Parameter Estimation of K Distribution Radar Clutter with the Gravity Searching Algorithm

Parameter estimation is an important task in the modeling, classification, and detection of radar clutters. Radar clutters have stochastic characteristics. Therefore, Statistical distributions are usually used to describe the features of clutters better. K distribution is one of the most common models utilized to the simulation of clutters. This distribution, which consists of scale and shape parameters, has two speckle and local power components. Because local power component is modeled by gamma distribution, the parameter estimation of K distribution is a high dimensional and nonlinear problem. In this paper, a novel method is proposed based on the gravity searching algorithm for the parameter estimation. This new method has high accuracy and validity in estimating parameters. For the evaluation of the proposed method, the estimated probability density function and power spectrum in two different experiments were compared to actual ones. Finally, the results of the new method are compared to the results of the maximum likelihood method. Furthermore, K-S test is performed to evaluate generated clutters with estimated parameters. Results prove the validity of the proposed method for the parameter estimation.