In this thesis two probabilistic model-based estimators are introduced that allow the reconstruction and identification of space-time continuous physical systems. The Sliced Gaussian Mixture Filter (SGMF) exploits linear substructures in mixed linear/nonlinear systems, and thus …