The only way is to derive it. The KF maximizes the likelihood of the observations.. You can easily form the optimization problem, "what state x maximizes the

The only way is to derive it. The KF maximizes the likelihood of the observations.. You can easily form the optimization problem, "what state x maximizes the likelihood of my observations z". I had a writeup to this effect in my class notes from Prof Roumeliotis' class. I'll see if I can post it, but you could probably derive it following the KF wiki page.