On Algorithm For Ballistic Target Tracking Filter Of Artillery Radar Based On Machine Learning

Artillery locating and fire adjusting radar is the main reconnaissance equipment of artillery.It mainly carries out two kinds of combat tasks:one is to track the ballistic target launched by the enemy’s artillery,rocket launcher and other weapons and equipment,and to position its launching point;The other is to track the ballistic target launched by our own and judge its landing point.It plays an important role in artillery operations.Existing artillery radar generally adopts the linear Kalman filter algorithm to realize the ballistic target filtering.During the filtering process,the state space linear equations are used to estimate the state and covariance of the target.Due to the nonlinear characteristics of the ballistic target trajectory large process errors are introduced in the filtering process.This paper focuses on the process error of trajectory target tracking filter algorithm and carries out the following research:First,the extended Kalman filter,unscented Kalman filter,particle filter algorithm and other nonlinear filter algorithms are combined with the ballistic differential equation model and applied to the ballistic target tracking filter.The performance and execution efficiency of different filtering algorithms are on a comparative research.Simulation results show that the nonlinear filtering algorithms have better tracking performance than the linear Kalman filtering algorithm.Second,the trajectory target movement is affected by ballistic wind and other environmental variables,which are not reflected in the nonlinear filtering model,and the process error is introduced into the filtering algorithm.In this paper,the trajectory wind identification method based on machine learning is introduced to design the model prediction filter algorithm to reduce the filtering process error under the influence of environmental parameters.Through the example simulation,the algorithm can effectively reduce the filtering error and improve the filter performance.Third,a tracking filter algorithm based on machine learning ballistic model is proposed.The nonlinear system modeling method based on machine learning was used to process the ballistic target data and establish the ballistic model.By comparing the ballistic model established by BP neural network and RBF neural network,the unscented Kalman filter algorithm based on machine learning was established.Simulation results show that the performance of this algorithm is better than that of the UKF.