This PR adds a new updater class: DynamicallyIteratedUpdater (DIEKF) - an implementation of algorithm 2: Dynamically iterated filter, from "Iterated Filters for Nonlinear Transition Models" by Anton Kullberg, Isaac Skog, and Gustaf Hendeby.
Additionally, this PR adds the following as required changes:
Added compatibility to pre-existing Kullback-Leibler divergence measure for GaussianState types.
Added prior state as a property of Prediction states. This is required in order for the DIEKF to access and smooth the prior state
Updated all relevant Predictor classes to assign the prior state to the output Prediction state.
This PR adds a new updater class: DynamicallyIteratedUpdater (DIEKF) - an implementation of algorithm 2: Dynamically iterated filter, from "Iterated Filters for Nonlinear Transition Models" by Anton Kullberg, Isaac Skog, and Gustaf Hendeby.
Additionally, this PR adds the following as required changes: