Track filters for correlated measurements and target maneuvers

The effect of correlated measurement errors on target tracking accuracies is studied by means of Kalman filter theory. The measurement error is modeled as a first-order Markov process characterized by a correlation time ( tau ) and a position error variance. The target motion is described by one of two dynamic models: the exponentially correlated velocity (ECV) model and the exponentially correlated acceleration (ECA) model. The analysis is based on two new solutions for the continuous-time, steady-state Kalman covariance. For both the ECV and ECA target models, the velocity variance exhibits resonant behavior whenever the target maneuver bandwidth is close to 1/ tau .<>