Study On Multi-station Radar Target Tracking Algorithm

Target tracking is a very important component in the field of radar signal processing.The target tracking technology obtains the state information of the target of interest through the relevant measurement values obtained by the sensors,and then realizes the prediction and estimation of the target motion state through an appropriate filtering algorithm,thereby achieving the purpose of stable tracking.Among them,the choice of filtering algorithm plays a decisive role in the accuracy of tracking results.Therefore,this thesis first studies the tracking performance of different filtering algorithms in linear scenes and nonlinear scenes,and analyzes their respective advantages and disadvantages.Secondly,in order to obtain higher target tracking accuracy,this thesis has also studied multi-sensor information fusion algorithms,including the covariance intersection fusion algorithm and information matrix track fusion algorithm suitable for synchronous observation scenarios,and the sequential measurement asynchronous fusion algorithm suitable for asynchronous scenarios.Based on this,the main work of this thesis is given:1.In terms of filtering algorithms: Firstly,this thesis introduces Bayesian filtering theory,which is the theoretical basis for subsequent research in this thesis.Subsequently,the filtering algorithms based on Bayesian theory are studied,including the Kalman filter algorithm(KF)for linear scenes,and the extended Kalman filter algorithm(EKF),the unscented Kalman filter algorithm(UKF)and the converted measurement Kalman filter algorithm(CMKF)for non-linear scenes.In order to analyze the tracking performance of each filtering algorithm in non-linear scenes,the comparative analysis of the simulation results shows that the Kalman filter can only obtain high tracking accuracy in a linear Gaussian scene;in a non-linear scene,because EKF,UKF and CMKF use different methods to deal with nonlinearity,the tracking performance is also different.Among them,EKF has the largest filtering error,and it is suitable for scenes with less nonlinearity;UKF has higher tracking accuracy than CMKF,but UKF has a larger calculation amount and longer filtering time than CMKF.2.In terms of multisensor information fusion: Because the cross-covariance matrix of the estimation error between the local sensors is unknown in the distributed tracking,and the calculation of the cross-covariance matrix is extremely complicated,this thesis studies two kinds fusion algorithms that can fuse information from local sensors well without the known cross-covariance matrix,which are the covariance intersection fusion algorithm(CI)and information matrix track fusion algorithm.Then these two algorithms are applied to the fusion of information from missile and ground-based radar and form ground-based radars.Simulation results show that these fusion algorithms can obtain good fusion tracking results.Then the fusion accuracy of these two fusion algorithms is compared.Secondly,this thesis also studies the difference between fusion with and without feedback.Simulation results show that although the fusion estimation solution is equivalent to the fusion center,the fusion algorithm with feedback can significantly improve the tracking performance of local sensors.Finally,a sequential measurement asynchronous fusion algorithm is introduced,and the effectiveness of the algorithm is verified by simulation.