Research On Underdetermined Blind Source Separation Method Based On Air Fretting Group Targets

At present,radar countermeasures are more and more abundant,and the situation of spatial dynamic group target recognition is more and more serious.How to identify the real target information from radar echo is an urgent problem to be solved.Blind source separation(BSS)is an effective method to solve the problem of the processing method of aliasing signals received by radar.It mixes several unknown source signals through some unknown method,and then uses less information to separate these several source signals from the mixed signals.Underdetermined blind source separation is a ill-posed problem,which can be regarded as an inhomogeneous system of linear equations with fewer equations than the number of unknowns.The solution of this problem is generally not unique,which is more difficult to solve,but also more practical.At present,the solution of underdetermined blind source separation is mainly the sparse component analysis method,which requires the signal to be a sparse signal and can recover the source signal after the mixed matrix is estimated.In this paper,the object of study is the target of airborne microdynamic group,and the separation method of the targets of the microdynamic group is introduced through the following aspects:The first part mainly introduces the research background and significance of underdetermined blind source separation,expounds the current research status,and then introduces the model of air micro-motion target,introduces the model and framework of underdetermined blind source separation,and conducts in-depth research on time-frequency analysis.Finally introduced the separation performance standard.The second part mainly studies the underdetermined blind source separation method of ballistic group targets.Firstly,the estimation method of the mixing matrix in the underdetermined blind source separation algorithm is introduced.The sparseness of the signal is improved by time-frequency transformation.The concept of directional amplitude ratio is proposed to process the signal.The mixing matrix is estimated by the clustering algorithm,and the process of the method is introduced and its limitations are analyzed.Then the difference between single source point and multi-source point is introduced.Single source point is selected according to the difference between the real and imaginary parts of the mixed signal time-frequency point,and the estimation of the mixed matrix is realized by density peak clustering.Compared with the previous method,this method has no limitation on the observation channel,so it is suitable for more different scenarios.On the basis of the estimation of the mixing matrix,the model of compressed sensing theory is analyzed,and the similarity between it and the model of underdetermined blind source separation is found.The signal reconstruction in compressed sensing theory is used to realize the recovery of source signal in underdetermined blind source separation.The third part studies the separation method of micro-motion UAV group targets.According to the characteristics of the different rotation speed of the rotor during the flight of the UAV,that is,the different period,the period of each target in the mixed signal can be estimated.Then,according to the difference of the period,each source signal can be separated by the method of segmental superposition of the mixed signal,and the separation of the UAV under different forms and vibration conditions is analyzed.