Research On Detection Methods For Radar Targets In Sea Clutter

The detection of moving targets in sea clutter has wide applications and important signifi-cance in military and civil fields,and it faces great difficulties and challenges.High resolution sea clutter generally presents non-Gaussian,non-stationary,and non-linear characteristics,it pos-es a great challenge for moving targets' detection.For the slow moving and stealth targets or floating targets with small radar cross section,the echo signal-to-clutter ratio is relatively low,which dramatically increases the difficulty of detection.In addition,in the case of a small number of the transmitted pulses,the signal-to-clutter ratio is difficult to improve,and detection perfor-mance is greatly restricted.The detection of targets in the background of high resolution sea clutter become a hot issue concerned by the researchers both at home and abroad.To deal with these difficulties,this dissertation studies the detection of moving targets in the background of high resolution sea clutter for the case of short pulse sequences,and explores new theories and methods for radar target detection under conditions of low signal-to-clutter ratio.Employing the correlation of received data,information geometry theory and clutter suppression methods,this dissertation explores several new radar moving target detection methods.The main contributions of this dissertation are as follows:(1)Fully mining and utilizing the information between data samples to improve detection performance for the case of small pulses is a difficult problem to be solved in radar target de-tection.Considering that sea clutter has temporal and spatial correlations,this dissertation uses eigenvalues of covariance matrix to capture the correlations between data and extract features to distinguish the target from clutter.We derive the relationship between the test statistic and eigenvalues from the perspective of likelihood ratio test in the eigenvalue domain and general-ized likelihood ratio.The scheme based on the maximum eigenvalue is designed.Simulation experiments show that the algorithm achieves better detection performance.In addition,the analysis using invariant theory shows that the proposed algorithm can maintain constant false alarm rate(CFAR)property with respect to the clutter covariance matrix.(2)Based on the theory of information geometry,we explore the matrix information geom-etry theory in-depth and design radar target detection methods.For geometric measure and mean matrix estimation on matrix manifold,we firstly exploit the matrix spectral norm based metric as the geometric measure on matrix manifold.The anisotropy of geometric measure is then an-alyzed to show the differences among the proposed geometric measure and existing geometric measures.Secondly,the mean matrix estimation associated with the new geometric measure is studied based on a geometric paradigm that does not rely on statistical characterization of data.The mean matrix estimation problem is transformed into a convex optimization problem on matrix manifold.Two effective methods are developed to deal with the optimization prob-lem of mean matrix estimation.And then two matrix detectors are designed using the proposed geometric measure and mean matrix estimators for target detection in sea clutter.Theoretical analysis and simulation results show that the proposed matrix detectors exhibit a bounded con-stant false alarm rate behavior and achieve performance improvement over the state-of-the-art matrix detectors based on other geometric measures.(3)To solve the problems of unsatisfactory detection performance and high calculation com-plexity of some existing matrix detectors,two new types of detection algorithms are proposed.In the first scheme,we exploit the Root-Euclidean distance,Power-Euclidean distance,Cholesky-Euclidean distance and new Riemannian distance to measure the difference between the two points on matrix manifold,and corresponding mean matrix and median estimation,and then design matrix detection algorithms based on mean and median matrix.In the second scheme,the method of solving median matrix in the existing matrix detectors is retained to maintain the advantages of the original algorithm,while the maximum eigenvalue is used as test statistic,and then six different median matrix detection methods are designed.Simulation experiments based on multiple scenarios show that the proposed algorithms achieve better detection perfor-mance and robustness,in particular,the new algorithm combining the maximum eigenvalue and the Hellinger median matrix obtains superior detection performance.What's more,the com-putational complexity analysis shows that the proposed algorithms have lower computational complexity than the existing geometric methods.(4)As coherent integration is difficult to improve signal-to-clutter ratio under the condition of short pulse,the clutter suppression methods are studied,and two kinds of maximum eigenval-ue matrix detectors using filter processing are proposed.To do this,in the case of small bunch of pulses,two kinds of detection algorithms based on filtering are proposed from the perspective of clutter suppression.In the first scheme,the preprocessing process is firstly realized by perform-ing fast Fourier transform on the received data,and then the maximum eigenvalue matrix detector based on the preprocessing is proposed by combing the frequency domain coherent integration with the maximum eigenvalue detection.This algorithm uses the prior information of target steering vector to weaken the influence of clutter and further improve detection performance.In the second scheme,the idea of subband decomposition is studied.Subband decomposition splits the received signal into multiple subband signals in the Doppler domain.This operation effectively suppresses clutter and increases the coherent integration time.A two-stage cascade algorithm combined subband decomposition and eigenvalue detector is proposed.The numeri-cal experimental results based on simulated data and real sea clutter data show that the proposed method achieves better detection performance.The above work makes a beneficial exploration in the detection algorithm,which enriches the theory of radar moving target detection under the background of sea clutter to some exten-t,improves the feasibility in the actual system,and has theoretical significance and practical applications.