Sea Clutter Characteristic Analysis And Simulation Based On Real Measured Data

The research of sea clutter characteristics and modeling is significant to radarsignal processing, especially to CFAR detector design and radar detection performance.Furthermore, it provides a theoretical basis on parameter optimization, radar signalprocessing method design and radar system design. This paper studies on the seaclutter characteristics before and after signal processing, which focuses on some keyissues of the analysis of sea clutter characteristics. And we propose a method of clutterparameter estimation following K-distribution for the case of thermal noise can’t beignored. For the disadvantage of the traditional clutter parameter estimation, a novelmethod for estimating the distribution parameters of sea clutter is presented based onapproximate Gamma distribution, which provides possible for vast amounts ofreal-time sea clutter data analysis. The main content of this paper specifically includethe following three aspects.1. The paper studies the correlation, and time correlation is analyzed with the realIPIX radar sea clutter data.2. Parameter estimation methods and distribution fitting methods which are basedon amplitude distribution model are discussed in the condition of non-noise. Then thedistribution fitting is analyzed with initial measured sea clutter data and real datawhich is separately processed by STAP and PD. In the absence of thermal noise, it hasbeen observed that the K-distribution doesn’t fit well to observed clutter data, so wepropose a method of clutter parameter estimation following K-distribution whenthermal noise can’t be ignored and adopt measured data to check up the estimatedeffect. For the disadvantage of the traditional clutter parameter estimation, a novelmethod for estimating parameters of K-distributed clutter is presented, and thesimulation results verify the accuracy and effectiveness of it.3. Two methods of generating relevant non-Gaussian clutter are investigated, likeZMNL and SIRP. According to K distribution based on SIRP, clutter simulations arerealized at last.