| The real clutter enviroment of airborne radar is nonhomogeneous, while the training samples which shall be iid with detecting samples are not enough for this case. The optimum performance of regular STAP algorithms needs suffice training samples to calculate covariance matrix, so the performance will worsen seriously in real clutter enviroment. New STAP algorithms for this problem are studied widely and the dissertation here discusses the problems such as STAP clutter suppression algorithms and targets detecting of phase array airborne radar in some kinds of nonhomogeneous clutter environments. The main contributions of this dissertation can be summarized as follows:1,Bistatic clutter model is set up to analyze some clutter nonhomogeneous problems caused by inner reasons of airborne radar, and the influence to STAP of clutter nonhomogeneous phenomena caused by outer reasons is analyzed quantitatively. 2,The received data model of airborne radar is studied, and the model of nonhomogeneous clutter environment is set up to analyze some questions when it is used. The model of nonhomogeneous clutter environment is the precondition of the following research.3,Clutter non-homogeneity detector is studied, and a new clutter non-homogeneity detector is proposed by using frequency center idea for clutter spectrum shifting and nonhomogeneity of clutter power. Another clutter non-homogeneity detector using the residual error idea based on the proposed data model is brought to interferential targets and isolated jamming, simulation results show that the new detector performs well for interferential targets and isolated jamming.4,The physical essences of two new STAP methods (PAMF and STAR) are analyzed based on the proposed data model. They are called as linear forecasting STAP methods because there are some inner relations between them. Linear forecasting STAP methods use the stationarity of pulse dimension data, which can be seen as a kind of reduced dimension combined with reduced rank STAP algorithm. The formula of calculating rank is shown to reduce the computational load of linear forecasting STAP methods.5,Structured reduced dimension STAP algorithm is proposed based on the proposed data model. Structured reduced dimension STAP algorithm which uses the characteristic of received data of airborne radar can get the good adaptive weights with very few samples. Because the computational load of structured reduced dimension STAP algorithm is too great to be deal with real time, a new structured reduced dimension STAP algorithm idea is discussed by combining conjugate grads and clutter non-homogeneity detector based on the residual error to reduce the computational load.6,Clutter parametric applied STAP method which shapes suppressing notch in clutter field by using estimated clutter parameters is proposed by analyzing frequency response of filter. Compared with the regular STAP method, this method performs worse in clutter field, but signal channel loss can be reduced in clear filed, which improves the performance about 2dB. The above method does not suppress interfere well because it only shapes notch to suppress clutter. In order to solve this problem the clutter suppressing filter in clutter parametric STAP method is used as advance filtering followed with STAP processing. After the above processing, clutter can be suppressed entirely and the following STAP processing will have more degree of freedom for interfere, so the proposed method has better nonhomogeneous processing performance.7,The measured data of airborne radar is analyzed theoretically. Based on the former research conclusions a STAP processing system is proposed for nonhomogeneous clutter environment. The experimental results show that the new system performs well for the current studied clutter nonhomogeneous cases, and the computational load is not great for realization.
|
PDF Full Text Request