Research On Radar Parmeter Estimation Method For High Dynamic Target

As the development of TBM and anti-TBM technology,flying with high-speed and maneuvers has become one of the primary means of missile penetration.Since radaris the main sensor in a missile defense system,radar signal processing and data processing methods are facing new challenges and new features.It is necessary to conduct in-depth research on parameter measurement and estimation methods for high-speed maneuvering targets,in order to enhance radar?s performance for high-dynamic target detection and tracking.For radar?s parameter estimation of high dynamic targets,this paper studies radar measurement error compensation,high-dynamic target filtering and tracking,motion model,mass-todrag ration estimation and class identification.Firstly,compensation methods of radar range-doppler coupling error and atmospheric refractive error are studied in this paper.Anopen loop range-doppler error compensation method and an Gauss iteration atmospheric refractive error compensation method are proposed,which are the foundation for accurate parameter estimation.In view of the high speed and high maneuvering characteristics of high dynamic targets,and the non-linearity of radar measurements,five filtering methods including ?-?-? filter,Kalman filter(KF),extended Kalman filter(EKF),converted measurement Kalman filter(CMKF),and Unscented Kalman filter(UKF),are studied.These filters are evaluated in filtering accuracy,computational efficiency and convergence speed.Simulation tests are performed,from which it can be seen that the UKF method has the highest filtering accuracy,and that CMKF and EKF methods have high filtering accuracy with a moderate amount of compuation.Filter tracking models of high dynamic targets are studied,which include ballistic motion model with J2 correction,the Singer model,reentry ballistic motion model with J2 correction andmaneuvering reentry ballistic motion model with J2 correction.Simuation tests are performed to compare their performce,from which it can be seen that in order to obtain high tracking filtering accuracy,it is necessary to model the acceleration of high dynamic targets in the motion model of a tracking filter.A parametric interactive multi-model filtering method is proposed,which is able to adjust the model structure according to task characteristics,target characteristics,and filght phases,to adjust the the number of models,model types,model variables according to task type and target class,to do model switch and interaction adaptively,and to obtain accuarate filtering effects under the nonlinear motion and mesuresurement models in a filter.Simulation tests have verified the effectiveness of the method,which can achieve high-precision filtering and tracking of high-dynamic targets.A method of mass-to-drag ratio estimation and class discrimination for ballistic targets is proposed.The mass-to-drag ratio is estimated by a nonlinear filter,in which it is taken as a state variable;based on the estimation,a dynamic Bayesian network model is built for discrimination of estimation convergence and target class,in which the posterior probabilities are calculated recursively,which are used for discrimination.Simulation has verified the effectiveness of the method: it is able to identify warheads and light bait targets effiently.