src/daComposant/daAlgorithms/ExtendedKalmanFilter.py

   1 # -*- coding: utf-8 -*-
   2 #
   3 # Copyright (C) 2008-2018 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  = ["APosterioriCorrelations", "APosterioriCovariance", "APosterioriStandardDeviations", "APosterioriVariances", "BMA", "CurrentState", "CostFunctionJ", "CostFunctionJb", "CostFunctionJo", "Innovation"]
  57             )
  58         self.defineRequiredParameter( # Pas de type
  59             name     = "Bounds",
  60             message  = "Liste des valeurs de bornes",
  61             )
  62         self.requireInputArguments(
  63             mandatory= ("Xb", "Y", "HO", "R", "B" ),
  64             optional = ("U", "EM", "CM", "Q"),
  65             )
  66
  67     def run(self, Xb=None, Y=None, U=None, HO=None, EM=None, CM=None, R=None, B=None, Q=None, Parameters=None):
  68         self._pre_run(Parameters, Xb, Y, R, B, Q)
  69         #
  70         if self._parameters["EstimationOf"] == "Parameters":
  71             self._parameters["StoreInternalVariables"] = True
  72         #
  73         # Opérateurs
  74         # ----------
  75         if B is None:
  76             raise ValueError("Background error covariance matrix has to be properly defined!")
  77         if R is None:
  78             raise ValueError("Observation error covariance matrix has to be properly defined!")
  79         #
  80         H = HO["Direct"].appliedControledFormTo
  81         #
  82         if self._parameters["EstimationOf"] == "State":
  83             M = EM["Direct"].appliedControledFormTo
  84         #
  85         if CM is not None and "Tangent" in CM and U is not None:
  86             Cm = CM["Tangent"].asMatrix(Xb)
  87         else:
  88             Cm = None
  89         #
  90         # Nombre de pas identique au nombre de pas d'observations
  91         # -------------------------------------------------------
  92         if hasattr(Y,"stepnumber"):
  93             duration = Y.stepnumber()
  94         else:
  95             duration = 2
  96         #
  97         # Précalcul des inversions de B et R
  98         # ----------------------------------
  99         if self._parameters["StoreInternalVariables"]:
 100             BI = B.getI()
 101             RI = R.getI()
 102         #
 103         # Initialisation
 104         # --------------
 105         Xn = Xb
 106         Pn = B
 107         #
 108         self.StoredVariables["Analysis"].store( Xn.A1 )
 109         if "APosterioriCovariance" in self._parameters["StoreSupplementaryCalculations"]:
 110             self.StoredVariables["APosterioriCovariance"].store( Pn.asfullmatrix(Xn.size) )
 111             covarianceXa = Pn
 112         Xa               = Xn
 113         previousJMinimum = numpy.finfo(float).max
 114         #
 115         for step in range(duration-1):
 116             if hasattr(Y,"store"):
 117                 Ynpu = numpy.asmatrix(numpy.ravel( Y[step+1] )).T
 118             else:
 119                 Ynpu = numpy.asmatrix(numpy.ravel( Y )).T
 120             #
 121             Ht = HO["Tangent"].asMatrix(ValueForMethodForm = Xn)
 122             Ht = Ht.reshape(Ynpu.size,Xn.size) # ADAO & check shape
 123             Ha = HO["Adjoint"].asMatrix(ValueForMethodForm = Xn)
 124             Ha = Ha.reshape(Xn.size,Ynpu.size) # ADAO & check shape
 125             #
 126             if self._parameters["EstimationOf"] == "State":
 127                 Mt = EM["Tangent"].asMatrix(ValueForMethodForm = Xn)
 128                 Mt = Mt.reshape(Xn.size,Xn.size) # ADAO & check shape
 129                 Ma = EM["Adjoint"].asMatrix(ValueForMethodForm = Xn)
 130                 Ma = Ma.reshape(Xn.size,Xn.size) # ADAO & check shape
 131             #
 132             if U is not None:
 133                 if hasattr(U,"store") and len(U)>1:
 134                     Un = numpy.asmatrix(numpy.ravel( U[step] )).T
 135                 elif hasattr(U,"store") and len(U)==1:
 136                     Un = numpy.asmatrix(numpy.ravel( U[0] )).T
 137                 else:
 138                     Un = numpy.asmatrix(numpy.ravel( U )).T
 139             else:
 140                 Un = None
 141             #
 142             if self._parameters["EstimationOf"] == "State":
 143                 Xn_predicted = numpy.asmatrix(numpy.ravel( M( (Xn, Un) ) )).T
 144                 if Cm is not None and Un is not None: # Attention : si Cm est aussi dans M, doublon !
 145                     Cm = Cm.reshape(Xn.size,Un.size) # ADAO & check shape
 146                     Xn_predicted = Xn_predicted + Cm * Un
 147                 Pn_predicted = Q + Mt * Pn * Ma
 148             elif self._parameters["EstimationOf"] == "Parameters":
 149                 # --- > Par principe, M = Id, Q = 0
 150                 Xn_predicted = Xn
 151                 Pn_predicted = Pn
 152             #
 153             if self._parameters["Bounds"] is not None and self._parameters["ConstrainedBy"] == "EstimateProjection":
 154                 Xn_predicted = numpy.max(numpy.hstack((Xn_predicted,numpy.asmatrix(self._parameters["Bounds"])[:,0])),axis=1)
 155                 Xn_predicted = numpy.min(numpy.hstack((Xn_predicted,numpy.asmatrix(self._parameters["Bounds"])[:,1])),axis=1)
 156             #
 157             if self._parameters["EstimationOf"] == "State":
 158                 d  = Ynpu - numpy.asmatrix(numpy.ravel( H( (Xn_predicted, None) ) )).T
 159             elif self._parameters["EstimationOf"] == "Parameters":
 160                 d  = Ynpu - numpy.asmatrix(numpy.ravel( H( (Xn_predicted, Un) ) )).T
 161                 if Cm is not None and Un is not None: # Attention : si Cm est aussi dans H, doublon !
 162                     d = d - Cm * Un
 163             #
 164             _A = R + Ht * Pn_predicted * Ha
 165             _u = numpy.linalg.solve( _A , d )
 166             Xn = Xn_predicted + Pn_predicted * Ha * _u
 167             Kn = Pn_predicted * Ha * (R + Ht * Pn_predicted * Ha).I
 168             Pn = Pn_predicted - Kn * Ht * Pn_predicted
 169             #
 170             self.StoredVariables["Analysis"].store( Xn.A1 )
 171             if "APosterioriCovariance" in self._parameters["StoreSupplementaryCalculations"]:
 172                 self.StoredVariables["APosterioriCovariance"].store( Pn )
 173             if "Innovation" in self._parameters["StoreSupplementaryCalculations"]:
 174                 self.StoredVariables["Innovation"].store( numpy.ravel( d.A1 ) )
 175             if self._parameters["StoreInternalVariables"] or "CurrentState" in self._parameters["StoreSupplementaryCalculations"]:
 176                 self.StoredVariables["CurrentState"].store( Xn )
 177             if self._parameters["StoreInternalVariables"] \
 178                 or "CostFunctionJ" in self._parameters["StoreSupplementaryCalculations"] \
 179                 or "CostFunctionJb" in self._parameters["StoreSupplementaryCalculations"] \
 180                 or "CostFunctionJo" in self._parameters["StoreSupplementaryCalculations"]:
 181                 Jb  = 0.5 * (Xn - Xb).T * BI * (Xn - Xb)
 182                 Jo  = 0.5 * d.T * RI * d
 183                 J   = float( Jb ) + float( Jo )
 184                 self.StoredVariables["CostFunctionJb"].store( Jb )
 185                 self.StoredVariables["CostFunctionJo"].store( Jo )
 186                 self.StoredVariables["CostFunctionJ" ].store( J )
 187                 if J < previousJMinimum:
 188                     previousJMinimum  = J
 189                     Xa                = Xn
 190                     if "APosterioriCovariance" in self._parameters["StoreSupplementaryCalculations"]:
 191                         covarianceXa  = Pn
 192             else:
 193                 Xa = Xn
 194             #
 195         #
 196         # Stockage supplementaire de l'optimum en estimation de parametres
 197         # ----------------------------------------------------------------
 198         if self._parameters["EstimationOf"] == "Parameters":
 199             self.StoredVariables["Analysis"].store( Xa.A1 )
 200             if "APosterioriCovariance" in self._parameters["StoreSupplementaryCalculations"]:
 201                 self.StoredVariables["APosterioriCovariance"].store( covarianceXa )
 202         #
 203         if "BMA" in self._parameters["StoreSupplementaryCalculations"]:
 204             self.StoredVariables["BMA"].store( numpy.ravel(Xb) - numpy.ravel(Xa) )
 205         #
 206         self._post_run(HO)
 207         return 0
 208
 209 # ==============================================================================
 210 if __name__ == "__main__":
 211     print('\n AUTODIAGNOSTIC \n')