Study On Robust Space-time Adaptive Processing For Airborne Radar

In recent years, the development of airborne radar has been drawn more and more attention from researchers and engineers. Although airborne radar has many advantages over ground radar such as flexibility and broad view, it faces much more serious clutter and significant clutter spectrum spread. In realistic projects, the simulation of clutter and space-time signal processing(STAP) have problems of computational complexity and huge number of training samples, so that it is necessary to research on more effective STAP method. This dissertation considers some issues in airborne radar, including clutter characteristic, the simulation of clutter, robust adaptive beamforming in 1-dimension, and space-time signal processing. The main contributions of this thesis are summarized as follows:1. Duo to the complicated environment, the airborne radar echos may be from different kinds of clutters, it is hard to simulate the clutter of the system of airborne radar. In order to solve it, firstly, the component of airborne clutter and the clutter scatter characteristic have been studied. Then the distributing models of amplitude of clutter are discussed, and clutter sequences obeying different distributions are simulated to validate the models. Finally, considering the requirement of project, a simple simulation software of airborne radar clutter is designed based on MATLAB GUI, it can not only simulate different clutter by importing reasonable parameters of airborne radar and scene, but also build airborne clutter database which can be used for the analysis and processing of received radar data.2. In conventional airborne radar, 2-dimension STAP filter is adopted for clutter mitigation. Exsiting STAP methods require enough training samples which need to satisfy IID condition and lead to computational complexity. In fact, it cannot be achieved. To deal with these problems, firstly, two new robust adaptive beamforming methods are proposed, they respectively use the methods of diagonal loading and covariance matrix taper to improve the covariance matrix and reduce the error of steering vector in the conditions of small samples or contaminated training samples; then a spatial temporal separable filter(STSF) is proposed which based on subspacetracking. New method can iteratively calculate the robust spatial and temporal weights with small samples. Our simulation results demonstrate the effectiveness of the proposed methods.