Research On The Space Time Adaptive Processing For Unmanned Airborne Early Warning Radar

Owing to its excellent characteristics such as mobility,strong survival ability,wide coverage area and no casualties,unmanned airborne early warning radar which can track the enemy target and command our own operations have played an important role in the field of future wars.Nowadays,unmanned airborne early warning radar is developing to be the high-altitude,long-endurance,near-space,stealth,and etc.With the further development of the theory on unmanned airborne early warning radar system,we believe that the application scope and function of unmanned airborne early warning radar system will expand continuously in modern wars.The progress of unmanned airborne early warning radar system also brings various challenges to the space-time adaptive processing(STAP)technology.Therefore,this requires that the STAP technology of unmanned airborne early warning radar system with higher performance and specifications is designed in clutter suppression and moving target detection(MTD).This dissertation regards the STAP method for unmanned airborne early warning radar as research interests.To be specific,the research content consists of Toeplitz covariance matrix reconstruction,clutter covariance matrix(CCM)filling method under space-time coprime sampling structure,second-order super-nested sampling structure STAP method,samples selection,and etc.The main contributions are summarized as follows:1.The STAP method for the reconstruction of Toeplitz covariance matrix is proposed,which can solve the CCM recovery with sparse linear structure.Specifically,using the CCM Toeplitz structure and noise prior information,the solution expression for CCM can be derived,which can give a efficient approximate solution.The proposed method can not only effectively gain high-resoltion beam pattern,but also improve target estimation accuracy and detection performance.2.The STAP method with coprime sampling structure in dual-frequency operation mode is presented.In simple terms,the proposed method can obtain the different coarray and copulse by choosing an appropriate additional operation frequency,thereby filling two missed virtual sensors and pulses of the difference structures in the single-frequency operation,which improves the degree of freedom(DOF)of the system filter.Moreover,the proposed method can promote the resolution of spatial and time.3.The STAP method with the second-order super-nested sampling structure is proposed,which can solve the space steering vector distortion caused by the the mutual coupling among the sensors and improve the system DOF and target detection capability.This method applies difference operation to construct virtual space-time snapshots in the second-order super-nested sampling structure.By the space-time smoothing technology,the clutter plus noise covariance matrix(CNCM)can be established,and the optimal virtual STAP weight can be obtained.4.The knowledge aided STAP method of sample selection based on the Gaussian kernel similarity is proposed.By calculating the Gaussian kernel similar degree between samples and clutter,the interference parts in the samples are filtered and eliminated with prior knowledge that the actual clutter components distribute on the clutter ridge.This system enhances the utilization of samples and the moving target detection ability.Especially in small samples size,the performance of the proposed method is more prominent.The methods mentioned above have been validated by both theoretical derivation and simulation experiments.The experiment results demonstrate that these methods are able to effectively solve some problems in STAP about CCM estimation,mutual coupling among the sensors and sample selection,thus improving the target detection capability.