The tl;dr of a Kalman filter is - you have observations that are normally distributed around a linear function of an unknown state - you have some state updat

The tl;dr of a Kalman filter is - you have observations that are normally distributed around a linear function of an unknown state - you have some state update function where the next state is normally distributed around a linear function of the unknown current state - your beliefs about your initial state are normally distributed then a kalman filter will update your beliefs about the state of your system and do as good a job as could be hoped for. There are also simple generalizations that relax some of those assumptions. This is one of those algorithms that anyone working with data should know. It answers the dynamical question of wtf is going on now, given what I’ve seen over time.