Target Polarimetric Detection And Signal Parameter Estimation Methods Based On Two Dimensional/Three Dimensional Polarization Sensitive Arrays

The polarization information of targets and environments can be obtained by two dimensional / three dimensional polarization sensitive arrays due to the polarization diversity features of array elements,which can be exploited to improve the detection and estimation performance,when detection and estimation are hard to implement by the spatial-Doppler processing.This dissertation has focused on target polarimetric detection and signal parameter estimation methods based on two dimensional / three dimensional polarization sensitive arrays:(1)Under the background of inhomogeneous and / or non-stationary clutters(particularly when the sample support is less than the number of array elements),and in the presence of either Gaussian or compound Gaussian clutters,an iterative shrinkage based regularized polarimetric target detection scheme is presented.The involved shrinkage parameter in the convex combined sample covariance matrix or fixed-point covariance matrix is herein adaptively determined by using a neighboring point matching technique developed in this dissertation.Performance assessment using synthetic and real data shows that the detection performance of the proposed method outperforms the existing detectors.(2)A generalized Rao test method is presented for polarimetric multichannel adaptive signal detection.By introducing several tunable parameters into the Rao test statistic,the proposed polarimetric detector is quite general with the modified Rao test detector,the polarimetric adaptive matched filter detector,and the adaptive subspace detector being its cases.Simulation results show that,in the presence of target steering vector mismatch,the proposed detector can offer more flexibility of being adjustable than the modified Rao test.(3)A robust polarimetric detector is presented to deal with the problem of target steering vector mismatch.Within the framework of the adaptive beamformer orthogonal rejection test,the test statistic of the proposed polarimetric detector involves the estimates of the polarimetric clutter-plus-noise covariance matrix of the secondary data,the target vector after primary data whitening and dual polarimetric channel unitary transformation,and the sidelobe interference vector which is polarimetricly and spatially orthogonal to the target vector after primary data whitening.With the polarimetric and spatial joint subspace constraint,the interference vector is estimated from the Bayesian estimate of the clutter-plus-noise covariance matrix.To obtain a more reliable estimate of the target vector after primary data whitening and unitary transformation,two separate conic uncertainty set constraints are incorporated into a least square fitting scheme.A modified version of the detector is also given for the case of ground clutter.Furthermore,the proposed detector is extended to the case of range-spread targets and full polarimetric channels.Simulations show that,in the presence of several types of target steering vector mismatch,the proposed detector achieves a considerable performance improvement over the traditional polarimetric adaptive detectors in terms of the probability of detection.(4)A method is presented to deal with the problem of polarimetric detection of a point-like target with energy spillover in non-Gaussian clutter.In the proposed method,consecutive range samples are employed to enhance the detection and range estimation performances.Within the framework of the generalized likelihood ratio test,the test statistic of the proposed polarimetric detector involves the estimates of the textures and the covariance structure of the clutter,the target return,and the residual delay.The final test statistic is achieved by using the effective estimates of these parameters.Simulations show that,the proposed detector achieves a considerable performance improvement over the polarimetric adaptive detectors that ignore either the non-Gaussian clutter or the spillover,in terms of both the probability of detection and the root mean squared error of range estimation.(5)A root manifold decoupling method is proposed for the estimation of direction-of-arrival and noncircularity coefficient of rectilinear signals(also known as second order fully noncircular signals,such as AM and BPSK signals)by using a uniform linear array of multiple noncollocated and diversely polarized antennas.In the proposed method,incorporated with eigendecomposition and the noncircular cum polarimetric rank reduction technique,the differences of signals in space,polarization,and noncircularity are jointly exploited in a closed-form widely linear framework,leading to an enhanced parameter estimation accuracy.Simulation results are included to validate the proposed method.(6)A direction-of-arrival and polarization parameter estimation method is presented using an imperfectly calibrated polarization sensitive antenna array.The model errors are calibrated by the proposed method,based on a series of separation of different types of model errors and a subsequent iterative and alternate self-calibration procedure.The proposed method works well in the case of coexistence of channel mismatch,antenna misorientation,and element position error.Simulation results show that the proposed method has a superior performance over the conventional uncalibrated counterpart in terms of incident angle and polarization estimation accuracy.