2 Copyright (C) 2008-2017 EDF R&D
4 This file is part of SALOME ADAO module.
6 This library is free software; you can redistribute it and/or
7 modify it under the terms of the GNU Lesser General Public
8 License as published by the Free Software Foundation; either
9 version 2.1 of the License, or (at your option) any later version.
11 This library 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 GNU
14 Lesser General Public License for more details.
16 You should have received a copy of the GNU Lesser General Public
17 License along with this library; if not, write to the Free Software
18 Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20 See http://www.salome-platform.org/ or email : webmaster.salome@opencascade.com
22 Author: Jean-Philippe Argaud, jean-philippe.argaud@edf.fr, EDF R&D
24 .. index:: single: KalmanFilter
25 .. _section_ref_algorithm_KalmanFilter:
27 Calculation algorithm "*KalmanFilter*"
28 --------------------------------------
33 This algorithm realizes an estimation of the state of a dynamic system by a
36 It is theoretically reserved for observation and incremental evolution operators
37 cases which are linear, even if it sometimes works in "slightly" non-linear
38 cases. One can verify the linearity of the observation operator with the help of
39 the :ref:`section_ref_algorithm_LinearityTest`.
41 In case of non-linearity, even slightly marked, it will be preferred the
42 :ref:`section_ref_algorithm_ExtendedKalmanFilter` or the
43 :ref:`section_ref_algorithm_UnscentedKalmanFilter`.
45 Optional and required commands
46 ++++++++++++++++++++++++++++++
48 .. index:: single: AlgorithmParameters
49 .. index:: single: Background
50 .. index:: single: BackgroundError
51 .. index:: single: Observation
52 .. index:: single: ObservationError
53 .. index:: single: ObservationOperator
54 .. index:: single: EstimationOf
55 .. index:: single: StoreSupplementaryCalculations
57 The general required commands, available in the editing user interface, are the
61 *Required command*. This indicates the background or initial vector used,
62 previously noted as :math:`\mathbf{x}^b`. Its value is defined as a
63 "*Vector*" or a *VectorSerie*" type object.
66 *Required command*. This indicates the background error covariance matrix,
67 previously noted as :math:`\mathbf{B}`. Its value is defined as a "*Matrix*"
68 type object, a "*ScalarSparseMatrix*" type object, or a
69 "*DiagonalSparseMatrix*" type object.
72 *Required command*. This indicates the observation vector used for data
73 assimilation or optimization, previously noted as :math:`\mathbf{y}^o`. It
74 is defined as a "*Vector*" or a *VectorSerie* type object.
77 *Required command*. This indicates the observation error covariance matrix,
78 previously noted as :math:`\mathbf{R}`. It is defined as a "*Matrix*" type
79 object, a "*ScalarSparseMatrix*" type object, or a "*DiagonalSparseMatrix*"
83 *Required command*. This indicates the observation operator, previously
84 noted :math:`H`, which transforms the input parameters :math:`\mathbf{x}` to
85 results :math:`\mathbf{y}` to be compared to observations
86 :math:`\mathbf{y}^o`. Its value is defined as a "*Function*" type object or
87 a "*Matrix*" type one. In the case of "*Function*" type, different
88 functional forms can be used, as described in the section
89 :ref:`section_ref_operator_requirements`. If there is some control :math:`U`
90 included in the observation, the operator has to be applied to a pair
93 The general optional commands, available in the editing user interface, are
94 indicated in :ref:`section_ref_assimilation_keywords`. Moreover, the parameters
95 of the command "*AlgorithmParameters*" allows to choose the specific options,
96 described hereafter, of the algorithm. See
97 :ref:`section_ref_options_Algorithm_Parameters` for the good use of this
100 The options of the algorithm are the following:
103 This key allows to choose the type of estimation to be performed. It can be
104 either state-estimation, with a value of "State", or parameter-estimation,
105 with a value of "Parameters". The default choice is "State".
107 Example : ``{"EstimationOf":"Parameters"}``
109 StoreSupplementaryCalculations
110 This list indicates the names of the supplementary variables that can be
111 available at the end of the algorithm. It involves potentially costly
112 calculations or memory consumptions. The default is a void list, none of
113 these variables being calculated and stored by default. The possible names
114 are in the following list: ["APosterioriCorrelations",
115 "APosterioriCovariance", "APosterioriStandardDeviations",
116 "APosterioriVariances", "BMA", "CostFunctionJ", "CostFunctionJb",
117 "CostFunctionJo", "CurrentState", "Innovation"].
119 Example : ``{"StoreSupplementaryCalculations":["BMA", "Innovation"]}``
121 Information and variables available at the end of the algorithm
122 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
124 At the output, after executing the algorithm, there are variables and
125 information originating from the calculation. The description of
126 :ref:`section_ref_output_variables` show the way to obtain them by the method
127 named ``get`` of the variable "*ADD*" of the post-processing. The input
128 variables, available to the user at the output in order to facilitate the
129 writing of post-processing procedures, are described in the
130 :ref:`subsection_r_o_v_Inventaire`.
132 The unconditional outputs of the algorithm are the following:
135 *List of vectors*. Each element is an optimal state :math:`\mathbf{x}*` in
136 optimization or an analysis :math:`\mathbf{x}^a` in data assimilation.
138 Example : ``Xa = ADD.get("Analysis")[-1]``
140 The conditional outputs of the algorithm are the following:
142 APosterioriCorrelations
143 *List of matrices*. Each element is an *a posteriori* error correlation
144 matrix of the optimal state.
146 Example : ``C = ADD.get("APosterioriCorrelations")[-1]``
148 APosterioriCovariance
149 *List of matrices*. Each element is an *a posteriori* error covariance
150 matrix :math:`\mathbf{A}*` of the optimal state.
152 Example : ``A = ADD.get("APosterioriCovariance")[-1]``
154 APosterioriStandardDeviations
155 *List of matrices*. Each element is an *a posteriori* error standard
156 deviation matrix of the optimal state.
158 Example : ``E = ADD.get("APosterioriStandardDeviations")[-1]``
161 *List of matrices*. Each element is an *a posteriori* error variance matrix
162 of the optimal state.
164 Example : ``V = ADD.get("APosterioriVariances")[-1]``
167 *List of vectors*. Each element is a vector of difference between the
168 background and the optimal state.
170 Example : ``bma = ADD.get("BMA")[-1]``
173 *List of values*. Each element is a value of the error function :math:`J`.
175 Example : ``J = ADD.get("CostFunctionJ")[:]``
178 *List of values*. Each element is a value of the error function :math:`J^b`,
179 that is of the background difference part.
181 Example : ``Jb = ADD.get("CostFunctionJb")[:]``
184 *List of values*. Each element is a value of the error function :math:`J^o`,
185 that is of the observation difference part.
187 Example : ``Jo = ADD.get("CostFunctionJo")[:]``
190 *List of vectors*. Each element is a usual state vector used during the
191 optimization algorithm procedure.
193 Example : ``Xs = ADD.get("CurrentState")[:]``
196 *List of vectors*. Each element is an innovation vector, which is in static
197 the difference between the optimal and the background, and in dynamic the
200 Exemple : ``d = ADD.get("Innovation")[-1]``
205 References to other sections:
206 - :ref:`section_ref_algorithm_ExtendedKalmanFilter`
207 - :ref:`section_ref_algorithm_UnscentedKalmanFilter`