Separated covariance filtering

A separated covariance filter (SCF) for estimating position and velocity given noisy measurements is discussed. The SCF calculates the state errors due to measurement errors and those due to target acceleration separately. The SCF overcomes two difficulties with the standard Kalman filter technique: (1) selection of the process noise covariance; and (2) interpreting the Kalman state covariance. The SCF is adaptable to a stream of unsequenced measurements.<
>