public void FitTest7()
{
UnivariateDiscreteDistribution target = CreateUnivariateDiscreteDistribution();
double[] observations = { 0, 1, 1, 1, 1 };
double[] weights = { 0.125, 0.125, 0.25, 0.25, 0.25 };
target.Fit(observations, weights);
double mean = Measures.WeightedMean(observations, weights);
Assert.AreEqual(mean, target.Mean);
}
public void FitTest4()
{
UnivariateDiscreteDistribution target = CreateUnivariateDiscreteDistribution();
double[] observations = { 0, 1, 1, 1, 1 };
IFittingOptions options = null;
target.Fit(observations, options);
double mean = Accord.Statistics.Tools.Mean(observations);
Assert.AreEqual(mean, target.Mean);
}
public void LogProbabilityMassFunctionTest()
{
UnivariateDiscreteDistribution target = CreateUnivariateDiscreteDistribution();
double p = 0.42;
double q = 1 - p;
double lnp = System.Math.Log(p);
double lnq = System.Math.Log(q);
Assert.AreEqual(lnq, target.LogProbabilityMassFunction(0));
Assert.AreEqual(lnp, target.LogProbabilityMassFunction(1));
}
/// <summary>
/// Construct a new Vehicle object from its initial state.
/// </summary>
/// <param name="initial">Initial pose.</param>
/// <param name="landmarks">Landmark 3d locations against which the measurements are performed.</param>
/// <param name="measurer">Measuring config and methods.</param>
public SimulatedVehicle(PoseT initial, List <double[]> landmarks, MeasurerT measurer)
: base(initial, measurer)
{
ClutterCount = ClutterDensity * measurer.Volume();
if (ClutterCount > 0)
{
clutterGen = new PoissonDistribution(ClutterCount);
}
else
{
clutterGen = new DegenerateDistribution(0);
}
Landmarks = landmarks;
HasDataAssociation = true;
DataAssociation = new List <int>();
}
public void ProbabilityMassFunctionTest()
{
UnivariateDiscreteDistribution target = CreateUnivariateDiscreteDistribution();
double p = 0.42;
double q = 1 - p;
Assert.AreEqual(q, target.ProbabilityMassFunction(0));
Assert.AreEqual(p, target.ProbabilityMassFunction(1));
double[] observations = { 0, 1, 0, 0, 1, 0 };
target.Fit(observations);
p = target.Mean;
q = 1 - p;
Assert.AreEqual(q, target.ProbabilityMassFunction(0));
Assert.AreEqual(p, target.ProbabilityMassFunction(1));
}
/// <summary>
/// Perform a deep copy of another general vehicle,
/// scaling the covariance matrices and setting a new detection probability.
/// </summary>
/// <param name="that">Copied vehicle.</param>
/// <param name="motioncovmultiplier">Scalar multiplier for the motion covariance matrix.</param>
/// <param name="measurecovmultiplier">Scalar multiplier for the measurement covariance matrix.</param>
/// <param name="pdetection">Probability of detection.</param>
/// <param name="clutter">Clutter density.</param>
/// <param name="copytrajectory">If true, the vehicle historic trajectory is copied. Relatively heavy operation.</param>
public SimulatedVehicle(Vehicle <MeasurerT, PoseT, MeasurementT> that,
double motioncovmultiplier, double measurecovmultiplier,
double pdetection, double clutter, bool copytrajectory = false)
: base(that, copytrajectory)
{
this.detectionProbability = pdetection;
this.ClutterDensity = clutter;
this.ClutterCount = this.ClutterDensity * this.Measurer.Volume();
this.motionCovariance = motioncovmultiplier.Multiply(that.MotionCovariance);
this.MeasurementCovariance = measurecovmultiplier.Multiply(that.MeasurementCovariance);
if (this.ClutterCount > 0)
{
this.clutterGen = new PoissonDistribution(this.ClutterCount);
}
else
{
this.clutterGen = new DegenerateDistribution(0);
}
}
public void EntropyTest()
{
UnivariateDiscreteDistribution target = CreateUnivariateDiscreteDistribution();
double q = 0.42;
double p = 1 - q;
double actual = target.Entropy;
double expected = -q *System.Math.Log(q) - p * System.Math.Log(p);
Assert.AreEqual(expected, actual);
target.Fit(new double[] { 0, 1, 0, 0, 1, 0 });
q = target.Mean;
p = 1 - q;
actual = target.Entropy;
expected = -q *System.Math.Log(q) - p * System.Math.Log(p);
Assert.AreEqual(expected, actual);
}
