Study On Statistical Characteristics And Processing Of Instantaneous Polarization

In modern warfare, the electromagnetic (EM) environments have been more and more complicated and adverse. In order to exploit and utilize the EM information utmost, that is, to let the radar sensing system can adapt adverse battleground environments with intelligent detection and recognition capability, some problems are becoming a groundwork and impendent mission in radar polarization information processing. These problems mainly include studying the polarization phenomena of time-varying EM waves from dynamic and statistical viewpoints, revealing the inherent rules of random EM waves, exploring polarization scattering characteristics of radar targets and environments, and founding effectual descriptive means of polarization statistics of radar targets.The dissertation can be divided into two sections. Firstly, the instantaneous polarization(InPol) statistical theory is founded, which is based on InPol theory. From the statistical viewpoint, polarization characteristics of time-varying EM waves and radar target scattering are studied. Thereby, using the theory above, target detection, multi-decoy discrimination, target recognition and other applications have been studied according to polarimetric radar systems and aerial targets (e.g. missiles). The methods and conclusions in this thesis can proffer some novel and promising techniques for weakening the influences of adverse EM environments, confronting active jamming, detecting concealed targets and recognizing targets and other applications.The main contributions of the thesis are as follows:In theory, the radar InPol statistical theory has been founded and InPol theory also has been developed deeply. The main contents include:1) The polarization characteristics of random EM waves are studied from dynamic and statistical viewpoints. Firstly, polarization similar degree, polarization fluctuant degree(PFD) and polarization sampling sequences and other conceptions are presented in this thesis. Secondly, the influences of orthogonal polarization basis to InPol descriptions have been discussed, and the properties of instantaneous Stokes vector or InPol projection vector (IPPV) and other descriptions have been explained. Thirdly, the probability density functions(PDFs) and joint PDFs of amplitude, phase, amplitude ratio, difference of phase, instantaneous Stokes vector and IPPV of random EM waves have been derived in the condition of Gaussian hypothesis, and one moment, variance, compounded moment characteristics of these parameters are also presented. Fourthly, the relation between moment characteristics of IPPV of random EM waves and its polarization parameters (such as degree of polarization) are analyzed. At last, statistical distributions and numerical characteristics of polarization sampling sequences are also analyzed.2) The description problem of time-varying polarization scattering of radar target have been respectively studied from determinate and statistical viewpoints. Firstly, the characteristic descriptions of instantaneous polarization scattering of radar target are derived. Secondly, InPol time-frequency distributions of radar target are presented and the properties of InPol Wigner-Ville distribution or general time-frequency distribution of radar target are also derived. Thirdly, statistical descriptions of instantaneous polarization scattering matrix of radar target have been presented and statistical properties of scattered wave of radar target inspirited by typical electromagnetic wave are given.These conceptions and conclusions proffer theoretical foundations and avail tools for polarization characteristic analyzing of random EM waves or radar target scattering, feature extraction and other applications. And these are possessed of practical significance to engineering applications.In application, the InPol statistical properties of radar targets and electro-jamming are presented, and their applications in target detection, discrimination of active-decoy and radar target, classification and recognition of targets in modern warfare are also studied. The main achievements include:1) Statistical and power properties of radar target scattering signal and receiver noise are analyzed respectively based on narrowband polarimetric radar systems. Then, their different statistical properties in IPPV series or instantaneous sub-Stokes series are analyzed and two kinds of detection problems are studied when signal polarization is known or not. Two novel and promising detection algorithms of dim targets are presented by polarization accumulating, which are based on IPVS and PFD respectively. Theories and simulation results show that it can improve the detection performance greatly by optimizing the receiver's bandwidth and transmit pulse width.2) Aiming at coherent pulse radar systems, utilizing the correlation difference between radar targets and receiver noise, a novel dim target detection algorithm which can extract targets' motion information has been presented by "polarization accumulation". Theoretical results and computer simulations show that the detection performance is greatly improved compared to the conventional detector.3) The discrimination problem of radar targets and any elliptic polarized active-decoys is studied. Firstly, polarization scattering characteristics of radar targets and active-decoys are analyzed. Aiming at different polarimetric measuring systems (such as narrowband full-polarimetric radar, narrowband timesharing polarization measurement radar and broadband timesharing polarization measurement radar), several polarimetric discriminators are designed. Then, the usage of these polarization discriminators are analyzed too. And computer simulations indicate that polarimetric radars have potential to distinguish active multi-decoy from radar targets.4) The problem of feature-extraction and recognition of ballistic missile targets is studied in this thesis. Firstly, based on instantaneous polarization scattering matrix of radar targets, the descriptions of its polarization scattering characteristics are presented by theory of matrices. Then, the first and second order polarization measures of radar targets are extracted with measured data of warhead targets, which are insensitive with radar targets' pose. Finally, missile targets classification and identification problem are studied through nearest neighbor recognition methods and satisfied recognition results are achieved.