src/daComposant/daAlgorithms/ExtendedKalmanFilter.py

   1 # -*- coding: utf-8 -*-
   2 #
   3 # Copyright (C) 2008-2019 EDF R&D
   4 #
   5 # This library is free software; you can redistribute it and/or
   6 # modify it under the terms of the GNU Lesser General Public
   7 # License as published by the Free Software Foundation; either
   8 # version 2.1 of the License.
   9 #
  10 # This library is distributed in the hope that it will be useful,
  11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13 # Lesser General Public License for more details.
  14 #
  15 # You should have received a copy of the GNU Lesser General Public
  16 # License along with this library; if not, write to the Free Software
  17 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
  18 #
  19 # See http://www.salome-platform.org/ or email : webmaster.salome@opencascade.com
  20 #
  21 # Author: Jean-Philippe Argaud, jean-philippe.argaud@edf.fr, EDF R&D
  22
  23 import logging
  24 from daCore import BasicObjects
  25 import numpy
  26
  27 # ==============================================================================
  28 class ElementaryAlgorithm(BasicObjects.Algorithm):
  29     def __init__(self):
  30         BasicObjects.Algorithm.__init__(self, "EXTENDEDKALMANFILTER")
  31         self.defineRequiredParameter(
  32             name     = "ConstrainedBy",
  33             default  = "EstimateProjection",
  34             typecast = str,
  35             message  = "Prise en compte des contraintes",
  36             listval  = ["EstimateProjection"],
  37             )
  38         self.defineRequiredParameter(
  39             name     = "EstimationOf",
  40             default  = "State",
  41             typecast = str,
  42             message  = "Estimation d'etat ou de parametres",
  43             listval  = ["State", "Parameters"],
  44             )
  45         self.defineRequiredParameter(
  46             name     = "StoreInternalVariables",
  47             default  = False,
  48             typecast = bool,
  49             message  = "Stockage des variables internes ou intermédiaires du calcul",
  50             )
  51         self.defineRequiredParameter(
  52             name     = "StoreSupplementaryCalculations",
  53             default  = [],
  54             typecast = tuple,
  55             message  = "Liste de calculs supplémentaires à stocker et/ou effectuer",
  56             listval  = [
  57                 "APosterioriCorrelations",
  58                 "APosterioriCovariance",
  59                 "APosterioriStandardDeviations",
  60                 "APosterioriVariances",
  61                 "BMA",
  62                 "CurrentState",
  63                 "CostFunctionJ",
  64                 "CostFunctionJb",
  65                 "CostFunctionJo",
  66                 "Innovation",
  67                 "PredictedState",
  68                 ]
  69             )
  70         self.defineRequiredParameter( # Pas de type
  71             name     = "Bounds",
  72             message  = "Liste des valeurs de bornes",
  73             )
  74         self.requireInputArguments(
  75             mandatory= ("Xb", "Y", "HO", "R", "B" ),
  76             optional = ("U", "EM", "CM", "Q"),
  77             )
  78
  79     def run(self, Xb=None, Y=None, U=None, HO=None, EM=None, CM=None, R=None, B=None, Q=None, Parameters=None):
  80         self._pre_run(Parameters, Xb, Y, R, B, Q)
  81         #
  82         if self._parameters["EstimationOf"] == "Parameters":
  83             self._parameters["StoreInternalVariables"] = True
  84         #
  85         # Opérateurs
  86         # ----------
  87         Hm = HO["Direct"].appliedControledFormTo
  88         #
  89         if self._parameters["EstimationOf"] == "State":
  90             Mm = EM["Direct"].appliedControledFormTo
  91         #
  92         if CM is not None and "Tangent" in CM and U is not None:
  93             Cm = CM["Tangent"].asMatrix(Xb)
  94         else:
  95             Cm = None
  96         #
  97         # Nombre de pas identique au nombre de pas d'observations
  98         # -------------------------------------------------------
  99         if hasattr(Y,"stepnumber"):
 100             duration = Y.stepnumber()
 101         else:
 102             duration = 2
 103         #
 104         # Précalcul des inversions de B et R
 105         # ----------------------------------
 106         if self._parameters["StoreInternalVariables"] \
 107             or self._toStore("CostFunctionJ") \
 108             or self._toStore("CostFunctionJb") \
 109             or self._toStore("CostFunctionJo"):
 110             BI = B.getI()
 111             RI = R.getI()
 112         #
 113         # Initialisation
 114         # --------------
 115         Xn = Xb
 116         Pn = B
 117         #
 118         self.StoredVariables["Analysis"].store( Xn.A1 )
 119         if self._toStore("APosterioriCovariance"):
 120             self.StoredVariables["APosterioriCovariance"].store( Pn.asfullmatrix(Xn.size) )
 121             covarianceXa = Pn
 122         Xa               = Xn
 123         previousJMinimum = numpy.finfo(float).max
 124         #
 125         for step in range(duration-1):
 126             if hasattr(Y,"store"):
 127                 Ynpu = numpy.asmatrix(numpy.ravel( Y[step+1] )).T
 128             else:
 129                 Ynpu = numpy.asmatrix(numpy.ravel( Y )).T
 130             #
 131             Ht = HO["Tangent"].asMatrix(ValueForMethodForm = Xn)
 132             Ht = Ht.reshape(Ynpu.size,Xn.size) # ADAO & check shape
 133             Ha = HO["Adjoint"].asMatrix(ValueForMethodForm = Xn)
 134             Ha = Ha.reshape(Xn.size,Ynpu.size) # ADAO & check shape
 135             #
 136             if self._parameters["EstimationOf"] == "State":
 137                 Mt = EM["Tangent"].asMatrix(ValueForMethodForm = Xn)
 138                 Mt = Mt.reshape(Xn.size,Xn.size) # ADAO & check shape
 139                 Ma = EM["Adjoint"].asMatrix(ValueForMethodForm = Xn)
 140                 Ma = Ma.reshape(Xn.size,Xn.size) # ADAO & check shape
 141             #
 142             if U is not None:
 143                 if hasattr(U,"store") and len(U)>1:
 144                     Un = numpy.asmatrix(numpy.ravel( U[step] )).T
 145                 elif hasattr(U,"store") and len(U)==1:
 146                     Un = numpy.asmatrix(numpy.ravel( U[0] )).T
 147                 else:
 148                     Un = numpy.asmatrix(numpy.ravel( U )).T
 149             else:
 150                 Un = None
 151             #
 152             if self._parameters["EstimationOf"] == "State":
 153                 Xn_predicted = numpy.asmatrix(numpy.ravel( Mm( (Xn, Un) ) )).T
 154                 if Cm is not None and Un is not None: # Attention : si Cm est aussi dans M, doublon !
 155                     Cm = Cm.reshape(Xn.size,Un.size) # ADAO & check shape
 156                     Xn_predicted = Xn_predicted + Cm * Un
 157                 Pn_predicted = Q + Mt * Pn * Ma
 158             elif self._parameters["EstimationOf"] == "Parameters":
 159                 # --- > Par principe, M = Id, Q = 0
 160                 Xn_predicted = Xn
 161                 Pn_predicted = Pn
 162             #
 163             if self._parameters["Bounds"] is not None and self._parameters["ConstrainedBy"] == "EstimateProjection":
 164                 Xn_predicted = numpy.max(numpy.hstack((Xn_predicted,numpy.asmatrix(self._parameters["Bounds"])[:,0])),axis=1)
 165                 Xn_predicted = numpy.min(numpy.hstack((Xn_predicted,numpy.asmatrix(self._parameters["Bounds"])[:,1])),axis=1)
 166             #
 167             if self._parameters["EstimationOf"] == "State":
 168                 d  = Ynpu - numpy.asmatrix(numpy.ravel( Hm( (Xn_predicted, None) ) )).T
 169             elif self._parameters["EstimationOf"] == "Parameters":
 170                 d  = Ynpu - numpy.asmatrix(numpy.ravel( Hm( (Xn_predicted, Un) ) )).T
 171                 if Cm is not None and Un is not None: # Attention : si Cm est aussi dans H, doublon !
 172                     d = d - Cm * Un
 173             #
 174             _A = R + numpy.dot(Ht, Pn_predicted * Ha)
 175             _u = numpy.linalg.solve( _A , d )
 176             Xn = Xn_predicted + Pn_predicted * Ha * _u
 177             Kn = Pn_predicted * Ha * (R + numpy.dot(Ht, Pn_predicted * Ha)).I
 178             Pn = Pn_predicted - Kn * Ht * Pn_predicted
 179             #
 180             self.StoredVariables["Analysis"].store( Xn.A1 )
 181             if self._toStore("APosterioriCovariance"):
 182                 self.StoredVariables["APosterioriCovariance"].store( Pn )
 183             if self._toStore("Innovation"):
 184                 self.StoredVariables["Innovation"].store( numpy.ravel( d.A1 ) )
 185             if self._parameters["StoreInternalVariables"] \
 186                 or self._toStore("CurrentState"):
 187                 self.StoredVariables["CurrentState"].store( Xn )
 188             if self._parameters["StoreInternalVariables"] \
 189                 or self._toStore("PredictedState"):
 190                 self.StoredVariables["PredictedState"].store( Xn_predicted )
 191             if self._parameters["StoreInternalVariables"] \
 192                 or self._toStore("CostFunctionJ") \
 193                 or self._toStore("CostFunctionJb") \
 194                 or self._toStore("CostFunctionJo"):
 195                 Jb  = 0.5 * (Xn - Xb).T * BI * (Xn - Xb)
 196                 Jo  = 0.5 * d.T * RI * d
 197                 J   = float( Jb ) + float( Jo )
 198                 self.StoredVariables["CostFunctionJb"].store( Jb )
 199                 self.StoredVariables["CostFunctionJo"].store( Jo )
 200                 self.StoredVariables["CostFunctionJ" ].store( J )
 201                 if J < previousJMinimum:
 202                     previousJMinimum  = J
 203                     Xa                = Xn
 204                     if self._toStore("APosterioriCovariance"):
 205                         covarianceXa  = Pn
 206             else:
 207                 Xa = Xn
 208             #
 209         #
 210         # Stockage supplementaire de l'optimum en estimation de parametres
 211         # ----------------------------------------------------------------
 212         if self._parameters["EstimationOf"] == "Parameters":
 213             self.StoredVariables["Analysis"].store( Xa.A1 )
 214             if self._toStore("APosterioriCovariance"):
 215                 self.StoredVariables["APosterioriCovariance"].store( covarianceXa )
 216         #
 217         if self._toStore("BMA"):
 218             self.StoredVariables["BMA"].store( numpy.ravel(Xb) - numpy.ravel(Xa) )
 219         #
 220         self._post_run(HO)
 221         return 0
 222
 223 # ==============================================================================
 224 if __name__ == "__main__":
 225     print('\n AUTODIAGNOSTIC \n')