Dynamic Feature Extraction And Modeling Of Target Electromagnetic Scattering Based On Mode Decomposition

In radar systems,the specific parameter characterization the electromagnetic scattering properties of a target is the Radar Cross Section(RCS).For reasons of cost and error,simulation prediction is becoming one of the most common means of obtaining the RCS of a target.The target is often disturbed by various factors in real flight,and its RCS distribution is consequently non-linear and non-smooth.Therefore,dynamic RCS feature extraction and analysis is of great significance for the analysis of radar target characteristics in complex environments,and even for the detection,tracking and identification of moving targets.In this paper,we focus on the differences between dynamic RCS curves before and after disturbance,and carry out research on dynamic feature extraction and fast reconstruction of dynamic RCS of targets,the main work is as follows:(1)Aiming at the problem of attitude uncertainty change due to environmental disturbance when the target is flying,the target flight trajectory is studied and a dynamic RCS modeling method applicable to practical engineering is proposed.The method first simulates the flight trajectory of the target by establishing its flight dynamics model,then establishes an attitude perturbation model to simulate the influence of the environment on the attitude angle to make the motion process of the target more realistic,and finally,establishes the common level flight trajectory,obtains the real-time attitude angle by attitude solution,and calculates the dynamic RCS database.The differences of dynamic RCS data are verified by comparing the dynamic RCS curves of different perturbations;by briefly analyzing the dynamic RCS curves of 1.0° perturbation,it is found that the overall statistical characteristics extracted directly from the dynamic RCS data are difficult to quantitatively describe the dynamic characteristics of the target under different perturbation disturbances.(2)To address the problem that it is difficult to quantify the target perturbation state by extracting statistical features directly from dynamic RCS data,the differences of different perturbed dynamic RCS curves are investigated,and a method based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN)is proposed.Noise,CEEMDAN)based target RCS dynamic feature extraction method is proposed.The method treats the target dynamic RCS curve as a non-smooth random process,and firstly decomposes the unperturbed RCS data and the perturbed dynamic RCS data by CEEMDAN to obtain multiple Intrinsic Oscillatory Mode(IMF)component matrices,extract their envelope difference features and distribution difference features,and integrate the two difference features The dynamic characteristics of the target RCS curves of different perturbation states are quantitatively described.The effectiveness of the proposed method is verified by comparing the two-dimensional distribution maps of the dynamic features of different perturbations,and the method is still valid for practical applications in engineering;the superiority of the method is verified by comparing different extraction methods,and the method has the best classification effect in different perturbation states.(3)Aiming at the problems of low computational accuracy of the current interpolation algorithm at low resolution(high angular interval)and the time-consuming establishment of the database,the spatial interpolation algorithm is studied,and a dynamic RCS reconstruction method based on dynamic features and spatial interpolation algorithm is proposed.The method does not need to establish a static RCS database,firstly,the real-time attitude angle of the corresponding perturbation is used as the interpolation point coordinates,and the unperturbed RCS data with the dynamic features of the corresponding perturbation are used as known sample points,and then the spatial interpolation method is used to reconstruct the target dynamic RCS data rapidly.The effectiveness of the method is verified by comparing Spline interpolation,which is the most effective among traditional interpolation methods,and the method can still maintain a certain reconstruction accuracy without establishing a static database.Considering the problems that IDW spatial interpolation is sensitive to perturbation leading to the increase of calculation error,the IDW interpolation algorithm is investigated,and the method of preprocessing the interpolated data by using CEEMDAN algorithm is proposed to combine CEEMDAN decomposition with IDW interpolation,and the unperturbed RCS data is decomposed by using CEEMDAN to avoid the mutual influence between different frequency components,which enhances the IDW interpolation stability for perturbation changes.The low error and stability of the method are verified by comparing the reconstructed values of different interpolation methods with the calculated values of electromagnetic simulations.The method not only has the highest computational accuracy and lowest error when reconstructing the dynamic RCS with different perturbations,but also has reduced sensitivity to perturbations.