states.vmf_state_entries
def vmf_state_entries(field: str, dim: int, meas_field: str | None = None, init_kappa: float = 20.0, kappa_obs: float = 20.0, tau: float = 10.0, kappa_min: float = 1.0) -> dict[str, State]Build the ``(direction, concentration)`` state entries for a vMF filter.
The direction entry holds a unit ``(dim,)`` mean — match it with the
existing :class:`~unitrack.costs.Cosine` cost. The concentration entry
holds a scalar, no-op'd through predict/update because the direction
entry's :class:`VonMisesFisherDecay` / :class:`VonMisesFisherUpdate`
write it as a side effect.
Parameters
| Name | Type | Description |
|---|---|---|
| field | str | Name of the direction field (and prefix for ``f"{field}_kappa"``). |
| dim | int | Embedding dimensionality. |
| meas_field = None | str | Detection field supplying the measured embedding. Defaults to :paramref:`field`. |
| init_kappa = 20.0 | float | Concentration assigned to a freshly spawned tracklet. |
| kappa_obs = 20.0 | float | Per-observation concentration used by the update. |
| tau = 10.0 | float | Concentration decay time constant used by the predict step. |
| kappa_min = 1.0 | float | Floor on the concentration after decay. |
Returns
dict — Two :class:`~unitrack.states.State` entries keyed by ``field`` and ``f"{field}_kappa"``.
Source: unitrack/states/directional.py:159