Skip to content

structs.ukf_data

Settings of the unscented Kalman filter.

SigmaPointsSettings

Parameters and weights of the Van der Merwe sigma points.

create(n, alpha, beta, kappa=0.0) classmethod

Compute the weights from the scaling parameters.

Parameters:

Name Type Description Default
n int

The dimension of the state.

 required
alpha float

The spread of the sigma points around the mean.

 required
beta float

Prior knowledge of the distribution, 2 is optimal for Gaussians.

 required
kappa float

Secondary scaling parameter.

 0.0

Returns:

Type Description
SigmaPointsSettings

The settings with the weights for the mean and the covariance.
Source code in drone_estimators/structs/ukf_data.py 
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
@classmethod
def create(cls, n: int, alpha: float, beta: float, kappa: float = 0.0) -> SigmaPointsSettings:
    """Compute the weights from the scaling parameters.

    Args:
        n: The dimension of the state.
        alpha: The spread of the sigma points around the mean.
        beta: Prior knowledge of the distribution, 2 is optimal for Gaussians.
        kappa: Secondary scaling parameter.

    Returns:
        The settings with the weights for the mean and the covariance.
    """
    lambda_ = alpha**2 * (n + kappa) - n
    c = 0.5 / (n + lambda_)
    Wc0 = np.array([lambda_ / (n + lambda_) + (1 - alpha**2 + beta)])
    Wm0 = np.array([lambda_ / (n + lambda_)])
    Wc = np.full(2 * n, c)
    Wm = np.full(2 * n, c)
    Wc = np.concat((Wc0, Wc))
    Wm = np.concat((Wm0, Wm))

    return cls(n, alpha, beta, kappa, lambda_, Wc, Wm)

UKFSettings

Settings of the unscented Kalman filter.

create(SPsettings, Q, R, fx, hx) classmethod

Create the settings from the sigma point settings, noise covariances, and the model.

Parameters:

Name Type Description Default
SPsettings SigmaPointsSettings

The sigma point settings.

 required
Q Array

The process noise covariance.

 required
R Array

The measurement noise covariance.

 required
fx Callable[[Array, Array, Array, Array, Array, Array, Array | None, Array | None], tuple[Array, Array, Array, Array, Array | None]]

The continuous dynamics function.

 required
hx Callable[[Array, Array, Array, Array, Array, Array, Array | None, Array | None], Array]

The observation function.

 required

Returns:

Type Description
UKFSettings

The settings.
Source code in drone_estimators/structs/ukf_data.py 
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
@classmethod
def create(
    cls,
    SPsettings: SigmaPointsSettings,
    Q: Array,
    R: Array,
    fx: Callable[
        [Array, Array, Array, Array, Array, Array, Array | None, Array | None],
        tuple[Array, Array, Array, Array, Array | None],
    ],
    hx: Callable[[Array, Array, Array, Array, Array, Array, Array | None, Array | None], Array],
) -> UKFSettings:
    """Create the settings from the sigma point settings, noise covariances, and the model.

    Args:
        SPsettings: The sigma point settings.
        Q: The process noise covariance.
        R: The measurement noise covariance.
        fx: The continuous dynamics function.
        hx: The observation function.

    Returns:
        The settings.
    """
    return cls(SPsettings, Q, R, fx, hx)