Pareto杂波模型的参数估计性能研究

The Performance Study of Parameter Estimation Based on Pareto Clutter Model

对于服从Pareto分布的杂波,采用一二阶矩估计法和<xln(x)>估计法存在分布参数的有效估计范围受限的问题。基于<xrln(x)>估计法进行估计,推导出形状参数的有效估计区间为(r,+∞)。在此基础上,利用负数阶矩估计的思想进一步提出<x-1ln(x)>估计法,使形状参数的有效估计范围扩大到(0,+∞)。仿真实验表明,<x-1ln(x)>估计法将形状参数的有效估计范围扩大到整个取值区间。和其他几种估计法的估计性能相比,<x-1ln(x)>估计法的运行速度更快,在区间(0,1)上的估计性能更好,在实际工作中的应用范围更广泛。

The distribution model of clutter and its parameter estimation have been the focus of radar signal processing. Compared with other distribution models, the radar clutter model in a state of subjection to the Pareto distribution is more dominant when it is fitted with the real clutter. As for clutters subject to the Pareto distribution, the range of effective estimation of the distributed parameters are restricted by adopting first two order moment or 〈xln（x）〉 estimation method. Therefore, a novel estimation method based on the 〈x＇ln（x）〉 is proposed, and the effective estimation interval of the shape parameter is derived as （r, ＋∞）. On the basis of such as these, an estimation method of 〈x ^-1ln（x）〉 is further proposed by using the idea of negative order moment estimation, making the range of effective estimation of shape parameters extend to （0, ＋∞）. The simulation results show that the estimation method of 〈x^-1 ln（x）〉 extends the range of effective estimation of shape parameters to the entire range. Compared with other estimators, the estimation method of 〈x-1ln（x）〉 runs faster and has better estimation performance in the interval （0, 1）. And the application scope is more extensive in actual work.