states.InformationProcess
class InformationProcess:Random-walk predict step in information form.
Adds isotropic process noise ``Q = q * dt * I`` to the covariance,
expressed on the information matrix via the Woodbury identity
.. math::
Y' = (Y^{-1} + cI)^{-1} = Y - Y (Y + c^{-1} I)^{-1} Y,
\quad c = q\,dt,
then rescales the information vector to preserve the mean
(``F = I`` leaves ``mu`` unchanged, so ``y' = Y' mu``). A non-positive
``dt`` is a no-op.
Parameters
| Name | Type | Description |
|---|---|---|
| field | str | Primary mean field. The information matrix lives in ``f"{field}_infomat"`` and the information vector in ``f"{field}_infovec"``. |
| q = 0.01 | float | Per-unit-time process-noise scale. |
Members
property
- field
- infomat_fieldReturn the information-matrix field name.
- infovec_fieldReturn the information-vector field name.
- q
Source: unitrack/states/kalman/information.py:50