I never really understood Kalman filters, but there was a time I knew how to design non-optimal state (Luenberger?) observers, which are a lot easier to design and implement. I wonder if discussing those first would make things easier for the audience.