ProcessNoiseSettings#

class ProcessNoiseSettings(type: str, Q_cont: CovarianceMatrix, beta: ndarray | None = None, max_interval=None, body=None)#

Bases: object

Process noise configuration for the filter.

A lightweight container that selects the process noise model type and stores its continuous-time covariance and model-specific parameters. Passed to FilterSettings and consumed by FilterOD subclasses to build the concrete ProcessNoise object.

Parameters:
typestr

Process noise model type: 'SNC' (State Noise Compensation), 'DMC' (Dynamic Model Compensation), or 'STOC' (stochastic).

Q_contCovarianceMatrix

Covariance \(Q_c\) of the input held constant over each timeline interval: for SNC the unmodeled acceleration (km²/s⁴), for DMC the input driving the stochastic acceleration (km²/s⁶).

betanp.ndarray, optional

DMC only — decay rates \(\beta\) (1/s) of the first-order Gauss–Markov stochastic acceleration. The filter takes them from the dynamics: the beta_fogm state when there is one, else the fogm_beta of the ForceModelTranslation. When given, they must match those, or the filter raises ProcessNoiseBetaMismatchError. Defaults to None.

max_intervalArrayWUnits or float, optional

Longest interval the noise is accumulated over, in seconds when a float. A longer interval, such as a data gap, receives the noise of one this long, so the covariance does not blow up across it. See _noise_intervals(). Defaults to None, no limit.

bodyBody or int, optional

Body whose dynamics the noise drives, or its SPICE ID. The noise acts on that body’s position and velocity (and a_fogm for DMC), wherever they sit in the state. Give the filter one settings object per body to noise several propagated bodies. Defaults to None, the primary body of the trajectory.

Attributes:
Q_cont

Covariance of the input held constant over each timeline interval: the unmodeled acceleration (SNC), or the input driving the stochastic acceleration (DMC).

beta

DMC only — decay rates \(\beta\) (1/s) of the Gauss–Markov stochastic acceleration, checked against the dynamics, or None to take them from the dynamics.

body

Body whose dynamics the noise drives (a Body or its SPICE ID), or None for the primary body of the trajectory.

max_interval

Longest interval the noise is accumulated over, in seconds, or None for no limit.

type

Process noise model type: 'SNC', 'DMC', or 'STOC'.

Raises:
ValueError

If type is not one of 'SNC', 'DMC', or 'STOC'.

ProcessNoise.NonPositiveIntervalError

If max_interval is not a positive time.

See also

scarabaeus.FilterSettings

Bundles this object with the initial covariance.

scarabaeus.StateNoiseCompensation

SNC process noise model.

scarabaeus.DynamicalModelCompensation

DMC process noise model.

Notes

The concrete ProcessNoise subclass (StateNoiseCompensation or DynamicalModelCompensation) is instantiated lazily by the filter, not here.

property Q_cont: CovarianceMatrix#

Covariance of the input held constant over each timeline interval: the unmodeled acceleration (SNC), or the input driving the stochastic acceleration (DMC).

property beta: ndarray | None#

DMC only — decay rates \(\beta\) (1/s) of the Gauss–Markov stochastic acceleration, checked against the dynamics, or None to take them from the dynamics.

property body#

Body whose dynamics the noise drives (a Body or its SPICE ID), or None for the primary body of the trajectory.

property max_interval: float | None#

Longest interval the noise is accumulated over, in seconds, or None for no limit.

property type: str#

Process noise model type: 'SNC', 'DMC', or 'STOC'.