AbstractThe raw radar data shows that the aIgorithms of enveloping aIignment and imaging tolSAR avallabIe nO suit to the maneuveri11g t3rgets. Some ai8orithms ofenveloPin8 atiWatand imaglng to maneuvering targets are proposed and verified by raw radaI data.The motion compensation of ISAN is considered fir8t in the dissertation. The enveIoPecorrelation algorithm is widely used in ranSe alignment of lSAR motion compensatio4however, this method often make the error accumu1ation, and what is more, ajump error wilImake when the refieCtion of target has strong scintillation. To reduce the error of mpealignment a ndnimum entroPy criterion of range alignment in ISax compenotion isproPOsed. USing this method the range bin can be alignment correctly which i8 verified byreal radar data.Range-Doppler imaging is the basic method of inverse synthetic aperture radar, it is basedon the unifOrm rotation of targets. The non-unifOrtn rotation may occur to manoringtargets, which will produce smeared images. ln Chapter 3 a constant acceleratively rotatingmode1 i8 used instead of the traditional unifOrm rotating model. The generalized marginalintegration along straight lines with varying slopes is calculated over the time-frequencyplane. The imaging quality is improved by using the new method.The echo Doppler of each scatter in a maneuvering target is a time variable. Moreover,the target imaging prOjection plane may also change with time. Thus the image wiII besmeared by traditional FFT based method. The commonIy-used rotating imaging algOrithmsare nOt avaiIable to the target with rotating direction changed. In Chapter 4, the conceptionof range-inStantaneous Doppler imaging is presented. Our simulation show that the novelimaging algorithm is effective to the maneuvering targets even when tbe rotating direction istime-varying. The method is verified by reaI radar data.Although range-instantaneous Doppler imaging algoritIun to maneuvering targats isgiven in Chapter 4, the calculation is very large, because it need to estimate the phase andamplitude of each scatterer's subecho. And what is more, the above algoritIun bnd on theassumption that the amplitude of each scatterer is the same during the interval of imaging. However, the amplitude of some scatterers will change with time because of the angular change of radar beam, etc., the above algorithm need to be improved. The change of each scatterer抯 Doppler is often first-order, because tli~ inertia of target. In Chapter 5, the image algorithm based on the filter pull-back from Radon-Wigner field is proposed to ianeuvering targets with approximate first-order Doppler change of each scatterer. The proposed algorithm is verified by raw radar data, which calculation is reduced.Although the Doppler of each scatterer generally fit a low order polynomial during the viewing time for imaging, the complex variation occur sometimes. In order to reduce the computational cost of imaging to maneuvering targets with complex Doppler variation of scatterers, the first-order approximate to the Doppler frequency variation of each scatterer subecho is used in Chapter 6 . By using an adaptive chirplet decomposition, the subecho of each scatterer is parted into several linear frequency modulation signals in different viewing time interval. Based on the above decomposition a imaging algorithm is presented. The proposed algorithm is verified by real data.Chapter 7 is the summary of the dissertation. It also discusses some important research areas in the future.