全局收敛高斯-牛顿法解非线性最小二乘定位问题

Global Convergence Gauss-Newton Method Applied to Nonlinear Least Square Estimation in Target Location

将全局收敛策略与牛顿法相结合得到对初值不敏感的迭代算法 ,将该算法应用于非线性最小二乘目标定位中 ,可以在初值受测量误差影响而估计不准的情况下 ,通过迭代得到精确解。

This paper proposes an iteration algorithm insensitive to initial guess which combines the global convergence strategy with Gauss Newton method. When applied to nonlinear least square estimation in target location, this algorithm can approach the precise guess by iterations even with the noise polluted initial guess. Some formulas are derived in this paper when this global convergene Gauss Newton method is applied in 2D target location in bistaric radar system.