Cognition-based Heterogeneous Clutter Suppression Method For Airborne Radar

In a heterogeneous environment,due to the lack of sufficient training samples that meet the conditions of independent and identical distribution with the samples to be tested,the performance of conventional space-time adaptive processing(STAP)methods will drop sharply.Therefore,it is particularly necessary to study the STAP method suitable for the actual heterogeneous environment.As one of the important development directions of future radar,cognitive radar can use the a prior knowledge of the environment to intelligently choose the working mode and improve the overall performance of the system.Among them,the knowledge-aided STAP(KA-STAP)using a prior knowledge is the key part of airborne cognitive radar,which is of great significance for improving the clutter suppression performance of airborne radar in a heterogeneous environment.Based on the above background,this paper conducts research on the KA-STAP method of airborne radar in a heterogeneous environment.The main content is as follows:1.The basic principles of the STAP method and two reduced-dimension STAP methods are studied.First,the signal model of the airborne radar is established,and the space-time coupling structure and low-rank characteristics of ground clutter are analyzed.On this basis,the principle of the fully adaptive STAP method,the limitations of the fully adaptive STAP method and the corresponding reduced-dimension STAP method are further studied.Finally,the 1DT and 3DT reduced-dimension STAP methods of the element-Doppler domain are studied.2.Two STAP methods in heterogeneous scenarios are studied.1)Aiming at the heterogeneousity problem caused by the interference target in the training sample,the nonhomogeneity detector method is studied,including the generalized inner product nonhomogeneity detector and the adaptive power residue nonhomogeneity detector.This method solves the heterogeneousity problem caused by the interference target by removing the singular samples contaminated by the interference target in the training samples.2)Aiming at the heterogeneousity problem caused by isolated interference in the cell under test(CUT),the direct data domain method is studied.This method only uses the data of the CUT to construct the STAP filter.Compared with the 3DT method,it has better performance in heterogeneous scenarios such as isolated interference.3.The KA-STAP method based on dynamic sensing of the environment is studied.1)Aiming at the problem of insufficient number of training samples that meet the conditions of independent and identical distribution in a heterogeneous environment,the colored loading in the pre-whitening class KA-STAP and the fast maximum likelihood with assumed clutter covariance algorithm are studied.The mathematical equivalent relationship between the two KA-STAP algorithms and the pre-whitening and rank reduction algorithm is established,and the principle and convergence speed of the two algorithms are analyzed based on this.Then through simulation experiments,the effectiveness of the KA-STAP algorithm is verified when the a prior knowledge is accurate;and the influence of the a prior knowledge mismatch on the two KA-STAP algorithms is further studied.2)Aiming at the problem of how to select the optimal colored loading factor in the colored loading algorithm,two colored loading factor optimization algorithms are studied,namely,the convex combination algorithm and the CUT a priori information involved pre-whitening algorithm.Through simulation experiments in homogeneous and heterogeneous environments,the performance of these two colored loading factor optimization algorithms is studied and compared.3)Aiming at the problem of mismatch between a prior knowledge and the actual environment,the KA-STAP method based on dynamic sensing of the environment is studied.This method can actively transmit signals to perceive the external environment,and use the sparse reconstruction technology to obtain real-time dynamic information of the environment from the echo signal.To a certain extent,it solves the problem of timeliness of static knowledge,and at the same time has a better effect on a prior knowledge mismatch.Strong robustness.