Study On Imaging Algorithms For Bistatic Radar

Bistatic imaging radar, which can be depicted as the imaging radar system with thedifferent locations for transmitter and receiver, takes many advantages over thetraditional monostatic imaging radar system, such as low cost, flexibility, reducedvulnerability, etc.. The radar system is the significant development in the radar imagingfield, and has drawn widespread attention. This thesis considers two issues including thebistatic Synthetic Aperture Radar (SAR) imaging algorithm and bistatic radar imagingalgorithm based on the narrow band continuous wave. The main contributions of thethesis are summarized as the following four parts:1. There two-dimensional(2-D) spectrum algorithms for bistatic SAR are proposed.Firstly, the simplified Fourth-order Extended Exact Transfer Function (EETF4)2-Dspectrum algorithm is considered. The algorithm, first of all, handles the slant rangehistory by using the fourth-order Taylor expansion, then obtains the more precise2-Dspectrum by solving the stationary points equation. Moreover, in order to obtain thefurther precise2-D spectrum, the optimal-square-approach-based2-D spectrumalgorithm is raised. Based on the simplified EETF42-D spectrum algorithm, thisalgorithm minimizes the error accumulation of the slant history expansion in theimaging process to improve the accuracy of the2-D spectrum by exploiting the optimalsquare approach. Furthermore, Micro-increment2-D spectrum algorithm is given. Onthe basis of the known low-order stationary phase point, this method indirectly gets theapproximate solution of the high-order stationary phase point by considering themicro-increment between the low-order and high-order stationary phase point, and thenthe2-D spectrum can be obtained. In fact, the three-order stationary phase point isknown, and it can be selected as the starting point to acquire the higher-order stationaryphase point. It is noted that the biggest characteristic of the micro-increment isexpandability, and so it can improve the accuracy of the2-D spectrum by raising theorder of stationary phase point to be solved.2. Based on the2-D spectrum algorithms mentioned above, the bistatic SAR dataprocessing algorithm is investigated. Due to that the result obtained by exploiting thealgorithms mentioned above is rather complicated and hard to be decomposed, while theOMEGA-K algorithm only needs the range variable which is separated between thephase in2-D spectrum and the other parts rather than makes the careful decomposition in the whole spectrum, and the OMEGA-K algorithm is exploited in the imagingprocess. The key process in the OMEGA-K algorithm is the stolt interpolation by usingthe range frequency mapping function, and employs the angle invariance hypothesis inthe process to acquire the range frequency mapping function. In order to guarantee thefocusing quality of the algorithm, the size of the invariance region is analyzed.3. The modified bistatic OMEGA-K algorithm based on MSR2-D spectrum isgiven. Based on the mentioned-above OMEGA-K algorithm, the higher-ordercancellation operation of slant range history is introduced to improve the focusingquality. By using this operation, the effect of the residual higher-order term of the slantrange history on the imaging result can be reduced. Compared to the2-D spectrummatches the echoes in one way in the traditional OMEGA-K algorithm, they match intwo way in the modified OMEGA-K algorithm. The simulation shows that the methodabove implements simply, has low computational complexity, and improves thefocusing effect significantly. In the case of illustrating the slant range history as thesecond-order Taylor expansion, the focusing effect of the modified OMEGA-Kalgorithm overmatches the ordinary methods, while approximates that of the ordinaryOMEGA-K algorithm based on the third-order Taylor expansion very closely.4. Based on the present bistatic radar frequency domain imaging algorithm in thecase of narrow band continuous wave, focusing on the new imaging model, a modifiedfrequency domain algorithm and a time domain algorithm are proposed, which arebased on the after-motion-compensation turntable model. Due to the echoes onlydistributes in one circle in the frequency domain, and then the modified frequency-domain algorithm handles the imaging process in the polar coordinate format, while thetime-domain considers it in the rectangular system. Although they are processed in thedifferent domain and coordinate system, the unit impulse response is identicalessentially, and the same in the resolution and the peak sidelobe ratio (PSLR) because ofexploiting the same imaging model and the signal form. The measurements of the effectcapacity of the two algorithms are obtained by analyzing the unit impulse response ofthe two algorithms carefully.