Improve the Tracking of the Target in Radar System Using the Particular Filter

Abstract

In this paper we are interested to improve the state estimation of a maneuvering target, which improve the tracking of it in real time in radar system. Trajectory estimation is posed as a continuous filtering problem with a simple closed –form solution, which is used a turn to update the label cost. Many of the most successful tracking methods at present perform tracking by detection, the target is presented by an object model that can be detected in every frame independently. The advantage of using an object detector are that it naturally handles re-initialization if a target has been lost, and that it avoids excessive model drift. The detector yields the per-frame evidence for the presence or the absence of the target using the false alarm in discrete time. The aim is to improve the process of the state estimation for too random target at the given instant in order to smooth their true state path for a long time, it simplifies the process of real-time tracking. In this frame, we propose the numerical methods presented by MONTE CARLO (MC) counter-part the methods conventionally used named Extended KALMAN Filter (EKF), in order we showed that the first are more successful.