Research On Algorithm For Tracking Moving Target By Radar And Tracking Based On Cubature Kalman Filter

The tracking problem of the target is the filtering problem of the moving target state.Using the radar to measure the relevant data of the moving target and the trajectory estimation of the moving target.Accurate tracking of moving targets is important both in the military and civil area.In the military field,radar target tracking can be applied to anti-missile defense,air defense early warning,battlefield area surveillance,low altitude penetration.In the civil sector,which can be used for air defense traffic control and ground traffic control,unmanned driving and tracking and so on.The observed value is a nonlinear function,so tracking filter is a nonlinear tracking filter problem,This paper explores the tracking and filtering algorithms with high positioning accuracy and strong stability.Based on the new Sigma point filtering algorithm-Cubature Kalman Filter(CKF),the moving target tracking is studied.The main contents of the thesis are as follows:First of all,it introduces the research background and significance of the paper,as well as the domestic and foreign research status of the motion model and algorithm,and the main models and algorithms used in the paper are covered.In the second chapter,it introduces several common and used more exercise models,including CV,CA,CT,Singer model,“current”statistical model,it introduces the state equation and some characteristics of these motion models clearly,analyzes the characteristics of the Singer model,learn some of the drawbacks of the Singer model,It is proposed that the “current”statistical model is compared with the Singer model and that the “current”statistical model has a wider applicability.It analyses the uniform movement and the establishment of uniform motion target mathematical model,The basic principle and algorithm of EKF,UKF,MVEKF and CKF are introduced in detail,And then apply them to the uniform motion target filter tracking,The CKF algorithm uses the Spherical-Radial criterion to compute the mean and variance of the nonlinear function.Then the simulation results show that the CKF algorithm is suitable for constant motion target tracking.And high positioning accuracy and tracking stablely.On the basis of this,we carry out more in-depth research on CKF algorithm,Use it or its improved form in a wider motion model.Moving targets is not always in the uniform movement,more is in the mobile,Aiming at the situation of target maneuver,a strong tracking square root cubature Kalman filter is proposed based on the “current” statistical model,The algorithm introduces the time-dependent fading factor in the mean square root Kalman filter,adjusts the state covariance of one-step prediction,it has a better filtering effect,And on this basis,the system noise added parameters to adjust,at every moment,real-time update system error,making it more close to the real environment,to achieve adaptive results.The interactive multi-model describes the state of motion of a target by using two or more models,the switch between each model is determined by the Mahrkov probabilities transfer matrix.Where multiple filters work at the same time and then estimate the system state by weighting the outputs of multiple filters,which overcomes the narrower scope of application of a single model.This will combine the SCKF algorithm with the IMM model to form the SCKF-IMM algorithm,to get a better tracking effect,and update on the basis of this algorithm,nonlinear measurement update base on the results of the output,Through the simulation analysis can get that its positioning accuracy is higher.In the process of moving target tracking,if the continuous wild value occurs,it will affect the estimation of the filter,which makes the filtering accuracy lower or unstable.So the SCKF algorithm is proposed.The algorithm is based on SCKF combined with the SCNM model,the posteriori probability based on the field value comes from the variance matrix of the adaptive adjustment measurement residual.The simulation results show the anti-robust SCKF algorithm has higher accuracy than ckf algorithm in the case of continuous wild value.Reducing the impact of wild value.Through the comparison and analysis of various filtering algorithms,it is concluded that if the target is doing uniform motion,the Cubature Kalman filter algorithm with high precision and stability can be selected.If the moving target is to be motorized,High tracking effect can choose SCKF-IMM algorithm,if the target tracking process in the continuous wild value of the interference,we can add anti-wild value algorithm to the filter algorithm to achieve more accurate tracking.