Research On Radar Key Technology For Small UAV Detection

Due to the low altitude environment,small radar cross section(RCS)and complex motion of target,conventional radar systems have very limited capability for drone detection.The research of novel radar theories and methods for drone detection is necessary and urgent.Regarding the requirements mentioned above,the key technologies of radar for drone detection is considered in this dissertation.The main works and contributions of this dissertation are as follows:1.For the problems of weak radar returns and strong clutter background in low altitude,this dissertation introduces the long time coherent integration method to detect drones via radar,and proposes an accelerating coherent integration algorithm,i.e.,parametric symmetric instantaneous autocorrelation function,keystone transform and scaled Fourier transform(PSIAF-KT-SFT).This algorithm first eliminates the second order range migration and Doppler frequency migration induced by the target motion via PSIAF.Then,the residual linear coupling is corrected by KT and SFT.PSIAF-KT-SFT realizes the target energy accumulation,and thus it can increase the output signal-to-noise ratio(SNR)and can improve the radar detecting ability for drones.Simulation and real data demonstrate the effectiveness of the proposed algorithm.2.For the problems of SNR loss and poor multiple targets detection ability induced by the cross-terms of PSIAF-KT-SFT and for removing Doppler frequency ambiguity,two linear coherent integration methods,i.e.,keystone transform and modified generalized Radon Fourier transform(KT-MGRFT)and two steps scaling and fractional Fourier transform(TSS-FrFT),are proposed.KT-MGRFT corrects linear range migration by KT,and then it achieves the residual migration correction and target energy integration via jointly searching range,Doppler frequency ambiguity coefficient and acceleration.KTMGRFT further improves the radar anti-noise ability,and it can detect multiple drones and obtain the motion parameters of each target in low SNR environment.TSS-FrFT utilizes FrFT to remove the second order range migration and Doppler frequency migration,then it realizes the coherent integration in range and Doppler frequency domain.Compared with KT-MGRFT,TSS-FrFT can remove the blind speed side-lobe caused by the joint search of range and velocity,and so it decreases the false alarms of radar system.3.For the drones with complex motion state,on the one hand,this dissertation introduces high-order motion parameters and constructs the radar returns by using the polynomial motion model.And a fast high-order range migration and Doppler frequency migration correction method is proposed to achieve the long time coherent integration.First,the auto-correlation function is introduced to correct the range migration.Then,the polynomial-phase transform(PPT)is employed to eliminate the high-order Doppler frequency migration.Finally,the Lv's Distribution(LVD)is presented to achieve coherent integration.The theoretical analysis and experimental results show that the computational complexity is low and the estimation accuracy of motion parameters is high.On the other hand,for multiple different motion states of drone in long coherent integration interval,this dissertation presents a multiple motion state coherent integration method.The reference signal is constructed to compensate the other motion states to be the initial state.After that,the GRFT is utilized to achieve coherent integration.This coherent integration method improves the radar detection ability for the drone with multiple different motion states.Furthermore,the adjacent state transition equations of the initial range,initial velocity and starting time are employed to decrease the computational cost.4.For drone recognition and detecting the hovering aircraft,this dissertation proposes a method for drone detection and micro-motion parameters estimation based on the micro-motion feature of rotor blades.Firstly,the radar returns of rotor blades are modeled as sinusoidal frequency-modulated(SFM)signals related to the rotation speed and blade length.Then,the theoretical analysis and experimental results indicate that the high carrier frequency of radar,long blade length and large difference of rotation angles of different blades within ? can obtain great expectation of the micro-Doppler frequency difference of blades.The larger the expectation is,the easier it is to distinguish radar echoes of different blades in time-frequency domain.Finally,based on the micro-Doppler frequency spectrum,the radar returns of different rotor blades can be separated in parameters domain via Hough-Radon transform(HRT)and two-dimensional constant false alarm rate(CFAR)detecting.Meanwhile,the drone detection and micro-motion feature extraction are achieved.