
This does NOT use the cache!
static FieldManager * getInstance()
Get singleton instance.
Measured coordinates on a plane.
double RKPropagate(M1x7 &state7, M7x7 *jacobian, M1x3 &SA, double S, bool varField=true, bool calcOnlyLastRowOfJ=false) const override
The actual Runge Kutta propagation.
double getCharge(const StateOnPlane &state) const override
Get the (fitted) charge of a state. This is not always equal the pdg charge (e.g. if the charge sign ...
const double m_mass
Definition MplTrackRep.h:55
const double m_magCharge
Definition MplTrackRep.h:54
StateOnPlane lastStartState_
Definition RKTrackRep.h:289
StateOnPlane lastEndState_
state where the last extrapolation has started
Definition RKTrackRep.h:290
A state with arbitrary dimension defined in a DetPlane.
const TVectorD & getState() const
void setRep(const AbsTrackRep *rep)
Defines for I/O streams used for error and debug printing.
RKMatrix< 1, 7 > M1x7
Definition RKTools.h:68
RKMatrix< 7, 7 > M7x7
Definition RKTools.h:71
std::ostream debugOut
RKMatrix< 1, 3 > M1x3
Definition RKTools.h:66