基于递归抗差估计的高精度GPS动态定位方法

GTP短基线动态定位中,单个历元求解时多余观测较少,求解精度不高,可以用后面历元的观测来增加多余观测,这就形成块模型。由于观测存在粗差,本文采用抗差估计递归求解块模型的浮点解,并且基于抗差估计的方差协方差矩阵,用LAMBDA方法搜索整周模糊度,然后固定基线向量,就可得到当前及以前历元较精确的基线向量的估计。最后利用块模型设计阵的结构特点改进算法提高抗差估计的计算效率。大量数值试验表明本文方法是有效的。

High Precise Method for GPS Kinematic Positioning Based on Recursive Robust Estimator

The precision is not enough when using observations of single time to position in GPS short baseline kinematic positioning.The latter times can be used to add redundant observations.That is the block-angular regression problem. There are outliers in GPS measurements.Recursive robust estimation technique is used to calculate the float resolution of the block-angular model.Based on the variance-covariance matrix,LAMBDA method is used to get the integer ambiguity,then fix the baseline vectors.And it can get more precise estimators of the times.To calculate more efficiently,a modified method is presented using the structure of the design matrix in block-angular model.