ParticleFilterFrameworkFloat.cpp

Go to the documentation of this file.

// ****************************************************************************
// This file is part of the Integrating Vision Toolkit (IVT).
//
// The IVT is maintained by the Karlsruhe Institute of Technology (KIT)
// (www.kit.edu) in cooperation with the company Keyetech (www.keyetech.de).
//
// Copyright (C) 2014 Karlsruhe Institute of Technology (KIT).
// All rights reserved.
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are met:
//
// 1. Redistributions of source code must retain the above copyright
//    notice, this list of conditions and the following disclaimer.
//
// 2. Redistributions in binary form must reproduce the above copyright
//    notice, this list of conditions and the following disclaimer in the
//    documentation and/or other materials provided with the distribution.
//
// 3. Neither the name of the KIT nor the names of its contributors may be
//    used to endorse or promote products derived from this software
//    without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE KIT AND CONTRIBUTORS “AS IS” AND ANY
// EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
// WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
// DISCLAIMED. IN NO EVENT SHALL THE KIT OR CONTRIBUTORS BE LIABLE FOR ANY
// DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
// (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
// LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
// ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
// THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
// ****************************************************************************
// ****************************************************************************
// Filename:  ParticleFilterFrameworkFloat.cpp
// Author:    Pedram Azad
// Date:      07.05.2009
// ****************************************************************************


// ****************************************************************************
// Includes
// ****************************************************************************

#include <new> // for explicitly using correct new/delete operators on VC DSPs

#include "ParticleFilterFrameworkFloat.h"

#include "Helpers/helpers.h"

#include <math.h>
#include <stdio.h>
#include <float.h>



// ****************************************************************************
// Constructor / Destructor
// ****************************************************************************

CParticleFilterFrameworkFloat::CParticleFilterFrameworkFloat(int nParticles, int nDimension)
{
        m_nParticles = nParticles;
        m_nDimension = nDimension;

        mean_configuration = new float[nDimension];
        last_configuration = new float[nDimension];

        sigma = new float[nDimension];
        lower_limit = new float[nDimension];
        upper_limit = new float[nDimension];

        int i;

        // allocate memory
        s = new float*[nParticles];
        for (i = 0; i < nParticles; i++)
                s[i] = new float[nDimension];

        s_temp = new float*[nParticles];
        for (i = 0; i < nParticles; i++)
                s_temp[i] = new float[nDimension];

        c = new double[nParticles];
        pi = new double[nParticles];

        temp = new float[nDimension];
        
        for (i = 0; i < nParticles; i++)
        {
                c[i] = 0.0;
                pi[i] = 0.0;
        }
}

CParticleFilterFrameworkFloat::~CParticleFilterFrameworkFloat()
{
        delete [] mean_configuration;
        delete [] last_configuration;
        delete [] sigma;
        delete [] lower_limit;
        delete [] upper_limit;

        int i;

        for (i = 0; i < m_nParticles; i++)
                delete [] s[i];
        delete [] s;

        for (i = 0; i < m_nParticles; i++)
                delete [] s_temp[i];
        delete [] s_temp;

        delete [] c;
        delete [] pi;

        delete [] temp;
}


// ****************************************************************************
// Methods
// ****************************************************************************

void CParticleFilterFrameworkFloat::GetConfiguration(float *pBestConfiguration, float fMeanFactor)
{
        double mean = 0;
        int nConfigurations = 0, i;
        
        for (i = 0; i < m_nDimension; i++)
                pBestConfiguration[i] = 0;

        for (i = 0; i < m_nParticles; i++)
                mean += pi[i];
                
        mean /= m_nParticles * double(fMeanFactor);
        
        for (i = 0; i < m_nParticles; i++)
                if (pi[i] > mean)
                {
                        for (int j = 0; j < m_nDimension; j++)
                                pBestConfiguration[j] += s[i][j];
                        
                        nConfigurations++;
                }

        if (nConfigurations > 0)
        {
                for (i = 0; i < m_nDimension; i++)
                        pBestConfiguration[i] /= nConfigurations;
        }
}

void CParticleFilterFrameworkFloat::GetBestConfiguration(float *pBestConfiguration)
{
        double max = -DBL_MAX;
        int best_i, i;

        for (i = 0; i < m_nParticles; i++)
                if (pi[i] > max)
                {
                        best_i = i;
                        max = pi[i];
                }

        
        for (i = 0; i < m_nDimension; i++)
                pBestConfiguration[i] = s[best_i][i];
}

void CParticleFilterFrameworkFloat::GetMeanConfiguration(float *pMeanConfiguration)
{
        for (int i = 0; i < m_nDimension; i++)
                pMeanConfiguration[i] = mean_configuration[i];
}

void CParticleFilterFrameworkFloat::GetPredictedConfiguration(float *pPredictedConfiguration)
{
        for (int i = 0; i < m_nDimension; i++)
                pPredictedConfiguration[i] = mean_configuration[i] + 0.8f * (mean_configuration[i] - last_configuration[i]);
}

int CParticleFilterFrameworkFloat::PickBaseSample()
{
        double choice = uniform_random() * c_total;

        int low = 0;
        int high = m_nParticles;

        while (high > (low + 1))
        {
                int middle = (high + low) >> 1;
                
                if (choice > c[middle])
                        low = middle;
                else
                        high = middle;
        }

        return low;
}

double CParticleFilterFrameworkFloat::ParticleFilter(float *pResultMeanConfiguration, float dSigmaFactor)
{
        // push previous state through process model
        // use dynamic model and add noise
        PredictNewBases(dSigmaFactor);

        // apply bayesian measurement weighting
        c_total = 0.0;
        int i;
        for (i = 0; i < m_nParticles; i++)
        {
                // update model (calculate forward kinematics)
                UpdateModel(i);

                // evaluate likelihood function (compare edges,...)
                pi[i] = CalculateProbability(false);
        }
        
        CalculateFinalProbabilities();
        
        for (i = 0; i < m_nParticles; i++)
        {
                c[i] = c_total;
                c_total += pi[i];
        }

        // normalize
        const double factor = 1.0 / c_total;
        for (i = 0; i < m_nParticles; i++)
                pi[i] *= factor;

        CalculateMean();

        GetMeanConfiguration(pResultMeanConfiguration);
        
        return CalculateProbabilityForConfiguration(pResultMeanConfiguration);
}

void CParticleFilterFrameworkFloat::CalculateMean()
{
        int i;

        for (i = 0; i < m_nDimension; i++)
        {
                last_configuration[i] = mean_configuration[i];
                mean_configuration[i] = 0.0f;
        }

        for (i = 0; i < m_nParticles; i++)
        {
                for (int j = 0; j < m_nDimension; j++)
                        mean_configuration[j] += float(pi[i] * s[i][j]);
        }
}

double CParticleFilterFrameworkFloat::CalculateProbabilityForConfiguration(const float *pConfiguration)
{
        // little "trick" for calculating the probability for the mean configuration
        // without introducing a new virtual function by using s[0]
        
        int i;
        
        for (i = 0; i < m_nDimension; i++)
        {
                temp[i] = s[0][i];
                s[0][i] = pConfiguration[i];
        }

        UpdateModel(0);

        for (i = 0; i < m_nDimension; i++)
                s[0][i] = temp[i];

        const double dResult = CalculateProbability(true);

        for (i = 0; i < m_nDimension; i++)
                s[0][i] = temp[i];
                
        return dResult;
}