Micro-motion Feature Analysis And Extraction Of Rotor UAV Radar Echoes

With the rapid development of the rotor UAV industry,the "black flying" UAVs have brought greater hidden dangers to the safe operation of airports,prisons,and nuclear power plants.Micro-Doppler features are the key to target recognition and classification,as a "low,slow,small" flying target,the UAV radar echoes is relatively weak and easy to be interfered by noise.Noise,motion state and initial phase are the main factors that make it difficult to estimate the micro-motion parameters of rotor UAV.Research on UAV micro-motion feature analysis and extraction methods to improve time-frequency spectrum resolution and micro-motion parameter estimation accuracy,which can provide corresponding theoretical foundation and technical support for UAV surveillance,identification and classification,and has certain theoretical and practical application significance.First,the micro-motion feature analysis methods are described,and the RSP(Reassigned Spectrogram)method is introduced.The performance of the RSP is analyzed and compared with the mainstream time-frequency analysis methods in terms of resolution and cross-term.At the same time,the influence of UAV micro-motion parameters on Doppler features is studied,which provides theoretical basis for the subsequent micro-motion feature extraction technology.Secondly,aiming at the problem of micro-motion feature extraction and parameter estimation of rotor UAV in hover state,a method based on RSP-CFD(Reassigned Spectrogram-Cadence Frequency Diagram)for micro-motion feature extraction and parameter estimation of rotor UAV in hover state is proposed.The RSP time-frequency diagram obtains the Doppler information of the UAV rotor blade,and on this basis,the CFD method is introduced to extract the UAV blade flash frequency,and finally the maximum value parameter estimation method is combined to estimate the UAV rotation speed and blade length.By comparing with traditional methods,the performance of feature extraction and parameter estimation of the proposed method is analyzed.Finally,aiming at the problem of micro-motion feature extraction and parameter estimation of rotor UAV in motion state,a method based on BPD-Cepstrum(Basis Pursuit Denoising-Cepstrum)for micro-motion feature extraction and parameter estimation of rotor UAV in motion state is proposed.First,the BPD method is used to reconstruct the echoes signal spectrum to achieve the purpose of denoising the UAV signal.Secondly,the maximum Doppler of the rotor blade echoes is extracted by the RSP combined with the maximum value parameter estimation method,the cepstrum method is applied to extract the blade flash frequency of UAV,and then realize the estimation of rotation speed and blade length.The experimental results of simulation and measured data demonstrate the effectiveness of the algorithm.