Research On Micro-Doppler Phenomenon And CLEAN Technology In ISAR Imaging

Inverse synthetic aperture radar (ISAR) is one kind of radar with high resolution in both two dimensions. There are two factors in the target interfering its ISAR image. One is micro-Doppler phenomenon, which is induced by the micro-motion structure of the target. This will induce additional modulation to the returns. Micro-Doppler Effect will lead to contaminations to the target's ISAR image. The other is high sidelobes induced by strong scatters, which will submerge weak scatters and become false targets. Based on turntable imaging model, micro-Doppler signatures in ISAR imaging, influence on the imaging process and suppression of micro-motion interference and CLEAN Technology are studied respectively in this paper.Firstly, based on the motion model, a point–scatter target with micro-motions is set up. The returns by the micro-motion scatter are deduced, then the signature of those returns and the influence of micro-motions on ISAR imaging is analyzed and simulated respectively.Secondly, micro-Doppler effect suppression methods are studied based on the difference of micro-Motion and bulk returns. Adaptive Gauss Short-time Fractional Fourier Transform is performed to decompose the short-time returns into a summation of linear frequency modulation signals, which are divided into two parts: bulk returns and micro-motion returns. Finally the bulk returns is compressed to obtain target image. In order to improve the speed of traversal search, two step searching is firstly analyzed and then random searching based on genetic algorithm is present. The validity of both methods is verified by simulation and measured data. Finally, the sidelobes in ISAR imaging are analyzed based on the point spread function. And then coherent CLEAN technology and sequence CLEAN technology are introduced. A detailed comparison is carried out between them. Both of technologies could reduce the sidelobes, and sequence CLEAN can reduce false alarm rate and is robust to noise. Limited by high time-space complexity of sequence CLEAN, optional path searching based on GA is proposed. The validity of both methods is verified by simulation and measured data.