Research On Target Tracking Algorithm Of Rocket Trajectory

As the only carrier for the human to explore the universe,the safe and stable flight of the rocket directly determines the successful completion of aviation missions of various spacecraft.In order to accurately track the state information of the rocket in its flight process,its trajectory information is obtained through a variety of measurement and control equipment in test field.High-precision and complete trajectory information provides important guarantee for mission decision-making,safety control,flight guidance,etc.With the development of measurement and control system,the use of various advanced measurement and control equipment,derived different forms of rocket trajectory measurement and control system,which put forward increasingly severe requirements for the rocket traj ectory tracking and positioning system.This thesis focuses on the external traj ectory measurement system of rocket and establishes the trajectory parameter fitting system by using the algorithms of multi-source measurement data location and trajectory smoothing filter,and then the target tracking model is established by using Unscented Kalman filter algorithm and Elman neural network.Finally,the software of trajectory tracking and positioning system is designed and developed based on Matlab GUI platform.The specific research content of this thesis is summarized as follows:(1)Aiming at the crossing location problem of multiple measuring devices,this thesis uses the least squares method and the recursive least squares method to realize multi-source trajectory data location.Besides,for a multi-source measurement system which only contains angle measurement information,this thesis uses the sum of the minimum distance between the target and each station to obtain the trajectory parameters.Finally,according to the actual engineering requirements,this thesis proposes a single/multiple automatic switching solver algorithm.The experiment shows that this algorithm is feasible and effective.(2)Aiming at the problem that trajectory calculation result is influenced by the random error of measurement equipment,which leads to a large deviation between estimated result and actual trajectory,based on the parameter regression model,this thesis respectively establishes a trajectory prediction model based on cubic B-spline and sliding polynomial,realizes smooth filtering of trajectory solution results.The experimental data shows that both algorithms can obtain higher accuracy trajectory estimation results.(3)Aiming at the problem of tracking flight target with strong maneuverability,this thesis presents an Unscented Kalman filter algorithm based on Elman neural network,which compensates the filtering error learned by the Elman neural network to the result of the Unscented Kalman filter.Based on experimental results,the algorithm proposed in this thesis has obvious improvement in filtering accuracy and stability compared with the Unscented Kalman filtering algorithm.(4)In this thesis,the trajectory tracking and positioning system is designed and developed,which mainly includes the functions of user login,data reading,trajectory solution,trajectory smoothing,velocity/acceleration solution and automatic report generation,which reduces the processing time of external trajectory measurement data and improves the efficiency of rocket trajectory information evaluation.