Research On Radar Target Tracking Method

Since the birth of target tracking theory in the 1950 s, it has been widely used in military and civilian fields, such as ballistic missile defense, battlefield surveil ance, precise missile guiding, air traffic control, navigation, robot vision, etc. The typical target tracking problem is essential y a optimal state estimate problem that target kinematic state can be estimated real-timely and accurately from its noisy corrupted measurements.To begin with, this paper describes the basic concepts, structure and principles of object tracking systems, as well as the fundamental issues need to be addressed. Then we discusses some results of linear estimation theory and the space model of the linear stochastic system which is the basic knowledge of tracking problems, and leads to the study of Kalman filter. Kalman filter, based on a minimum mean square error criterion, is an optimal linear state filter. Furthermore, as the most commonly used radar tracking algorithm at present, Kalman filter has become a primary and dominant tracking algorithm. However, Kalman filter can only be applied in the linear system case, which is not usual y seen in reality. Therefore research is made on the extended Kalman filter for nonlinear systems. We can obtain extended Kalman filter by using Taylor series expansion to transform a nonlinear model to a approximate linear model. Extended Kalman filter is a sub-optimal filter. Besides, due to the great calculation amount, Jacobian matrix is difficult to get. To solve the defect of the huge calculation of Kalman filter algorithm, this paper studies the α-β filter algorithm, which is simple in structure and commonly used in engineering. Afterwards, this paper discussed the preprocessing of radars’ raw data which is previous to the filtering process, including outliers identification and elimination, de-biasing converted measurement. This paper also studies the data association method of information fusion theory. Data association process is a process to determine the associated relationship between measurement information that radar receives and target source. This paper deals with the nearest neighbor method, which is easy to implement and calculate. However, the result of this method will be poor, if the working environment has dense clutter and low signal-to-noise ratio. In addition, combined with a project, specific and practical application target tracking theory is carried out. Finally, a summary and outlook of this paper is made, aiming at its drawbacks.