Study on a predictive filter based maneuvering target tracking algorithm

For the uncooperative flying objects, it is hard to describe the object behavior by only one dynamic model as the intentional maneuver of the object is stochastic. So the acceleration of the flying target is hard to be estimated accurately. A state estimation algorithm is proposed here by combining nonlinear predictive algorithm and extended Kalman filter (EKF) algorithm for maneuvering object, and the estimation problem for the uncertain nonzero mean acceleration in "current" model is taken for example. The acceleration of the maneuvering object is estimated by predictive filter adaptively, such that the state model of the system can be modified firstly. Then, the object states are estimated by EKF. Finally, the correctness as well as validity of the algorithm is demonstrated through a numerical simulation.