Research On Imaging Algorithm For Mono-and Bistatic Synthetic Aperture Radar

Synthetic aperture radar (SAR) is capable of greatly improving radar’s informationacquirability to obtain high-resolution microwave images of the observed scene,regardless of meteorological conditions and solar illumination; hence it has great valuein both civilian and military applications.In recent years, widely used in various platforms, such as plane, missile, andsatellite etc., there are increasing number of problems required to be solved in specialSAR imaging systems. Firstly, after the linear reange cell migration correction (LRCMC)in highly-squinted configuration, the invariant regions, in which the data can bewell-focused, shrunk considerably, refer to as focus-depth problem. Secondly, how tofocus forward-looking data is a crucial problem in air-and missile-borne SAR systems.Thirdly, in terms of atmospheric turbulence and aircraft maneuvers, the moving sensorcannot flight along a straight track with a constant velocity, which will result in imagingdegradation. Furthermore, the SAR system on-board hypersonic vehicle or on-boardseparately platform, reffered to as Bistatic SAR (BiSAR), has new problems needed tobe solved. Therefore, research on SAR imaging algorithms in special cases is of greatimportance.Here we start our work with mono-and bistatic SAR geometry analysis and signalprocessing. The contents include highly-squinted imaging, motion compensation(MoCo), imaging analysis in the case on-board hypersonic vehicle, BiSAR imagingalgorithm in general case (GC), which are proved through computer simulations. Thekey innovations in this thesis are following.An imaging algorithm in highly-squinted case based on azimuth nonlinear chirpscaling (NLCS) is proposed. Linear range cell migration correction (LRCMC) can beused to accommodate the spectrum overlapping problem, which needs high PRF or/andextra computational load in this case. However, the LRCMC introduces azimuth-variantproblem, or referred to as focus-depth problem, which shrinks the imaging regions inazimuth direction significantly. This algorithm derives an extended azimuth nonlinearchirp scaling (ENLCS) algorithm to accommodate the problem based on a newprojection of the data from acquisition sub-geometry to imaging one. Compared withoriginal NLCS, the proposed ENLCS is capable of compensation of the azimuth-variantcomponents with neglectable residual phase.Modified range Doppler algorithm (RDA) is proposed to focus forward targets in curvilinear track using the method of series reversion (MSR). To obtain theforward-looking image, the algorithm uses a curvilinear track instead of ideal track totransform forward-looking mode into highly-squinted case. The special track iscompensated in2-dimension spectrum based on MSR, and then the data are focused bythe modified RDA. Simulated results are utilized to validate the effectiveness of theproposed algorithm.Modified chirp scaling algorithm (CSA) is proposed to image the forward targets incurvilinear track. In the new configuration, LRCMC is used to remove highly-squintedDoppler center; then a power series slant range is used to express the curvilinear trackand the method of series reversion (MSR) is adopted to derive precise2-D spectrum.Based on the spectrum, the coefficients used for algorithm are extracted and modifiedCSA ultilized to process the data. The simulations are used to illustrate theeffectiveness.Azimuth two-step motion compensation (MoCo) is proposed based on modifiednonlinear chirp scaling (MNLCS). The curvilinear track caused by aircraft maneuversand the motion errors caused by atmospheric turbulence can be seen as the same one, anideal track with motion errors. We give the instantaneous slant range by vectorexpression, and analyze that the traditional Two-step MoCo strategy just compensatesthe second order motion errors. In the squint mode, however, high order motion errorsalong range direction degrade the imaging results; on the other hand, theazimuth-variant problem caused by the varying magnitude and direction of sensorvelocity vector shrinks the invariant region. Analogous to Two-step procedure, theproposed algorithm compensates the motion errors by two steps. In azimuth timedomain, the azimuth-variant components can be eliminated by modified NLCSalgorithm and the high order components are processed in the azimuth spectrum.Simulated results show that the imaging quality is improved considerably.A SAR imaging algorithm is proposed to focus to data acquired from the systemon-board hypersonic vehicle. With the increasing of velocity, there are two intolerablemotion components of the sensor in original SAR systems should be taken into account,the sensor motion by transmitting signal and the other one between transmitting andreceiving. The former will degrade imaging results with the increasing of pulse durationand/or squint angle; the latter is required to be compensated in high velocity case. Inthis chapter, both the motion components are expressed by modified instantaneous slantrange. The former is extracted by introducing fast time into slant range expression;whereas the latter is covered in equi-slant range by calculating transmitting and receiving slant range respectively. In the imaging procedures of extended frequencyscaling algorithm (EFSA), the motion components are compensated step-by-step.Simulated results are used to illustrate the effectiveness of the proposed analysis.An imaging algorithm is proposed to accommodate the azimuth-variant bistaticdata in general case (GC). In GC, there is azimuth-variant problem caused by theconfiguration. The spectrum derived by MSR can just process the data acquired in asmall region centered on a reference target; whereas the one far away from the referencewill suffer more degradations associated with the increasing of imaging region. Thischapter extracts the azimuth-variant components by vector expression of instantaneousslant range. Then based on the principle of nonlinear chirp scaling, a new perturbation isderived to compensate these azimuth-variant components. A spaceborne/airborne and anairborne/airborne configuration are simulated to illustrate the effectiveness of theperformance.