Handle the case where the measurement errors are non additive
In the current state, the filter codes are written under the assumption that the measurement errors enter linearly in the measurement equation. For instance in sequential_importance_particle_filter we compute the variance of the prediction errors, PredictedObservedVariance, by adding the covariance matrix of the measurement errors, H, to the covariance of the predicted observed variables.
To upload designs, you'll need to enable LFS. More information