最小二乘方差分量估计在GNSS差分定位随机模型精化中的应用

Least-Squares Variance Component Estimation Applied to Stochastic Model Refinement of GNSS Difference Positioning

使用最小二乘方差分量估计法对基于卫星高度角随机模型中的未知参数进行估计,通过实际观测数据,以负方差为指标,分析指数模型和各种三角函数模型的适用性,计算比较等权模型、指数模型、正切模型及余弦模型的基线向量解算精度,指出应根据实际使用情况合理选择随机模型,以提高定位精度。

The authors apply least-squares variance component estimation to evaluate unknown parameters of five different elevation-dependent stochastic models with double differenced GPS observables, and then use negative variance components as indicators to detect the applicability of models. Afterwards, the tangent, cosine, exponential function and the identical weight models are compared by calculating the baseline components. Test results indicate that the positioning accuracy and efficiency are dependent on elevation, and that one should choose the realistic stochastic model.