Two-stage Kalman estimator for tracking maneuvering targets

A two-stage Kalman estimator was investigated to solve an estimation problem in the presence of random bias. Under an algebraic constraint, the optimal estimate of the system state can be obtained as a linear combination of the output of the first stage (a `bias-free´ filter) and the second stage (a bias filter). By treating the bias vector as a target acceleration, the two-stage Kalman estimator can be used for tracking maneuvering targets. The first stage contains a constant velocity target model and produces the target position and velocity estimates. The second stage provides estimates of the target acceleration. During the maneuver, the output of the second stage is used to correct the output of first stage. In the absence of the maneuver, the second stage is turned off and the first stage provides the target state estimates