ShortPeriodicsInterpolatedCoefficient.java

  1. /* Copyright 2002-2018 CS Systèmes d'Information
  2.  * Licensed to CS Systèmes d'Information (CS) under one or more
  3.  * contributor license agreements.  See the NOTICE file distributed with
  4.  * this work for additional information regarding copyright ownership.
  5.  * CS licenses this file to You under the Apache License, Version 2.0
  6.  * (the "License"); you may not use this file except in compliance with
  7.  * the License.  You may obtain a copy of the License at
  8.  *
  9.  *   http://www.apache.org/licenses/LICENSE-2.0
  10.  *
  11.  * Unless required by applicable law or agreed to in writing, software
  12.  * distributed under the License is distributed on an "AS IS" BASIS,
  13.  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  14.  * See the License for the specific language governing permissions and
  15.  * limitations under the License.
  16.  */
  17. package org.orekit.propagation.semianalytical.dsst.utilities;

  19. import java.io.Serializable;
  20. import java.util.ArrayList;

  22. import org.hipparchus.analysis.interpolation.HermiteInterpolator;
  23. import org.hipparchus.util.FastMath;
  24. import org.orekit.time.AbsoluteDate;

  26. /** Interpolated short periodics coefficients.
  27.  * <p>
  28.  * Representation of a coefficient that need to be interpolated over time.
  29.  * </p><p>
  30.  * The short periodics coefficients can be interpolated for faster computation.
  31.  * This class stores computed values of the coefficients through the method
  32.  * {@link #addGridPoint} and gives an interpolated result through the method
  33.  * {@link #value}.
  34.  * </p>
  35.  * @author Nicolas Bernard
  36.  *
  37.  */
  38. public class ShortPeriodicsInterpolatedCoefficient implements Serializable {

  40.     /** Serializable UID. */
  41.     private static final long serialVersionUID = 20160319L;

  43.     /**Values of the already computed coefficients.*/
  44.     private ArrayList<double[]> values;

  46.     /**Grid points.*/
  47.     private ArrayList<AbsoluteDate> abscissae;

  49.     /**Number of points used in the interpolation.*/
  50.     private int interpolationPoints;

  52.     /**Index of the latest closest neighbor.*/
  53.     private int latestClosestNeighbor;

  55.     /**Simple constructor.
  56.      * @param interpolationPoints number of points used in the interpolation
  57.      */
  58.     public ShortPeriodicsInterpolatedCoefficient(final int interpolationPoints) {
  59.         this.interpolationPoints = interpolationPoints;
  60.         this.abscissae = new ArrayList<AbsoluteDate>();
  61.         this.values = new ArrayList<double[]>();
  62.         this.latestClosestNeighbor = 0;
  63.     }

  65.     /**Compute the value of the coefficient.
  66.      * @param date date at which the coefficient should be computed
  67.      * @return value of the coefficient
  68.      */
  69.     public double[] value(final AbsoluteDate date) {
  70.         //Get the closest points from the input date
  71.         final int[] neighbors = getNeighborsIndices(date);

  73.         //Creation and set up of the interpolator
  74.         final HermiteInterpolator interpolator = new HermiteInterpolator();
  75.         for (int i : neighbors) {
  76.             interpolator.addSamplePoint(abscissae.get(i).durationFrom(date), values.get(i));
  77.         }

  79.         //interpolation
  80.         return interpolator.value(0.0);

  82.     }

  84.     /**Find the closest available points from the specified date.
  85.      * @param date date of interest
  86.      * @return indices corresponding to the closest points on the time scale
  87.      */
  88.     private int[] getNeighborsIndices(final AbsoluteDate date) {
  89.         final int sizeofNeighborhood = FastMath.min(interpolationPoints, abscissae.size());
  90.         final int[] neighborsIndices = new int[sizeofNeighborhood];

  92.         //If the size of the complete sample is less than
  93.         //the desired number of interpolation points,
  94.         //then the entire sample is considered as the neighborhood
  95.         if (interpolationPoints >= abscissae.size()) {
  96.             for (int i = 0; i < sizeofNeighborhood; i++) {
  97.                 neighborsIndices[i] = i;
  98.             }
  99.         } else {
  100.             // get indices around closest neighbor
  101.             int inf = getClosestNeighbor(date);
  102.             int sup = inf + 1;

  104.             while (sup - inf < interpolationPoints) {
  105.                 if (inf == 0) { //This means that we have reached the earliest date
  106.                     sup++;
  107.                 } else if (sup >= abscissae.size()) { //This means that we have reached the latest date
  108.                     inf--;
  109.                 } else { //the choice is made between the two next neighbors
  110.                     final double lowerNeighborDistance = FastMath.abs(abscissae.get(inf - 1).durationFrom(date));
  111.                     final double upperNeighborDistance = FastMath.abs(abscissae.get(sup).durationFrom(date));

  113.                     if (lowerNeighborDistance <= upperNeighborDistance) {
  114.                         inf--;
  115.                     } else {
  116.                         sup++;
  117.                     }
  118.                 }
  119.             }

  121.             for (int i = 0; i < interpolationPoints; ++i) {
  122.                 neighborsIndices[i] = inf + i;
  123.             }

  125.         }

  127.         return neighborsIndices;
  128.     }

  130.     /**Find the closest point from a specific date amongst the available points.
  131.      * @param date date of interest
  132.      * @return index of the closest abscissa from the date of interest
  133.      */
  134.     private int getClosestNeighbor(final AbsoluteDate date) {
  135.         //the starting point is the latest result of a call to this method.
  136.         //Indeed, as this class is meant to be called during an integration process
  137.         //with an input date evolving often continuously in time, there is a high
  138.         //probability that the result will be the same as for last call of
  139.         //this method.
  140.         int closestNeighbor = latestClosestNeighbor;

  142.         //case where the date is before the available points
  143.         if (date.compareTo(abscissae.get(0)) <= 0) {
  144.             closestNeighbor = 0;
  145.         }
  146.         //case where the date is after the available points
  147.         else if (date.compareTo(abscissae.get(abscissae.size() - 1)) >= 0) {
  148.             closestNeighbor = abscissae.size() - 1;
  149.         }
  150.         //general case: one is looking for the two consecutives entries that surround the input date
  151.         //then one choose the closest one
  152.         else {
  153.             int lowerBorder = latestClosestNeighbor;
  154.             int upperBorder = latestClosestNeighbor;

  156.             final int searchDirection = date.compareTo(abscissae.get(latestClosestNeighbor));
  157.             if (searchDirection > 0) {
  158.                 upperBorder++;
  159.                 while (date.compareTo(abscissae.get(upperBorder)) > 0) {
  160.                     upperBorder++;
  161.                     lowerBorder++;
  162.                 }
  163.             }
  164.             else {
  165.                 lowerBorder--;
  166.                 while (date.compareTo(abscissae.get(lowerBorder)) < 0) {
  167.                     upperBorder--;
  168.                     lowerBorder--;
  169.                 }
  170.             }

  172.             final double lowerDistance = FastMath.abs(date.durationFrom(abscissae.get(lowerBorder)));
  173.             final double upperDistance = FastMath.abs(date.durationFrom(abscissae.get(upperBorder)));

  175.             closestNeighbor = (lowerDistance < upperDistance) ? lowerBorder : upperBorder;
  176.         }

  178.         //The result is stored in order to speed up the next call to the function
  179.         //Indeed, it is highly likely that the requested result will be the same
  180.         this.latestClosestNeighbor = closestNeighbor;
  181.         return closestNeighbor;
  182.     }

  184.     /** Clear the recorded values from the interpolation grid.
  185.      */
  186.     public void clearHistory() {
  187.         abscissae.clear();
  188.         values.clear();
  189.     }

  191.     /** Add a point to the interpolation grid.
  192.      * @param date abscissa of the point
  193.      * @param value value of the element
  194.      */
  195.     public void addGridPoint(final AbsoluteDate date, final double[] value) {
  196.         //If the grid is empty, the value is directly added to both arrays
  197.         if (abscissae.isEmpty()) {
  198.             abscissae.add(date);
  199.             values.add(value);
  200.         }
  201.         //If the grid already contains this point, only its value is changed
  202.         else if (abscissae.contains(date)) {
  203.             values.set(abscissae.indexOf(date), value);
  204.         }
  205.         //If the grid does not contain this point, the position of the point
  206.         //in the grid is computed first
  207.         else {
  208.             final int closestNeighbor = getClosestNeighbor(date);
  209.             final int index = (date.compareTo(abscissae.get(closestNeighbor)) < 0) ? closestNeighbor : closestNeighbor + 1;
  210.             abscissae.add(index, date);
  211.             values.add(index, value);
  212.         }
  213.     }
  214. }
Created with JaCoCo 0.7.9.201702052155