基于精密单点定位和自适应最小二乘配置建立新疆水平速度场格网模型

针对双差计算测站坐标效率低下的问题，采用GNSSer精密单点定位方式多核并行处理测站数据，加快解算速度；针对测站速度含有随机信号导致速度场建模精度降低的问题，采用最小二乘配置来估计随机信号，并引入Helmert方差分量估计来调整噪声与信号协方差矩阵间的不合理关系，建立精度更高的速度场模型。以新疆陆态网络2011~2017年连续运行基准站为例，利用GNSSer精密单点定位方式获取坐标时间序列，大幅提升了解算效率，验证了GNSSer的可靠性和高效性，证实精密单点定位可获得与双差定位基本一致的速度信息（差异在1.5 mm/a以内）；建立新疆水平速度场格网模型，结果表明，新疆水平运动速度为27.1~34.8 mm/a，整体趋势为北向偏东，自西南到东北部，运动方向由北偏东向东偏转。

Establishment of Horizontal Velocity Field Grid Model in Xinjiang Based on Precise Point Positioning and Adaptive Least-Squares Collocation

In order to solve the inefficient problem in calculating station coordinates with double-difference positioning, GNSSer PPP mode is introduced to speed up the solution process by multicore parallel operation. The precision of velocity field model is reduced due to the random signal contained in the station velocity, so least-squares collocation is used to estimate the random signal, and the Helmert estimation of variance components is introduced to adjust the unreasonable relationship between the noise covariance matrix and the signal covariance matrix. Then, a more accurate velocity field model can be established. Taking the continuously operating reference stations of the CMONOC in Xinjiang from 2011 to 2017 as an example, GNSSer PPP mode is used to obtain the coordinate time series to improve the efficiency of the calculation. The reliability and efficiency of GNSSer are verified. It is also confirmed that the PPP can obtain the velocity information, which is substantially consistent with the double-difference positioning, the difference is within 1.5 mm/a. The established horizontal velocity field grid model in Xinjiang shows that the movement velocity in Xinjiang is 27.1-34.8 mm/a, and the overall trend is north to east. From southwest to northeast in Xinjiang, the direction of motion is deflected from north to east.