Radar Clutter Analysis And Modleing Based On Compound Gaussian Distributian

Radar systems usually need to work against strong clutter. Modeling and estimation the clutter accurately can improve the radar detection performance,and improve the radar’s capability of clutter suppression. One of the important applications of the radar system is the sea detection. For modern high resolution radar, the sea clutter’s amplitude distribution probability density function usually has a long "tail", which led to the traditional statistical model mismatch, but compound Gaussian distribution model can match with the measured data very well in this case. So it becomes the focus of marine-radar researching in recent years. But the compound Gaussian distribution still has some problems in application. But the structure of the Compound Gaussian distribution is complex, The classic parameter estimation algorithms are ineff icient and very easy to fall into local optimal solution. Aimed at this problem, we study the characteristics of Compound Gaussian distribution model, proposed a parameter estimation method based on component separation method. By this method, the Compound Gaussian distribution model parameters can be quickly estimated.On the other hand, the modern radar system manufacture and testing costs lot of time and money, using the computer system simulate the radar system simulation, based on the simulation system can effectively improve the development efficiency of radar system designing, and reduce the cost of manufacture and testing at the same time. We can see that this kind of simulation is quite valuable. The compound Gaussion distribution can fit well with the actual situation, so the simulation method of the compound Gaussion distribution has a strong practical value.This paper mainly divides into three parts, which are the statistical characteristics of sea clutter, the fast parameter estimation of the compound Gaussian distribution, and the compound Gaussian distribution clutter simulation. The second chapter introduced sea clutter’s characteristics and it’s formation principle, and then studied the models of sea clutter’s amplitude probability density distribution model and the modles of sea clutter’s power spectrum, and analyzed their application range. Got this conclusions: the complex Gauss model fit the high-resolution radar clutter best. At the beginning of the third chapter, we introduced some traditional models’ classical parameters estimation algorithm, and then a detailed analysis of classical parameter estimation algorithm of compound Gaussian distribution, pointed out the problems in those algorithm, finally, using the idea of component separation and the Scale-Independent-ShapeEstimation(SISE) proposed a new parameter estimation algorithm: the compound Gaussian distribution fast estimation of the model parameters. In the fourth chapter, we used several statistical distribution models fitting the measured data of Canadian IPIX radar. To comparison the goodness of fit three kinds of test was used: the mean square error test, the K-S test and the CHIsquare test. Proved that the compound Gauss distribution model fit the highresolution radar clutter best. And proved that the fast parameter estimation algorithm putting forward in this paper is useful. In the fifth chapter, two common methods of clutter simulation are introduced. After analyzing the advantages and disadvantages of those methods, the method based on the ZMNL simulating the compound Gaussian distribution with pointed characteristics was studied and presented.