On the covariance of the steady-state filter predictor and smoother
H. James Rome · IEEE Transactions on Automatic Control · 1970
The pertinent algorithms are developed for determining the steady-state covariance matrices associated with continuous measurement Kalman filtering, prediction, and smoothing of time-invariant systems. The formulas represent closed-form solutions to the problem as opposed to recursively derived asymptotic solutions. They can be used as a single unified set of algorithms.