CPose3DPDFSOG.h

Go to the documentation of this file.

/* +------------------------------------------------------------------------+
   |                     Mobile Robot Programming Toolkit (MRPT)            |
   |                          https://www.mrpt.org/                         |
   |                                                                        |
   | Copyright (c) 2005-2020, Individual contributors, see AUTHORS file     |
   | See: https://www.mrpt.org/Authors - All rights reserved.               |
   | Released under BSD License. See: https://www.mrpt.org/License          |
   +------------------------------------------------------------------------+ */
#pragma once

#include <mrpt/core/aligned_allocator.h>
#include <mrpt/poses/CPose3DPDF.h>
#include <mrpt/poses/CPose3DPDFGaussian.h>

namespace mrpt::poses
{
/** Declares a class that represents a Probability Density function (PDF) of a
 * 3D(6D) pose \f$ p(\mathbf{x}) = [x ~ y ~ z ~ yaw ~ pitch ~ roll]^t \f$.
 *   This class implements that PDF as the following multi-modal Gaussian
 * distribution:
 *
 * \f$ p(\mathbf{x}) = \sum\limits_{i=1}^N \omega^i \mathcal{N}( \mathbf{x} ;
 * \bar{\mathbf{x}}^i, \mathbf{\Sigma}^i )  \f$
 *
 *  Where the number of modes N is the size of CPose3DPDFSOG::m_modes. Angles
 * are always in radians.
 *
 *  See mrpt::poses::CPose3DPDF for more details.
 * \ingroup poses_pdf_grp
 * \sa CPose3DPDF
 */
class CPose3DPDFSOG : public CPose3DPDF
{
    DEFINE_SERIALIZABLE(CPose3DPDFSOG, mrpt::poses)

   public:
    /** The struct for each mode:
     */
    struct TGaussianMode
    {
        TGaussianMode() : val() {}
        CPose3DPDFGaussian val;

        /** The log-weight
         */
        double log_w{0};
    };

    using TModesList = std::vector<
        TGaussianMode, mrpt::aligned_allocator_cpp11<TGaussianMode>>;
    using const_iterator = TModesList::const_iterator;
    using iterator = TModesList::iterator;

   protected:
    /** Assures the symmetry of the covariance matrix (eventually certain
     * operations in the math-coprocessor lead to non-symmetric matrixes!)
     */
    void enforceCovSymmetry();

    /** Access directly to this array for modify the modes as desired.
     *  Note that no weight can be zero!!
     *  We must use pointers to satisfy the mem-alignment of the matrixes
     */
    TModesList m_modes;

   public:
    /** Default constructor
     * \param nModes The initial size of CPose3DPDFSOG::m_modes
     */
    CPose3DPDFSOG(size_t nModes = 1);

    /** Clear all the gaussian modes */
    void clear();
    /** Set the number of SOG modes */
    void resize(const size_t N);
    /** Return the number of Gaussian modes. */
    size_t size() const { return m_modes.size(); }
    /** Return whether there is any Gaussian mode. */
    bool empty() const { return m_modes.empty(); }
    iterator begin() { return m_modes.begin(); }
    iterator end() { return m_modes.end(); }
    const_iterator begin() const { return m_modes.begin(); }
    const_iterator end() const { return m_modes.end(); }

    void getMean(CPose3D& mean_pose) const override;
    std::tuple<cov_mat_t, type_value> getCovarianceAndMean() const override;

    /** Normalize the weights in m_modes such as the maximum log-weight is 0. */
    void normalizeWeights();
    /** Return the Gaussian mode with the highest likelihood (or an empty
     * Gaussian if there are no modes in this SOG) */
    void getMostLikelyMode(CPose3DPDFGaussian& outVal) const;

    /** Copy operator, translating if necesary (for example, between particles
     * and gaussian representations) */
    void copyFrom(const CPose3DPDF& o) override;

    /** Save the density to a text file, with the following format:
     *  There is one row per Gaussian "mode", and each row contains 10
     * elements:
     *   - w (The linear weight)
     *   - x_mean (gaussian mean value)
     *   - y_mean (gaussian mean value)
     *   - x_mean (gaussian mean value)
     *   - yaw_mean (gaussian mean value, in radians)
     *   - pitch_mean (gaussian mean value, in radians)
     *   - roll_mean (gaussian mean value, in radians)
     *   - C11,C22,C33,C44,C55,C66 (Covariance elements)
     *   - C12,C13,C14,C15,C16 (Covariance elements)
     *   - C23,C24,C25,C25 (Covariance elements)
     *   - C34,C35,C36 (Covariance elements)
     *   - C45,C46 (Covariance elements)
     *   - C56 (Covariance elements)
     *
     */
    bool saveToTextFile(const std::string& file) const override;

    /** this = p (+) this. This can be used to convert a PDF from local
     * coordinates to global, providing the point (newReferenceBase) from which
     *   "to project" the current pdf. Result PDF substituted the currently
     * stored one in the object. */
    void changeCoordinatesReference(const CPose3D& newReferenceBase) override;

    /** Bayesian fusion of two pose distributions, then save the result in this
     * object (WARNING: Currently p1 must be a mrpt::poses::CPose3DPDFSOG object
     * and p2 a mrpt::poses::CPose3DPDFSOG object) */
    void bayesianFusion(const CPose3DPDF& p1, const CPose3DPDF& p2) override;

    /** Draws a single sample from the distribution */
    void drawSingleSample(CPose3D& outPart) const override;
    /** Draws a number of samples from the distribution, and saves as a list of
     * 1x6 vectors, where each row contains a (x,y,z,yaw,pitch,roll) datum */
    void drawManySamples(
        size_t N,
        std::vector<mrpt::math::CVectorDouble>& outSamples) const override;
    /** Returns a new PDF such as: NEW_PDF = (0,0,0) - THIS_PDF */
    void inverse(CPose3DPDF& o) const override;

    /** Append the Gaussian modes from "o" to the current set of modes of "this"
     * density */
    void appendFrom(const CPose3DPDFSOG& o);

};  // End of class def.
}  // namespace mrpt::poses