Radar Target Detection Under Complex And Non-homogeneous Clutter

The complex and non-homogeneous clutter background is the main problem that the target detection technology is facing. In non-sidelooking mode, the range-invariance of clutter is violated, which is known as geometry-induced range dependence. This range dependence limits the practical implementation of the standard STAP approaches which rely on the covariance matrix estimation by using the training data from adjacent ranges. In this paper, we propose to use the statistical properties of samples to test the homogeneity of the clutter background. Then use the same identification to mark the samples which have the same statistical characteristics. The samples which have the same identification can be used to choose the reference units to improve the accuracy of estimation of the covariance matrix and improve the robustness of STAP.The detailed research work of article is as follows:1、Analyze the characteristics of airborne radar echo data and introduce the principle of the traditional STAP method. In forward-looking mode, the range Doppler dependence result the clutter background become complex and non-homogeneous. From the introduction of the STAP, we know that estimated samples are required to be independent and identically distributed. The accuracy of the estimation of covariance matrix determines the STAP clutter suppression performance.2、A thorough study on the method of homogeneity test. Introduce the principle of E distance homogeneity test, AD homogeneity test and K sample AD test. Through simulation experiments, it is proved that they have the ability to distinguish different probability distributions, and the performances of them are also discussed. After comparing the experimental results and the running time, this paper decides to use the K sample AD test as the homogeneity test method.3、A partition algorithm based on K-AD homogeneity test is proposed. The K-AD homogeneity test is to determine whether the two sequence is satisfied with the same probability distribution by calculating the value of the K-AD detected between the two sequences, and comparing the size of the detection with the empirical threshold value. In this paper, according to the form of the actual radar echo data, the application of the partition algorithm is developed b y using the reference window, and the partition results of the whole data sample are obtained.4、A STAP algorithm based on K-AD homogeneity test is proposed. The partition algorithm based on the K-AD homogeneity test can achieve the division of the data samples with the same statistical characteristics in the actual radar echo data. According to this partition, we can choose the same probability distribution of the samples as the reference unit to estimate the covariance matrix. The focus of this approach is the choice of reference unit, and the subsequent STAP algorithm is used to the superior performance of 3DT-STAP. By using the simulation data and the measured data that the simulation targets are added, observe the SINR changes before and after processing. The algorithm has the ability of clutter suppression, and can achieve the effective suppression of clutter.5、Process the real radar echo data with the GIP-STAP、APR-STAP、PST-STAP、K-AD-STAP. By comparing the value of SINR with different processing methods, the performances of them are analyzed, and the clutter suppression ability of the proposed algorithm is observed.