Technology Of Radar Target Tracking

In modern wars,the target localization and tracking is closely related to the informatization development of modern warfare.According to the large amount of information,data obtained on the battlefield,the goal of the detection,localization and tracking of target can be achieved,and the most timely target motion information for the weapons can be provided,so the weapon can be controlled to achieve the precision strike and intercept of the target.Therefore,it is one of the most basic functions of the modern target tracking technology to realize the accurate positioning and tracking of the target,so as to control the development of the war.The tracking is done in the discrete time system,and the tracking is the estimation of the state parameters of the target using the estimated value and the measured value of the target.From the perspective of the target tracking,the complexity of the target tracking has two reasons,one is the uncertainty of the measurement source,which is multi-objective and false alarm,the other is due to the uncertainty of target model parameters,which is produce of the motor phenomenon.From the point of system,the tracking performance of the system is not only related to the algorithm itself,but also to the linearit y of the system itself.Firstly,the method of parameter estimation is discussed in this thesis.The methods include least squares estimation,minimum mean square error estimation and linear minimum mean square error.Linear mean square filter is optimal linear filter,kalman filter(KF)is based on the minimum mean square error estimation algorithm,which uses recursive method to achieve.Secondly,the linear kalman filter(KF)and the extended kalman filter(EKF)are described in detail and the azimuth measured datum are used for the simulation.Kalman filter(KF)uses the state variable method to describe the system.Under the state variable method,the input and output of the system are expressed by state transition model and output observation model.Then the optimal filtering method of linear system can be used to estimate the target state.The basic theory of kalman filter(KF)is only applicable to linear system at first.In the practical application,because the azimuth measured datum are composed of oblique distance and target angle,whereas the linear kalman filter(KF)is based on the horizontal and vertical coordinates to set up state transition model and output observation model.Therefore the nonlinear problem is transformed into an approximate linear problem by linearization method,and then the linear kalman filter(KF)is used to filter it.The most commonly used method is the Taylor series expansion,in order to establish the extended kalman filter(EKF).The extended kalman filter(EKF)is the most commonly used nonlinear filtering method,and it requires the solution of the Jacobian matrix or Hessian matrix.It is suitable for nonlinear systems with small linearization errors and higher filtering accuracy can be obtained.At the same time,the computation is ver y small,and it is very widely used.Thirdly,the probabilistic data association algorithm is introduced and simulated in detail.Probabilistic data association algorithm is a kind of quasi Bayesian algorithm,which considers all the plots in association gate are likely derived from the target,but the probabilities that derived from the target are different,and filters according to the trace information and probability.Finally,the azimuth measured datum are validated based on the extended kalman filter(EKF)algorithm.The azimuth measured datum in clutter environment are processed with probabilistic data association algorithm,and the tracking effect is good.