]>
Commit | Line | Data |
---|---|---|
d20df6fc | 1 | /**************************************************************************\r |
2 | * Copyright(c) 1998-1999, ALICE Experiment at CERN, All rights reserved. *\r | |
3 | * *\r | |
4 | * Author: The ALICE Off-line Project. *\r | |
5 | * Contributors are mentioned in the code where appropriate. *\r | |
6 | * *\r | |
7 | * Permission to use, copy, modify and distribute this software and its *\r | |
8 | * documentation strictly for non-commercial purposes is hereby granted *\r | |
9 | * without fee, provided that the above copyright notice appears in all *\r | |
10 | * copies and that both the copyright notice and this permission notice *\r | |
11 | * appear in the supporting documentation. The authors make no claims *\r | |
12 | * about the suitability of this software for any purpose. It is *\r | |
13 | * provided "as is" without express or implied warranty. *\r | |
14 | **************************************************************************/\r | |
15 | \r | |
4e459a9d | 16 | /* $Id: AliTRDtrackingChamber.cxx 23810 2008-02-08 09:00:27Z hristov $ */\r |
d20df6fc | 17 | \r |
18 | ////////////////////////////////////////////////////////////////////////////\r | |
19 | // //\r | |
20 | // Tracking in one chamber //\r | |
21 | // //\r | |
22 | // Authors: //\r | |
23 | // Alex Bercuci <A.Bercuci@gsi.de> //\r | |
24 | // Markus Fasel <M.Fasel@gsi.de> //\r | |
25 | // //\r | |
26 | ////////////////////////////////////////////////////////////////////////////\r | |
27 | \r | |
28 | #include "AliTRDtrackingChamber.h"\r | |
29 | \r | |
30 | #include "TMath.h"\r | |
31 | #include "TMatrixTBase.h"\r | |
32 | #include <TTreeStream.h>\r | |
33 | \r | |
34 | #include "AliTRDReconstructor.h"\r | |
35 | #include "AliTRDrecoParam.h"\r | |
36 | #include "AliTRDtrackerV1.h"\r | |
37 | #include "AliTRDgeometry.h"\r | |
38 | #include "AliTRDpadPlane.h"\r | |
39 | #include "AliTRDcalibDB.h"\r | |
40 | \r | |
41 | ClassImp(AliTRDtrackingChamber)\r | |
42 | \r | |
43 | //_______________________________________________________\r | |
44 | AliTRDtrackingChamber::AliTRDtrackingChamber(Int_t det) :\r | |
45 | fDetector(det)\r | |
46 | ,fX0(0.)\r | |
47 | {} \r | |
48 | \r | |
49 | //_______________________________________________________\r | |
50 | void AliTRDtrackingChamber::Clear(const Option_t *opt)\r | |
51 | {\r | |
52 | for(Int_t itb=0; itb<kNTimeBins; itb++) fTB[itb].Clear(opt);\r | |
53 | }\r | |
54 | \r | |
55 | //_______________________________________________________\r | |
56 | void AliTRDtrackingChamber::InsertCluster(AliTRDcluster *c, Int_t index)\r | |
57 | {\r | |
3a039a31 | 58 | fTB[c->GetPadTime()].InsertCluster(c, index);\r |
d20df6fc | 59 | }\r |
60 | \r | |
61 | //_______________________________________________________\r | |
62 | Bool_t AliTRDtrackingChamber::Build(AliTRDgeometry *geo)\r | |
63 | {\r | |
64 | // Init chamber and all time bins (AliTRDchamberTimeBin)\r | |
65 | // Calculates radial position of the chamber based on \r | |
66 | // radial positions of the time bins (calibration/alignment aware)\r | |
67 | //\r | |
053767a4 | 68 | Int_t stack = geo->GetStack(fDetector);\r |
69 | Int_t layer = geo->GetLayer(fDetector);\r | |
70 | AliTRDpadPlane *pp = geo->GetPadPlane(layer, stack);\r | |
d20df6fc | 71 | Double_t zl = pp->GetRow0ROC() - pp->GetRowEndROC();\r |
053767a4 | 72 | Double_t z0 = geo->GetRow0(layer, stack, 0) - zl;\r |
d20df6fc | 73 | Int_t nrows = pp->GetNrows();\r |
74 | \r | |
75 | Int_t index[50], jtb = 0;\r | |
76 | for(Int_t itb=0; itb<kNTimeBins; itb++){ \r | |
77 | if(!fTB[itb]) continue;\r | |
78 | fTB[itb].SetRange(z0, zl);\r | |
79 | fTB[itb].SetNRows(nrows);\r | |
80 | fTB[itb].BuildIndices();\r | |
81 | index[jtb++] = itb;\r | |
82 | } \r | |
83 | if(jtb<2) return kFALSE;\r | |
84 | \r | |
85 | \r | |
86 | // ESTIMATE POSITION OF PAD PLANE FOR THIS CHAMBER\r | |
87 | Double_t x0 = fTB[index[0]].GetX();\r | |
88 | Double_t x1 = fTB[index[1]].GetX();\r | |
89 | Double_t dx = (x0 - x1)/(index[1] - index[0]); \r | |
90 | fX0 = x0 + dx*(index[0] - (Int_t)AliTRDcalibDB::Instance()->GetT0Average(fDetector)); \r | |
91 | return kTRUE;\r | |
92 | }\r | |
93 | \r | |
94 | //_______________________________________________________ \r | |
95 | Int_t AliTRDtrackingChamber::GetNClusters() const\r | |
96 | {\r | |
97 | // Returns number of clusters in chamber\r | |
98 | //\r | |
99 | Int_t n = 0;\r | |
100 | for(Int_t itb=0; itb<kNTimeBins; itb++){ \r | |
101 | n += Int_t(fTB[itb]);\r | |
102 | }\r | |
103 | return n; \r | |
104 | } \r | |
105 | \r | |
106 | //_______________________________________________________\r | |
107 | Double_t AliTRDtrackingChamber::GetQuality()\r | |
108 | {\r | |
109 | //\r | |
110 | // Calculate chamber quality for seeding.\r | |
111 | // \r | |
112 | //\r | |
113 | // Parameters :\r | |
114 | // layers : Array of propagation layers for this plane.\r | |
115 | //\r | |
116 | // Output :\r | |
117 | // plane quality factor for seeding\r | |
118 | // \r | |
119 | // Detailed description\r | |
120 | //\r | |
121 | // The quality of the plane for seeding is higher if:\r | |
122 | // 1. the average timebin population is closer to an integer number\r | |
123 | // 2. the distribution of clusters/timebin is closer to a uniform distribution.\r | |
124 | // - the slope of the first derivative of a parabolic fit is small or\r | |
125 | // - the slope of a linear fit is small\r | |
126 | //\r | |
127 | \r | |
128 | Int_t ncl = 0;\r | |
129 | Int_t nused = 0;\r | |
130 | Int_t nClLayer;\r | |
131 | for(int itb=0; itb<kNTimeBins; itb++){\r | |
132 | if(!(nClLayer = fTB[itb].GetNClusters())) continue;\r | |
133 | ncl += nClLayer;\r | |
134 | for(Int_t incl = 0; incl < nClLayer; incl++){\r | |
135 | if((fTB[itb].GetCluster(incl))->IsUsed()) nused++;\r | |
136 | }\r | |
137 | }\r | |
138 | \r | |
139 | // calculate the deviation of the mean number of clusters from the\r | |
140 | // closest integer values\r | |
141 | Float_t nclMed = float(ncl-nused)/AliTRDtrackerV1::GetNTimeBins();\r | |
142 | Int_t ncli = Int_t(nclMed);\r | |
143 | Float_t nclDev = TMath::Abs(nclMed - TMath::Max(ncli, 1));\r | |
144 | nclDev -= (nclDev>.5) && ncli ? 1. : 0.;\r | |
145 | return TMath::Exp(-5.*TMath::Abs(nclDev));\r | |
146 | \r | |
147 | // // get slope of the derivative\r | |
148 | // if(!fitter.Eval()) return quality;\r | |
149 | // fitter.PrintResults(3);\r | |
150 | // Double_t a = fitter.GetParameter(1);\r | |
151 | // \r | |
152 | // printf("ncl_dev(%f) a(%f)\n", ncl_dev, a);\r | |
153 | // return quality*TMath::Exp(-a);\r | |
154 | \r | |
155 | }\r | |
156 | \r | |
157 | \r | |
158 | //_______________________________________________________\r | |
3a039a31 | 159 | Bool_t AliTRDtrackingChamber::GetSeedingLayer(AliTRDchamberTimeBin *&fakeLayer, AliTRDgeometry *geo, const AliTRDReconstructor *rec)\r |
d20df6fc | 160 | {\r |
161 | //\r | |
162 | // Creates a seeding layer\r | |
163 | //\r | |
164 | \r | |
165 | // constants\r | |
166 | const Int_t kMaxRows = 16;\r | |
167 | const Int_t kMaxCols = 144;\r | |
168 | const Int_t kMaxPads = 2304;\r | |
3a039a31 | 169 | Int_t timeBinMin = rec->GetRecoParam()->GetNumberOfPresamples();\r |
170 | Int_t timeBinMax = rec->GetRecoParam()->GetNumberOfPostsamples();\r | |
4e459a9d | 171 | \r |
d20df6fc | 172 | // Get the geometrical data of the chamber\r |
053767a4 | 173 | Int_t layer = geo->GetLayer(fDetector);\r |
174 | Int_t stack = geo->GetStack(fDetector);\r | |
d20df6fc | 175 | Int_t sector= geo->GetSector(fDetector);\r |
053767a4 | 176 | AliTRDpadPlane *pp = geo->GetPadPlane(layer, stack);\r |
d20df6fc | 177 | Int_t nCols = pp->GetNcols();\r |
178 | Float_t ymin = TMath::Min(pp->GetCol0(), pp->GetColEnd());\r | |
179 | Float_t ymax = TMath::Max(pp->GetCol0(), pp->GetColEnd());\r | |
180 | Float_t zmin = TMath::Min(pp->GetRow0(), pp->GetRowEnd());\r | |
181 | Float_t zmax = TMath::Max(pp->GetRow0(), pp->GetRowEnd());\r | |
182 | Float_t z0 = -1., zl = -1.;\r | |
183 | Int_t nRows = pp->GetNrows();\r | |
184 | Float_t binlength = (ymax - ymin)/nCols; \r | |
185 | //AliInfo(Form("ymin(%f) ymax(%f) zmin(%f) zmax(%f) nRows(%d) binlength(%f)", ymin, ymax, zmin, zmax, nRows, binlength));\r | |
186 | \r | |
187 | // Fill the histogram\r | |
188 | Int_t nClusters; \r | |
189 | Int_t *histogram[kMaxRows]; // 2D-Histogram\r | |
190 | Int_t hvals[kMaxPads]; memset(hvals, 0, sizeof(Int_t)*kMaxPads); \r | |
191 | Float_t *sigmas[kMaxRows];\r | |
192 | Float_t svals[kMaxPads]; memset(svals, 0, sizeof(Float_t)*kMaxPads); \r | |
193 | AliTRDcluster *c = 0x0;\r | |
194 | for(Int_t irs = 0; irs < kMaxRows; irs++){\r | |
195 | histogram[irs] = &hvals[irs*kMaxCols];\r | |
196 | sigmas[irs] = &svals[irs*kMaxCols];\r | |
197 | }\r | |
4e459a9d | 198 | for(Int_t iTime = timeBinMin; iTime < kNTimeBins-timeBinMax; iTime++){\r |
199 | if(!(nClusters = fTB[iTime].GetNClusters())) continue;\r | |
d20df6fc | 200 | z0 = fTB[iTime].GetZ0();\r |
201 | zl = fTB[iTime].GetDZ0();\r | |
202 | for(Int_t incl = 0; incl < nClusters; incl++){\r | |
203 | c = fTB[iTime].GetCluster(incl); \r | |
204 | histogram[c->GetPadRow()][c->GetPadCol()]++;\r | |
205 | sigmas[c->GetPadRow()][c->GetPadCol()] += c->GetSigmaZ2();\r | |
206 | }\r | |
207 | }\r | |
208 | \r | |
209 | // Now I have everything in the histogram, do the selection\r | |
210 | //Int_t nPads = nCols * nRows;\r | |
211 | // This is what we are interested in: The center of gravity of the best candidates\r | |
212 | Float_t cogyvals[kMaxPads]; memset(cogyvals, 0, sizeof(Float_t)*kMaxPads);\r | |
213 | Float_t cogzvals[kMaxPads]; memset(cogzvals, 0, sizeof(Float_t)*kMaxPads);\r | |
214 | Float_t *cogy[kMaxRows];\r | |
215 | Float_t *cogz[kMaxRows];\r | |
216 | \r | |
217 | // Lookup-Table storing coordinates according to the bins\r | |
218 | Float_t yLengths[kMaxCols];\r | |
219 | Float_t zLengths[kMaxRows];\r | |
220 | for(Int_t icnt = 0; icnt < nCols; icnt++){\r | |
221 | yLengths[icnt] = pp->GetColPos(nCols - 1 - icnt) + binlength/2;\r | |
222 | }\r | |
223 | for(Int_t icnt = 0; icnt < nRows; icnt++){\r | |
224 | zLengths[icnt] = pp->GetRowPos(icnt) - pp->GetRowSize(icnt)/2;\r | |
225 | }\r | |
226 | \r | |
227 | // A bitfield is used to mask the pads as usable\r | |
228 | Short_t mask[kMaxCols]; memset(mask, 0 ,sizeof(Short_t) * kMaxCols);//bool mvals[kMaxPads];\r | |
229 | for(UChar_t icount = 0; icount < nRows; icount++){\r | |
230 | cogy[icount] = &cogyvals[icount*kMaxCols];\r | |
231 | cogz[icount] = &cogzvals[icount*kMaxCols];\r | |
232 | }\r | |
233 | // In this array the array position of the best candidates will be stored\r | |
234 | Int_t cand[AliTRDtrackerV1::kMaxTracksStack];\r | |
235 | Float_t sigcands[AliTRDtrackerV1::kMaxTracksStack];\r | |
236 | \r | |
237 | // helper variables\r | |
238 | Int_t indices[kMaxPads]; memset(indices, -1, sizeof(Int_t)*kMaxPads);\r | |
239 | Int_t nCandidates = 0;\r | |
240 | Float_t norm, cogv;\r | |
241 | // histogram filled -> Select best bins\r | |
242 | Int_t nPads = nCols * nRows;\r | |
243 | TMath::Sort(nPads, hvals, indices); // bins storing a 0 should not matter\r | |
244 | // Set Threshold\r | |
245 | Int_t maximum = hvals[indices[0]]; // best\r | |
3a039a31 | 246 | Int_t threshold = Int_t(maximum * rec->GetRecoParam()->GetFindableClusters());\r |
d20df6fc | 247 | Int_t col, row, lower, lower1, upper, upper1;\r |
248 | for(Int_t ib = 0; ib < nPads; ib++){\r | |
249 | if(nCandidates >= AliTRDtrackerV1::kMaxTracksStack){\r | |
250 | printf("Number of seed candidates %d exceeded maximum allowed per stack %d", nCandidates, AliTRDtrackerV1::kMaxTracksStack);\r | |
251 | break;\r | |
252 | }\r | |
253 | // Positions\r | |
254 | row = indices[ib]/nCols;\r | |
255 | col = indices[ib]%nCols;\r | |
256 | // here will be the threshold condition:\r | |
257 | if((mask[col] & (1 << row)) != 0) continue; // Pad is masked: continue\r | |
258 | if(histogram[row][col] < TMath::Max(threshold, 1)){ // of course at least one cluster is needed\r | |
259 | break; // number of clusters below threshold: break;\r | |
260 | } \r | |
261 | // passing: Mark the neighbors\r | |
262 | lower = TMath::Max(col - 1, 0); upper = TMath::Min(col + 2, nCols);\r | |
263 | lower1 = TMath::Max(row - 1, 0); upper1 = TMath::Min(row + 2, nCols);\r | |
264 | for(Int_t ic = lower; ic < upper; ++ic)\r | |
265 | for(Int_t ir = lower1; ir < upper1; ++ir){\r | |
266 | if(ic == col && ir == row) continue;\r | |
267 | mask[ic] |= (1 << ir);\r | |
268 | }\r | |
269 | // Storing the position in an array\r | |
270 | // testing for neigboring\r | |
271 | cogv = 0;\r | |
272 | norm = 0;\r | |
273 | lower = TMath::Max(col - 1, 0);\r | |
274 | upper = TMath::Min(col + 2, nCols);\r | |
275 | for(Int_t inb = lower; inb < upper; ++inb){\r | |
276 | cogv += yLengths[inb] * histogram[row][inb];\r | |
277 | norm += histogram[row][inb];\r | |
278 | }\r | |
279 | cogy[row][col] = cogv / norm;\r | |
280 | cogv = 0; norm = 0;\r | |
281 | lower = TMath::Max(row - 1, 0);\r | |
282 | upper = TMath::Min(row + 2, nRows);\r | |
283 | for(Int_t inb = lower; inb < upper; ++inb){\r | |
284 | cogv += zLengths[inb] * histogram[inb][col];\r | |
285 | norm += histogram[inb][col];\r | |
286 | }\r | |
287 | cogz[row][col] = Float_t(cogv) / norm;\r | |
288 | // passed the filter\r | |
289 | cand[nCandidates] = row*nCols + col; // store the position of a passig candidate into an Array\r | |
290 | sigcands[nCandidates] = sigmas[row][col] / histogram[row][col]; // never be a floating point exeption\r | |
291 | // Analysis output\r | |
292 | nCandidates++;\r | |
293 | }\r | |
d611c74f | 294 | if(!nCandidates) return kFALSE;\r |
d20df6fc | 295 | \r |
296 | Float_t pos[3], sig[2];\r | |
297 | Short_t signal[7]; memset(&signal[0], 0, 7*sizeof(Short_t));\r | |
d611c74f | 298 | \r |
299 | new(fakeLayer) AliTRDchamberTimeBin(layer, stack, sector, z0, zl);\r | |
43d6ad34 | 300 | fakeLayer->SetReconstructor(rec);\r |
d20df6fc | 301 | AliTRDcluster *cluster = 0x0;\r |
302 | if(nCandidates){\r | |
303 | UInt_t fakeIndex = 0;\r | |
304 | for(Int_t ican = 0; ican < nCandidates; ican++){\r | |
305 | row = cand[ican] / nCols;\r | |
306 | col = cand[ican] % nCols;\r | |
307 | //temporary\r | |
308 | Int_t n = 0; Double_t x = 0., y = 0., z = 0.;\r | |
309 | for(int itb=0; itb<kNTimeBins; itb++){\r | |
310 | if(!(nClusters = fTB[itb].GetNClusters())) continue;\r | |
311 | for(Int_t incl = 0; incl < nClusters; incl++){\r | |
312 | c = fTB[itb].GetCluster(incl); \r | |
313 | if(c->GetPadRow() != row) continue;\r | |
314 | if(TMath::Abs(c->GetPadCol() - col) > 2) continue;\r | |
315 | x += c->GetX();\r | |
316 | y += c->GetY();\r | |
317 | z += c->GetZ();\r | |
318 | n++;\r | |
319 | }\r | |
320 | }\r | |
321 | pos[0] = x/n;\r | |
322 | pos[1] = y/n;\r | |
323 | pos[2] = z/n;\r | |
324 | sig[0] = .02;\r | |
325 | sig[1] = sigcands[ican];\r | |
326 | cluster = new AliTRDcluster(fDetector, 0., pos, sig, 0x0, 3, signal, col, row, 0, 0, 0., 0);\r | |
327 | fakeLayer->InsertCluster(cluster, fakeIndex++);\r | |
328 | }\r | |
329 | }\r | |
330 | fakeLayer->SetNRows(nRows);\r | |
331 | fakeLayer->SetOwner();\r | |
332 | fakeLayer->BuildIndices();\r | |
333 | //fakeLayer->PrintClusters();\r | |
334 | \r | |
3a039a31 | 335 | if(rec->GetStreamLevel(AliTRDReconstructor::kTracker) >= 3){\r |
d20df6fc | 336 | //TMatrixD hist(nRows, nCols);\r |
337 | //for(Int_t i = 0; i < nRows; i++)\r | |
338 | // for(Int_t j = 0; j < nCols; j++)\r | |
339 | // hist(i,j) = histogram[i][j];\r | |
340 | TTreeSRedirector &cstreamer = *AliTRDtrackerV1::DebugStreamer();\r | |
341 | cstreamer << "GetSeedingLayer"\r | |
053767a4 | 342 | << "layer=" << layer\r |
d20df6fc | 343 | << "ymin=" << ymin\r |
344 | << "ymax=" << ymax\r | |
345 | << "zmin=" << zmin\r | |
346 | << "zmax=" << zmax\r | |
347 | << "L.=" << fakeLayer\r | |
348 | //<< "Histogram.=" << &hist\r | |
349 | << "\n";\r | |
350 | }\r | |
351 | \r | |
d611c74f | 352 | return kTRUE;\r |
d20df6fc | 353 | }\r |
354 | \r |