抗差主成分估计在板块运动参数解算中的应用

为解决在采用空间大地测量数据求解板块运动欧拉参数过程中可能出现的矩阵病态问题,以及消除粗差对解算结果的影响,利用抗差主成分估计方法,采用IGG3方案的选权迭代方案对北美板块的2 577个台站进行检测,剔除包含粗差的异常站数据,求出欧拉参数最优解,并进行误差估计,建立北美板块的运动模型。计算结果表明,剔除异常台站前后精度有一定的提高,参数结果与其他模型的解算结果具有较强的一致性。

Application of principal components estimate to the calculation of plate motion parameters

To solve the problem of motion matrix morbid appearing in the process of solving the Euler parameters of plate motion using the space geodetic data ,and to eliminate the influence of gross error on the results ,this paper detects the 2 577 stations of North American plate by using principal component estimation method with IGG3 iteration scheme .The abnormal station data containing gross error is eliminated .Then the North American plate motion model is established with the optimal solution of Euler parameters and error estimation .The result shows that the accuracy increases to a certain extent before and after excluding abnormal stations and the results have strong consistency with other models .