K-distribution parameters estimation based on the Nelder-Mead algorithm in presence of thermal noise

In this paper, we propose an efficient approach to the estimation of the compound K-distribution parameters in presence of additive thermal noise. This is acquired by means of a multidimensional unconstrained nonlinear minimization algorithm based upon the Nelder-Mead direct search method. In doing this, we minimize the sum of squared residuals. The best fit is simply achieved by a direct comparison of the experimentally measured cumulative distribution function (CDF) of the recorded data with the set of curves derived from the model of interest. A good minimization can be reached only if the real CDF is accurately estimated. We show, particularly, that the new approach yields the best spiky clutter parameter estimation. The proposed estimator is more efficient than existing estimation methods.