Interesting approach to unifying motion models. Handling continuity and temporal consistency in these models is always the hardest part, especially when dealing with noise or sparse data over time. Have you looked into how state verification handles drift across long trajectories?