Study On Micro-motion Characteristic And Recognition Of Space Cone-shaped Targets

Since the1950s, the space cone target recognition has been a hot topic of theradar automatic target recognition (RATR) study. And the recognition abilityrepresents the detection performance of the defense system to a great extent. With thedevelopment of the radar technology and advanced signal processing technology, themicro-motion and micro-Doppler are the focuses in current researches and haveshown enormous potential in the domain of target detection, resolution andrecognition. During a flight in the exoatmosphere, the space cone targets havedifferent micro-motion, such as a spin, a precession, a wobble and a nutation.Making research on micro-motion characteristic and micro-motion-based recognitionof space cone targets is of great significance for the development of the new-typeradar and for the promotion of the defense capability.By considering the engineering background of space cone targets recognition,this dissertation gives our researches on theories and techniques of RATR, mainlyfocusing on the following three aspects, i.e., signal preprocessing, feature analysisand extraction and narrow band imaging, which are supported by Advanced DefenseResearch Programs of China. The main work of this dissertation is as follows:1．The principles of the micro-motion and micro-Doppler are introduced. Firstly,the typical micro-motion induced by space cone targets is analyzed. Then theequivalent scattering point model and the fixed scattering point model are established.At last, the influence of occlusion effect on micro-Doppler is studied.2．The signal preprocessing is researched. Firstly, the influence of the highspeed and acceleration on the echo envelope, phase and micro-Doppler is analyzed.Secondly, it is pointed out that keystone algorithm or range alignment method can beused to correct distance migration without changing the micro-Doppler characteristic.Thirdly, a new method is proposed to estimate and compensate acceleration. Basedon the minimum entropy criterion, the estimation accuracy of acceleration isoptimized by picking up the acceleration in a certain range.3．A novel method is proposed to implement the multi-targets resolution basedon the micro-motion characteristic. The essence and the limitation of traditionalmulti-targets resolution algorithms are analyzed. Then the analysis shows that themicro-motion offers a new way to resolve targets with the same velocity and the same range. The proposed method combines the B-Distribution(BD) and the Viterbialgorithm to extract the micro-Doppler of the multi-targets, and then the DFT isapplied to extract the micro-motion parameters.4．Two different methods based on the difference of micro-motion are proposedto recognize the space cone targets. When a cone-shaped target is flying inexoatmosphere, the micro-motion of the target belongs to a precession and that of thedecoy belongs to a spin or a wobble. Based on the difference of the target and thedecoy in the micro-motion, the micro-Doppler signature of the target and the decoy isanalyzed in detail. The analysis shows that the micro-Doppler Spectra of precessiontargets，wobble targets and spin targets is approximated as a line-Spectrum, but thereare significant differences. Based on the difference, the recognition feature isextracted from the micro-Doppler Spectrum and the eigenvalue spectrum.In micro-Doppler domain, the entropy and standard deviation are extracted asthe recognition feature. Based on the harmonic model, eigenvalue decomposition isapplied to compute the eigenvalue as the recognition feature. At last, the influence ofthe pulse repetition frequency (PRF) and the accumulating time on recognition rate isanalyzed.5．Narrow band imaging and micro-motion parameters estimation of space coneprecession targets are studied. Firstly, according to the micro-Doppler characteristicof a precession target, a method of matching filter is applied. And the result showsthat it is a simple and efficient algorithm in general applications. Secondly, a newmethod is proposed to estimate micro-motion parameters of space precession targets.The proposed method is as follows: firstly, based on the micro-Doppler characteristicof a precession target, the Fourier transformation is applied to estimate precessionparameters coarsely; secondly, based on the time-frequency spectrum characteristicof the echo signal, coherent accumulation is applied to obtain a3-D high resolutiondistribution of the target. The estimation accuracy of the parameters is refined bycalculating the whole image entropy.