Key Techniques On GMTI For High-speed Platform Radar

Ground moving target indication (GMTI) is an important capability of the modern airborne/spaceborne radar in civilian and military areas. It has attracted the attention of researchers all over the world for practical applications such as ground moving targets imaging, positioning and motion parameter estimation. This thesis investigates the challenging problem in the filed of moving target detection, parameter estimation and target imaging. The main work can be summarized as follows:1. The accurate estimation of Doppler centroid is essential for fine image quality since it influences the range cell migration, the azimuth compression and the image registration, etc. For a multi-channel SAR together with GMTI system, the imprecise or wrong Doppler centroid estimate results in SAR images with relatively low correlation and affects the image coregistration. Therefore, clutter rejection algorithms cannot be well performed, which will degrade the GMTI performance. In Chapter two, two different approaches are addressed for Doppler ambiguity resolution. The first method exploits the fact that, in compressed range time and special scaled azimuth time domain, all the trajectories exhibit the same slope which is proportional to the Doppler ambiguity number. The method can directly estimate the ambiguity number without a prior knowledge on the baseband Doppler centroid. The second method adopts the fact that the absolute Doppler centroid is a linear function of range frequency. Based on this concept, an alternative azimuth compression method which is without a prior knowledge on the motion parameters is addressed and carried out in range frequency domain to accumulate the target energy along azimuth. The resulting trajectories become straight lines with the same slope proportional to the unwrapped Doppler centroid. Results of real measured data processing show that both methods work well in medium- to high-contrast scenes.2. For the sake of the robustness to GMTI for SAR system in heterogeneous environment, in Chapter three, a new approach using multi-channel and multi-look SAR images is presented for slowly moving target. This approach has been shown to attain improved suppression of monostatic clutter over its counterpart (the multi-channel and single look algorithm). For fast moving target, a robust method is addressed based on subaperture processing. In this method, multilook Doppler beam sharpening (DBS) images with different illuminated angles are formed by dividing the data sequence in azimuth. Compared with tranditional methods, the detection probability with multiple looking angles can be improved under the same condition. The robustness and practicability of the proposed algorithms are verified by processing the measured data of the experiment with three-aperture airborne radar.3. Fast time (range delay time) information has been well exploited for the terrain scattered interference or hot clutter mitigation. In Chapter four, an approach incorporating fast time in space-time adaptive processing (STAP) is introduced for monostatic (cold) clutter rejection. This method takes advantage of the coherence information of adjacent range bins. Compared to traditional STAP algorithms, the performance of monostatic clutter mitigation can be improved due to the fact that the additional range degree of freedom (DoF) can mitigate various deleterious factors in real scenarios which may increase the DoFs of clutter patches. Results of real measured data processing using airborne X-band and MCARM radar systems demonstrate the effectiveness of the proposed processor. When the supported samples are rare and limited, we develop a new approach to STAP which is based on the iterative worst-case optimization for the spatial-temporal separable filter. It is confirmed that this method belongs to the class of colored loading algorithms, where the optimal loading factor can be efficiently calculated via the so-called second-order cone programming based on the known level of the uncertainty of spatial temporal steering vectors. Computer simulations demonstrate that this two-dimensional (2-D) beamformer has attained better performance as compared to the conventional algorithm.4. After clutter cancellation and target detection, we propose an approach for ambiguity-free motion parameter estimation of ground moving targets with arbitrary velocity and constant acceleration from dual-channel airborne SAR systems. For targets with high and low signal-to-clutter-noise ratios (SCNRs), different analytic expressions for along/cross-track velocities and accelerations estimation are derived, respectively. We use the nth-order keystone transform to mitigate the range migration induced by the target's residual motion after the ambiguity removal. The validity of the proposed algorithm is verified by processing both the simulated and measured airborne stripmap SAR data.5. It is well known that, due to the temporal and spatial sampling, conventional slant-range velocity estimation approaches suffer the limitation of pulse repetition frequency (PRF) for single-channel SAR or the antenna spacing for multi-channel SAR systems. Chapter six describes an ambiguity resolving approach for slant-range velocity estimation which utilizes the wideband characteristic of the transmitted signal (multiple wavelengths). On the basis of the wavelength diversity concept, we extract two looks in range frequency domain to form dual-wavelength radar data. Then two effective approaches are introduced to focus the moving target whether the Doppler ambiguity exists or not. The true slant-range velocity can be estimated by the number of azimuth cell displacement between the two focused images. Both two imaging methods have different properties and advantages. Performance analysis is made and deleterious factors in practical applications are analyzed detailedly. The effectiveness of the unambiguous slant-range velocity estimation method is demonstrated by simulated and real measured data processing.6. The final resultant procedure is the target refocusing. As is well known, the motion of a target induces range migration, especially for the high-resolution SAR system. Ground moving target imaging necessitates the correction of the unknown range migration. To finely refocus a moving target, one must accurately obtain the motion parameters for compensating the target trajectory. However, in practice, these parameters usually cannot be precisely estimated. This thesis proposes a new imaging approach for ground moving targets without a priori knowledge of their motion parameters. In the devised method, the azimuth compression function is constructed in range frequency domain, which can eliminate the coupling effect between range and azimuth. The methodology can be considered as an extension of the keystone formatting technique, which can remove the linear and quadratic components of range migration simultaneously. Theoretical analysis confirms that this method can precisely focus targets without interpolation procedure. The practicability of the proposed imaging technique is demonstrated by real airborne SAR data.