摘要
考虑了双基地雷达目标定位问题中的非线性最小二乘方程组的迭代解法。用高斯 牛顿迭代法解非线性最小二乘方程组计算量小、收敛快,但所得解的正确性及精度依赖于选取的迭代初值与真值的靠近程度,及方程组的非线性强度。给出了两种变步长全局收敛策略,与高斯-牛顿法相结合可得到对初值不敏感的迭代算法。仿真结果表明,用全局收敛的高斯 牛顿法解最小二乘方程组能得到更准确的解,且迭代次数较少。
The nonlinear least square estimation of target location in bistatic radar system and its iterative solution are concerned. The Gauss-Newton (G-N) iterative algorithm applied to solve the nonlinear least square equations is of the advantages in computation and convergence rate. Nonetheless the correctness and accuracy of the solution depend on the distance of the initial iteration guess and the true value as well as the nonlinearity degree of the equations. Two global convergence strategies of changing step are given to combined with G-N method that can perform the virtue of insensitiveness to initial guess in iteration algorithm. The simulation results show that the global convergent G-N method can approach a much accurate solution with fewer iteration step.;
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2004年第1期30-33,43,共5页
Systems Engineering and Electronics
关键词
目标定位
高斯-牛顿法
双基地雷达
target location
G-N method
bistatic radar systemlet power spectrum
