The Kalman filter30 is a minimum

The Kalman filter30 is a minimum-variance filter in which time-series measurements are incorporated recursively into estimates of state variables it is the optimal, Bayesian least-squares estimator for linear dynamic systems.-The Kalman filter30 is a minimum-variance filter in which time-series measurements are incorporated recursively into estimates of state variables it is the optimal, Bayesian least-squares estimator for linear dynamic systems.

Roweis_1999Neural Computation_A Unifying Review of Linear Gaussian Models.pdf