Complex Target RCS Estimation Based On Deep Learning

RCS(Radar Cross Section)is a measure of scattering performance of a target.In the real battlefield,radar stealth technology of military targets(such as aircraft,ships,etc.)directly affects their combat defense capabilities and assault capabilities.For large-scale targets such as aircraft and ships,the radar stealth performance mainly reduces RCS by various methods,that is,reduces the effective reflection power of the radar detection wave,thereby reducing the enemy radar detection range.In the field of electromagnetic scattering calculation,the acquisition of complex target RCS is through theoretical analysis,formula calculation and actual measurement.However,such methods are computationally complex,have long time periods,and have poor timeliness.They cannot meet the simple,efficient,and accurate requirements for RCS estimation and prediction of radar targets in real battlefields.In recent years,deep learning has been widely used in image recognition,speech recognition,and recommendation systems.By constructing a multi-layer neural network model,deep learning can accurately and efficiently extract the feature information of the data.The main research goal of this thesis is to realize the estimation and prediction of complex target RCS based on deep learning.This paper proposes to build a multi-level model based on LSTM-Encoder-Decoder neural network structure to learn RCS data for known azimuth and elevation directions,and to estimate and predict RCS of target unknown angle,and finally calculate with electromagnetic scattering calculation method FEKO The results were compared to verify the feasibility and effectiveness of the method.The main research work and innovations of this paper consist of the following:· A complex target RCS estimation algorithm based on deep learning is proposed.Compared with the RCS estimation method in the field of electromagnetic scattering calculation,this algorithm makes full use of the spatial information of the target,can extract and extract key information from large-scale data,estimate directly complex target RCS,and reduce the memory reduction computing ability of real-time calculation and cost,which method is simple,efficient and accurate.· Establish a complex target RCS estimation model based on deep learning.Based on recurrent neural network and long-term and short-term memory model,this algorithm of this paper extracts feature information from some complex target RCS data with different azimuth and elevation directions,and uses this feature information to achieve omnidirectional estimation and prediction of complex target RCS,which satisfies the simple,efficient and accurate requirements for complex target RCS estimation in engineering application.It has a good application prospect.· The experimental results based on the deep learning model were compared with the complex target RCS data calculated by FEKO software.This paper designs and builds a two-layer LSTM-Encoder-Decoder model,and uses the complex target RCS data calculated by FEKO to train to verify the feasibility and effectiveness of the method which estimate complex target RCS based on deep learning.