Research On High Speed Maneuvering Target Tracking Model And State Estimation Algorithm

With the continuous updating and breakthrough of technology in the near space target field,the maneuvering forms of hypersonic vehicles are more complex and changeable,and their powerful penetration capabilities pose severe challenges to traditional defense systems.Therefore,the research on the tracking filtering method with high tracking accuracy and strong robustness has become the focus of the proximity defense technology.This article takes boost gliding aircraft as the research object,and conducts in-depth research on the tracking filtering method of its trajectory:In order to fully understand the motion process of the free gliding segment of the hypersonic vehicle,considering the earth’s rotation and flatness ratio,the differential equation of motion of the gliding aircraft is established under the ballistic system,and the aerodynamic model of the aircraft is mainly studied,and the aerodynamic parameters are given.According to the fitting formula,according to the safe flight and combat requirements of the aircraft,the Gaussian pseudo-pseudo method is used to optimize the design of the free cruise section of the aircraft,and the change process of the ballistic characteristics and c ontrol quantity is analyzed through numerical simulation.In order to accurately describe the motion characteristics of high-speed maneuvering targets,this article establishes the target motion description from two aspects of kinematics and dynamics.Summarized several traditional motion model modeling methods,and compared the advantages and disadvantages of each model.Aiming at the shortcomings of low tracking accuracy and poor adaptability of the kinematics model,a dynamic tracking model is further pr oposed,and the method of joint state and parameter estimation is used to realize online modeling of target motion.Determine the measurement model of the system,complete the three-dimensional positioning of the target based on the least square algorithm and analyze the accuracy of the two-station positioning.Due to the strong nonlinearity of the tracking system,the paper comprehensively compares the characteristics of the nonlinear filtering algorithm,uses cubature Kalman filtering,and combines each tracking model to optimize the ballistic tracking simulation,and analyzes the convergence speed and convergence accuracy of each model.In order to improve the filtering accuracy,the fifth-order spherical-phase-diameter volume rule is used to improve the sampling strategy,and a high-degree cubature Kalman filter is proposed.In practical engineering applications,due to the complex working environment of the detector,the detection data will be disturbed by non-Gaussian noise.For the case where the measurement noise is non-Gaussian,The maximum likelihood estimation method proposed by Huber is used to improve the high-degree cubature Kalman filter and convert the measurement update problem into linear regression.Solve the problem,and the simulation results verify the advantages of the algorithm in terms of accuracy and robustness under mixed Gaussian noise.When tracking unknown targets,aiming at the uncertainty of the system model caused by the target maneuver: the unknown nature of the target motion m odel and the unpredictability of the target maneuver,an online feedback filter based on the Elman network is designed to monitor the target maneuver online,and the state estimation is corrected in real time.The network structure and algorithm implementation process are elaborated in detail,and the network training process is derived.Embed the trained network into the filter to form a hybrid filter,and predict the estimated error based on the input.Aiming at the problem of accuracy of network prediction,threshold detection is set for the predicted value of the network,and simulations verify the effectiveness of the algorithm.