Ballistic Target Geometry Parameter Estimation Based On Time-frequency Analysis

In modern military warfare,ballistic missiles have a long attack range,fast strike speed,and good maneuverability,and have a strong strategic deterrent ability as the country's most important weapon.The flight process of ballistic missiles is generally divided into three stages: initial stage,middle stage and reentry stage.Among them,the middle flight is generally outside the atmosphere and takes a long time,which is the best time to identify the target.Narrowband radar has a long detection range and can work around the clock.It is an important early warning and detection device in the defense system.It plays an irreplaceable role in target recognition and tracking detection.This article takes the mid-course ballistic missile warhead as the research object,and uses the time-frequency analysis method to analyze the micro-Doppler information of the warhead target by estimating the micro-movement and structural parameters of the warhead target with a flat bottom cone structure.The information extraction and convolutional neural network based micro-motion and structural parameter estimation methods of cone warheads lay the foundation for subsequent radar target recognition and interception.This article mainly introduces the method of space frustum target fretting and structural parameter estimation through the following sections:The first part includes the first and second chapters of this article.It mainly introduces the background of ballistic target research,the current status of domestic and foreign research,the theory of time-frequency analysis,and the derivation of the target precession micro-Doppler frequency.The second part is the third chapter of this paper,which mainly studies the method of cone target precession and structure parameter estimation based on micro-Doppler information estimation.The principle of different time-frequency transform methods is introduced first,and then the method of extracting the instantaneous micro Doppler frequency of the target from the time-frequency distribution map of the target is introduced.Finally,the precession and structural parameters of the cone warhead target are estimated based on the extracted instantaneous micro-Doppler frequency.The third part is the fourth chapter of this paper,which mainly studies the fretting and structure parameter estimation of cone target based on convolutional neural network.Firstly,the basic framework of CNN is introduced,including convolutional layer,pooling layer and activation function,etc.The forward propagation process of neural network is mainly deduced.Then it introduces the regression model of neural network and how to use CNN to estimate the fretting and structural parameters of the warhead target.Then use CNN to estimate the structure and fretting parameters of the cone target in a more realistic multi-precession cycle scenario.The fourth part is the fifth chapter of this article,which summarizes the work of the full text and looks forward to the follow-up work.