基于最小二乘支持向量机的SAR平台定位

将SAR平台定位分解为粗略定位和精确定位两个阶段。首先，在SAR正侧视成像的条件 下，利用某个方位门内的所有控制点，采用非线性最小二乘平差粗略估计出SAR平台在该方 位时刻的空间位置，并从理论上推导了控制点的误差协方差矩阵到SAR平台定位的误差协方 差矩阵的传递规律。其次，利用各个方位时刻粗略估计的SAR平台位置，采用最小二乘支持 向量回归机精确估计SAR平台的运动方程，从而精确估计SAR平台在某个时刻的空间位置。仿 真试验表明，本文提出的方法能够精确地反演出SAR平台的空间位置。

SAR Platform Geo-location Based on Least Squares Support Vector Machines

The SAR geo-location method is composed of two stages. The first stage employs a nonlinear least squares adjustment method which using all control points in an azimuth door to coarsely estimate the SAR platform location at the azimuth time under zero imaging squint angle. And the error propagation law from control point to coarse SAR position is deduced theoretically. The second stage employs a least squares support vector machine which based on coarsely evaluated SAR positions at each azimuth time to finely estimate the SAR platform motion equation in the imaging process. And the accurate SAR platform position at any time is calculated by the least squares support vector machine. Simulated experiment results show the excellent performance of the SAR geo-location method proposed in this paper.