Study On Polarization Space Time Adaptive Processing For Airborne Radar

The clutter suppression and detection of the weak target are serious problem forairborne radars worked in the down state. Space-time adaptive processing (STAP)technology may be adopted by a new generation of airborne phased array radar.However, the clutter canceller is a problem for STAP when the target is slowly ortangentially moving, because the target and clutter in the angle domain and Dopplerdomain are similar or identical. At this point, the difference of the polarization domaintogether with STAP can be used to suppress clutter, enhance target signal and improvethe weak target detection in strong clutter background.The paper studies tightly around the clutter suppression and weak target detection,divided into two levels with theory and application and three aspects with radar echomodel, fully adaptive polarization-space-time processing and reduced rank processing.The study on the radar parameters, target parameters and environmental parameterseffects on the performance of polarization-space-time adaptive processing (PSTAP).Then, it is used to guide engineer practice with adjustment of the radar parameters toachieve performance index quickly in the case of given radar performance metric.Specific research and results of thesis are as follows:1. Echo characteristics modeling and analysis for airborne polarization array radarsFirstly, the antenna beamformer model is established according to the airbornepolarized array radar on the ground/sea space geometric relation. The result is that usingpolarized array can not only control the beam direction, but also can control the antennapolarization of electromagnetic waves. Secondly, the author establishes the polarizationecho model of the received signal: the target, clutter and noise model. Then, two kindsof clutter secondary data models are established in the case of the lack of measured data.Finally, the joint distribution of the clutter characteristic features in polarization, spaceand time domain is studied. The clutter may be equivalent to some non-coherentinterference using resolution grid method.2. Polarization-space-time joint filter and detectionPolarization-space-time joint filter:(1) Using clutter equivalent non-coherentinterference with resolution grid idea and the partly polarized wave into completelypolarized wave analysis method, the output signal-to-interfere-plus-noise (SINR)mathematical analytical expression of the optimum polarization-space-time joint filter isderived, which reveals the mechanism that the radar parameters and target parameters,and clutter parameters effects on the airborne polarized array radar adaptive polarizationspace-time processing performance.(2) A polarization-space-time sample matrixinverse (SMI) algorithm is proposed in the case of unknown clutter covariance andknown polarization-space-time steering vector. Statistical distribution of output SINR is derived.(3) The author proposes the minimum variance unbiased estimator and theorthogonal projection estimator to estimate the target polarization vector in the case ofunknown target polarization, which is not using any filter bank to cover the polarizationdomain, greatly reducing the amount of computation. Also, the filter after estimation ofPSTAP outperforms the STAP.Polarizatio-space-time joint detection:(1) An optimal polarization-space-timedetection algorithm is proposed where the clutter covariance matrix (CCM) andpolarization space time steering vector are known. The author deduces the intrinsicrelation between optimal filter and optimal detection.(2) Polarization-space-timeadaptive matched filter detection algorithm with specified steering vector (SPST-AMF)is proposed. The author deduces the test statistical, the probability of detection and theprobability of false alarm of SPST-AMF. At the same time, the author proves that itensures CFAR property with respect to the covariance matrix of noise.(3) A novelpolarization-space-time adaptive matched filter (PST-AMF) detector is proposed withunknown target polarization steering vector and clutter covariance matrix. Moreremarkably, it has the same performance, but a lower complexity, than thecorresponding polarization-space-time generalized likelihood ratio test (PST-GLRT).(4)The unified framework and internal relation between the polarization-space-timeadaptive filtering and adaptive detection are studied systematically. The difference andthe relation based on adaptive matched filter (AMF) detector and based on thegeneralized likelihood ratio test (GLRT) detector are comparatively analysed. Thedetection probability and the false alarm probability expression can be used topolarimetric spatial domain joint detection, space-time adaptive CFAR detection,polarization-space-time adaptive CFAR detection and polarimetric SAR detection inGaussian clutter background.3. Polarization-space-time reduced-rank processing (PST-RRP)The author analyses polarization-space-time reduced-rank processing methods,aiming to the practical problems of large amount of calculation, independent andidentically distributed (IID) secondary data limited. Firstly, Unified framework on filterand detection of PST-RRP is established. Depending on the difference of the structureof the reduced rank processing method, the unified models based on the direct form(DFP) and based on the generalized sidelobe cancellation structure (GSC) reduced rankprocessing are derived. Secondly, six PST-RRP approaches (cross spectral metric,principle component cross spectral metric, SINR metric, small eigenvalue, modifiedcross spectral metric and principle component with direct form processing) based ondata characteristics decomposition and seven PST-RRP approaches(Doppler-polarization-space Cascade, beam-polarizaton-time Cascade andpolarization-space-time multiple beamformer, et al.) based on discrete Fourier transformdomains are proposed. Finally, the optimal reduced-rank methods and optimal reduced-rank order selection metric are established in the case of auxiliary training datalimited.