Font Size: a A A

Research On Radar Recognition Technology Of Rotorcraft UAV Based On Deep Learning

Posted on:2023-12-07Degree:MasterType:Thesis
Country:ChinaCandidate:G L RenFull Text:PDF
GTID:2532306905472864Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
Although the use of rotorcraft UAV makes people’s life become convenient and quick,but it also brings a lot of social harm,so it is very important to accurately identify and detect multiple types of rotorcraft UAV.Based on the differences of micro-Doppler characteristics of different rotorcraft UAVs,this thesis realizes the classification of multiple rotorcraft UAVs through deep learning method.The main research results of this thesis are as follows:(1)Firstly,the micro-Doppler effect produced by the fretting is studied.On this basis,the mathematical model of the echo signal of the rotor blade of the rotorcraft UAV is established,and the time-frequency diagram of its radar echo is analyzed.On the basis of the measured data,the Short-Time Fourier Transform(STFT)isotime-frequency analysis methods are analyzed and compared,which not only confirms the accuracy of the rotor uav mathematical model,but also provides a sufficient theoretical basis for the identification method in this thesis.(2)Aiming at the recognition problem of single-target rotorcraft UAV,a classification recognition method based on dual-channel GoogLeNet network is proposed according to the radar measured echo data.Firstly,STFT was carried out on the echo signal of rotorcraft UAV to obtain the time-frequency spectrum of the signal,and then the Cadence-Velocity Diagram(CVD)was obtained by Fourier transform along the time axis.Then the time-frequency diagram and CVD are used as input of two-channel GoogLeNet network for feature extraction to obtain the time-frequency domain and rhythm and speed domain features of echo signal.Finally,the obtained features are input into Softmax classifier to realize the classification and recognition of rotorcraft UAV.The accuracy of the proposed classification method can reach98.9%.The proposed method is compared with single channel network to ensure the superiority of the proposed method.(3)For the recognition of multi-target rotorcraft UAV,ResNet-50 network is used to extract the features of time-frequency diagram,and then Support Vector Machine(SVM)is used to classify and identify the extracted features.It can be seen from the comparison test of classification accuracy.The proposed rotorcraft UAV classification method can reach 97.36%,which is the most suitable for multi-target rotorcraft UAV identification.
Keywords/Search Tags:Rotor UAV recognition, Micro Doppler effect, Short-Time Fourier Transform, Dual channel GoogLeNet network
PDF Full Text Request
Related items