Optimal Tracking of Maneuvering Targets
James S. Thorp · IEEE Transactions on Aerospace and Electronic Systems · 1973
Maneuvering target motion is modeled by introducing a binary random variable in the target state equation. The optimal estimate is shown to be a weighted combination of two Kalman filter estimates with weights depending on the likelihood ratio for the detection of a maneuver. A tracking scheme is proposed for maneuvering target tracking and illustrated in an example.