Research On Heterogeneous Clutter Suppression For Airborne Digital Array Radar

Airborne radar is an important sensor for obtaining battlefield information, which plays a significant role in the war. The clutter echoes received by airborne radar are often very strong and bury the weak signals returned from interested targets, and thus dramatically degrade the detection performance of airborne radar.The space time adaptive processing (STAP) technique can effectively suppress the clutter and evidently improve the radar detection performance since it can adaptively combine the temporal and spatial samples, and thus has been of great interest. The clutter covariance matrix required by statistical STAP methods is commonly estimated from a number of independent and identically distributed (â…¡D) samples. However, in practice, the real clutter is often heterogeneous, which may be induced by many factors, such as fast variation of radar scenarios, discrete clutter sources, the non-sidelooking or bistatic configuration, etc., and the clutter suppression performance of STAP methods is thus degraded significantly. Focusing on the heterogeneous clutter suppression, some improved STAP algorithms are developed in this dissertation. The main contents of this dissertation can be summarized as follows:1. The characteristics of clutter received by non-sidelooking airborne radar are investigated, and the causes inducing the heterogeneity of clutter are analyzed. We find that for airborne radar with medium pulse repetition frequency, the near-range clutter is the main reason for clutter heterogeneity, whereas the far-range clutter can be considered as homogeneity. Based on this fact, a cascaded STAP method that is applicable to non-sidelooking radar is developed. The near-range clutter is firstly extracted from the original echoes by using oblique projection technique and is then mitigated by using a robustly adaptive beamformer, such that the homogeneity of clutter is improved greatly. After that, an azimuth STAP algorithm is employed to suppress the residual far-range clutter to further increase the signal-to-clutter-plus-noise ratio (SCNR). Simulation results show that the proposed method can improve the detection performance significantly.2. The application of the geography knowledge in STAP is studied. Based on the fact that the statistics of clutter is strong correlation with the terrain of the radar scenarios, a knowledge-aide STAP algorithm is proposed. With the aid of a prior knowledge on the terrain, the representative samples are selected as the training data to estimate the clutter covariance matrix. Then, a spectrum-based covariance matrix is constructed by using the clutter spectrum estimated from the cell under test (CUT). Eventually, the two covariance matrix is incorporated to form the final clutter matrix required by STAP algorithm. The proposed method can not only improves the homogeneities of the training samples, but also reduces the requirement on the number of training samples, and results in favorable performance improvements.3. The problem of estimating the covariance matrix of nonhomogeneous clutter is studied. When the clutter follows the Gaussian distribution, the covariance matrix can be formulated as the sample covariance matrix (SCM). However, if the clutter is seriously heterogeneous, the Gaussian distribution assumption would be violated, and the SCM estimation may lead to significant errors. Here, two covariance matrix estimation methods are developed. By making use of the compound Gaussian clutter model, a weighting covariance matrix is constructed, which can effectively eliminate the clutter fluctuation, and thus is constant false alarm rate (CFAR) with respect to clutter power. This method is applicable to partially homogeneous clutter environment. The second method is developed to estimate the covariance matrix of arbitrarily heterogeneous clutter. The samples whose clutter spectrum is evidently different from that of the CUT are censored from the training data to enhance the homogeneity, and then using the updating training sets to estimate the clutter covariance matrix. The experiments on simulation and measured data demonstrate that the proposed methods can obtain excellent performances.4. Adaptive coherence estimator (ACE) often suffers considerable performance degradation in the presence of steering vector errors. To address this problem, a robust ACE detector based on the ellipsoid uncertainty set constraint is proposed. A detailed analysis of ACE detector is first conducted, which results in an interesting observation that scaling of the steering vector does not affect the statistical test of ACE. With this property exploited, a model for designing robust ACE detector is constructed and is subsequently converted into a convex optimization problem. Then, the solution to the problem is given with the powerful Newton-Raphson method. Simulation results show that the robustness of the ACE detector against the steering vector errors can be improved significantly.