系统工程与电子技术

• 传感器与信号处理 • 上一篇    下一篇

基于粒子群优化算法的KK分布参数估计方法

高彦钊, 占荣辉   

  1. 国防科学技术大学电子科学与工程学院, 湖南 长沙 410073
  • 出版日期:2013-12-24 发布日期:2010-01-03

Parameter estimation of KK distribution based on particle swarm optimization algorithm

GAO Yan-zhao, ZHAN Rong-hui   

  1. School of Electronic Science and Engineering, National University of Defense Technology, Changsha 410073, China
  • Online:2013-12-24 Published:2010-01-03

摘要:

针对KK分布的参数估计,首先介绍了半经验估计法,然后提出了一种基于粒子群优化的估计方法。该方法将杂波数据统计直方图与KK分布概率密度函数在部分采样点上的差异作为代价函数,通过粒子群优化搜索参数的最优值。通过蒙特卡罗方法对半经验估计法在权重参数不同时的性能进行了仿真,然后分析了杂波数据样本点数的多少等因素对所提算法精度的影响,最后基于实测合成孔径雷达图像杂波数据对该算法进行了验证。仿真结果表明,该算法对KK分布参数具有良好的估计性能,KK分布与K分布等相比,对合成孔径雷达图像杂波数据具有更强的拟合能力。

Abstract:

Aiming at the parameter estimation of KK distribution, a semi-experiential algorithm is introduced firstly, and then a new algorithm based on particle swarm optimization is presented. This algorithm takes the discrepancies between the histogram of the clutter data and probability density function of KK distribution as the cost function to search the optimal parameters of KK using the particle swarm optimization algorithm. The performance of the semi-experiential algorithm is analyzed for different mixing coefficient using Monte Carlo simulations. And then the influence of some factors, such as the number of clutter data samples, in the new algorithm on the estimation precision is evaluated. At the end, this new algorithm is applied to some real synthetic aperture radar clutter data. The simulation results clearly show the good performance of this algorithm, and the KK distribution is validated to model the radar clutter much better than some common used models, such as K-distribution.