摘要
散射物体自然频率 (极点 )是目标识别可以利用的一个基本特征 ,传统的理论求解极点的方法一般是用迭代法求取由矩量法得到的矩阵方程系数行列式的零点。矩阵方程系数矩阵的病态性是其最大的缺点。文章提出了用遗传算法获得极点的方法 ,对遗传算法应用于求取雷达目标极点方面的问题做了理论和算法分析。由于球体目标的极点可以解析获得 ,所以利用了球体作为实验验证的例子。并计算了有限长圆柱的结果。实验结果表明算法相对于传统的围线积分法不仅避免了病态矩阵的问题 ,而且简单易行 ,结果可靠。
The natural frequencies (poles) of a scattering object is one of the basic characteristics in application of target recognition. Usually, an iterative method is used to solve the matrix equation of the moment of method (MoM). However, the ill-condition of the coefficient matrix in the matrix equation is the greatest shortcoming. In this paper, a technique is proposed to extract the poles with genetic algorithm. Since the poles can be obtained by an analytic expression, a perfect conducting sphere is employed as an example to demonstrate the validity of the algorithm. The poles of a finite cylinder is also calculated. The simulation results indicate that comparing to the contour-integral method, the genetic algorithm not only avoids the trouble of ill-condition matrix, but also is much simpler in the implementation.
出处
《电波科学学报》
EI
CSCD
2004年第2期157-160,214,共5页
Chinese Journal of Radio Science
基金
国家高科技发展计划项目 (编号 :2 0 0 1AA132 0 2 0 )
关键词
遗传算法
雷达
目标极点分布
散射物体
目标识别
电磁散射
Electromagnetic wave scattering
Genetic algorithms
Integral equations
Iterative methods
Method of moments
Natural frequencies
