Research On Integration And Super-resolution Detection Methods Of UAV Swarms Based On Radar

With the development of communication and control technology,group targets,such as formations,targets with passive or active interference and unmanned aerial vehicle(UAV)swarms,have become more common,which poses a great challenge to the resolution of radar.In recent years,the use of UAV swarms to achieve swarm intelligence and swarming UAVs combats has attracted the attention of many countries.Because of the capabilities of all-day and all-weather operations and long-distance detection,radar has become an indispensable way to detect UAV swarms.However,it is an arduous task for radar with limited bandwidth(transmitted)signal,coherent processing interval and aperture to detect UAVs,because of the low signal-to-noise ratio(SNR)of radar echo,the small distances between UAVs and the similar movement states of them.In recent years,the development of long-time coherent integration technology and sparse signal recovery technology has brought new solutions for processing radar echoes in the above scenarios.The long-time coherent integration technology can achieve a long-time integration of weak radar echoes,while it can not improve the resolution of range and angle dimensions.The sparse signal recovery technology is not sensitive to the correlation between sources,and can break the limitation of resolution to achieve the accurate estimation of targets within one beam.However,these methods based on sparse signal recovery technology also require high SNR and are sensitive to the model of the estimated signal.A large deviation between the form of the signal to be estimated and that of preset signal will lead to the degradation of the performance of these methods.To this end,the problem of radar detection and parameter estimation of weak group target is studied in this dissertation to develop some robust,low-complexity and accurate group target detection and parameter estimation methods.The main contributions and innovations of this dissertation are as follows:1.Aiming at the problem of radar detection and parameter estimation of UAV swarms,a framework for processing radar echo of weak group target is proposed.The characteristics of long-time coherent integration technology and sparse representation technology are considered to promote the strengths of these two technologies and avoid their drawbacks.In the proposed framework,long-time coherent processing technology is adopted first to correct the range cells to cause range cell migration(RCM)and Doppler Frequency Migration(DFM)of the target and to achieve long-time energy focusing of weak target.Meanwhile,the signal form changes caused by RCM and DFM are compensated to meet the requirements of the subsequent sparse representation technology.Then,the super-resolution parameter estimation based on sparse representation technology is used to achieve the super-resolution parameter estimation of the group target,which makes up for the shortcomings of the traditional coherent processing lack of resolution,and also avoids the drawbacks of traditional super-resolution algorithms,such as,these methods require the knowledge of targets number and are sensitive to signal correlation.2.Based on the framework mentioned above,an adaptive group target super-resolution detection method is proposed.This method first uses long-time coherent accumulation processing to focus the energy of the signals of each channel to improve SNR,and tries to distinguish multiple targets in different dimensions as much as possible and distinguishes the target echo signal from clutter.Then,the signal energy is further integrated by conventional beamforming technology,and the constant false alarm rate(CFAR)detection technology is used to detect whether the resolution cell contains a target(targets).If the target is detected,a binary detection with an adaptive threshold is adopted to determine the target state in the current cell,that is,whether it contains a single target or multiple targets.If there is a single target,the monopulse method(with extremely low computational complexity)is used to estimate the target angle;if there are multiple targets,the algorithm based on sparse representation technology is used to estimate the angles of targets.The computational complexity analysis shows the low computational complexity of the proposed method.Extensive simulations and processing results based on measured data demonstrate the effectiveness and practicability of the proposed method.3.Aiming at the problem of grid mismatch in grid-based methods,a frequency-selective gridless angle super-resolution estimation method based on prior information is proposed in this dissertation to achieve super-resolution estimation of the target angle.By the zero-padding Fourier transformation,the maximum extraction of target information is realized and the dimension of the signal to be processed is reduced,which benefits the processing in subsequent steps.Based on the existing frequency-selective atomic l0 norm,a group target angle super-resolution estimation problem is formulated and some specific solutions to this problem are proposed.The computational complexity analysis illustrates the low computational complexities of the solutions and the applicable scenarios based on the properties of various solutions are pointed out.Extensive simulations and processing results based on measured data demonstrate the effectiveness and practicability of the proposed solutions.4.Aiming at the problem of poor range resolution of radar under the small transmission signal bandwidth,a two-dimensional super-resolution accurate estimation method is proposed.For the group target echo signal under the uniform acceleration model,a method named as Dechirp-KT is proposed to eliminate the influence of target RCM and DFM on the energy integration,to realize the effective focusing of the echo energy and to realize the rough estimation of parameters,providing a priori information for the subsequent steps.Based on the frequency-selective two-dimensional atomic l0 norm,the specific problem of two-dimensional super-resolution estimation of the range and angle of the group target is constructed,and the specific solution method,i.e.,frequency-selective reweighted trace minimization method is given.The processing results based on simulation data demonstrate the effectiveness and practicability of the proposed method.