Study On STAP Methods For Airborne Radar In Nonhomogeneous Environment

In the nonhomogeneous environments, the performance of conventional statistical space time adaptive processing (STAP) often declines sharply due to the lack of independent identically distributed training samples. This dissertation focused on the investigation of clutter suppressing techniques for airborne radar in nonhomogeneous environment. The solutions to the nonhomogeneity about interfering target and isolated interference are studied from chapter 2 to 5, the ones to range dependence of non-sidelooking array are investigated in chapter 6.Chapter 1 is the introduction. The histories of STAP and related techniques, and the latest developments in China and abroad are briefly reviewed. Then, the research background and main contents are addressed.Chapter 2 focuses on the Joint Domain Localized (JDL) and gain and phase error calibration. The JDL adaptive processing algorithm is very suitable for nonhomogeneous environment due to its low computational load and small sample support. Howerer, the gain and phase errors will be inevitable and will result in the JDL performance degradation. In order to improve the JDL performance, these errors should be properly calibrated. In an ideal condition, the clutter covariance matrix of linear equispaced array has Block Toeplitz with Toeplitz Block(BTTB) structure. However, this structure is not preserved in a practical system due to the existence of the gain and phase errors.According to this property, a simple and efficient gain and phase error calibration method is proposed. The error estimation is realized by solving the linear equations of gain and phase errors based on different diagonal lines of the covariance matrix.Charper 3 studies the interfering target detection algorithm. The Nonhomogeneous Detector (NHD) acting as an effective tool to detect interfering target, generally works under the condition of interfering target existence. The performance of adaptive NHD algorithm is affected by these outliers. A new interfering target detection algorithm MGIP is developed in this paper, which uses the sample data phase information and is not dependent on the sample amplitude information. The improved algorithm can recognize the weaker outliers effectively, although the stronger outliers exist, and remove them from the training data.Charpter 4 investigates the discrete interference suppression. The surrounding range cells do not possess information about discrete interference in the primary range cell under test. To suppress it,"single data set"(SDS) STAP algorithms are required, witch only operates on the test data.The direct data domain(DDD) algorithms is one of these algorithms, it has favorable performance under ideal environment. However, when the array errors exist or the interelements is not equispaced, the signal will be cancelled and the detection performance will degrade. In this paper, the space-time 2-dimensional amplitude and phase estimation of a sinusoid (2D APES) filter is applied to suppress interference and estimate the signal amplitude, and the maximum likelihood estimation detector (MLED) is applied to detect target. The covariance matrix estimation is improved through sliding window partitioning, and the output SCNR is improved by space-time 2-dimensional adaptive processing. Compared with the DDD algorithm, the 2D APES algorithm is more robust to amplitude and phase errors.Charpter 6 sdudies the performance of multistage Wiener filter (MWF) under nonhomogeneous environment. The MWF does not utilize the clutter variance matrix estimation, inversion and eigendecomposition techniques, which can dramatically reduce the computational cost and enhence the convergence speed. However, the interfering targets included in training samples will lead to the MWF performance degradation. In order to inherit the MWF's merit and overcome its shortcoming simultaneously, the MLED algorithm which only operates on single range gate data is applied to eliminate the outliers in this paper. Then, the adaptive weight is computed via the reduced rank MWF. Moreover, the modified Concurrent Block Processing (CBP) is applied to reduce the computational cost.Charper 6 investigates the short range clutter suppression technique for non-sidelooking array. The performance of Doppler warping (DW) and angle-Doppler compensation (ADC) algorithms are analyzed, based on these two algorithms, the multiple space angle compensation algorithm MDW and MADC are proposed. Compared with DW and ADC methods, the improved algorithms can further reduce the geometry-induced nonhomogeneity by accomplishing multiple space angle alignment at multiple Doppler frequencies from reference cell to test cell. Additionally, the DW and ADC algorithms are respectively incorporated with reduced dimension, reduced rank, clutter covariance matrix estimation and STAR algorithms. The clutter suppression performance can be further improved.The work of the whole thesis is concluded in chapter 7, and the issues to be further studied are pointed out as well.