Ma Hongyang, Verhagen Sandra
Geoscience and Remote Sensing, Delft University of Technology, 2628CK Delft, The Netherlands.
Sensors (Basel). 2020 Mar 14;20(6):1634. doi: 10.3390/s20061634.
Precise point positioning (PPP) is one of the well-known applications of Global Navigation Satellite System (GNSS) and provides precise positioning solutions using accurate satellite orbit and clock products. The tropospheric delay due to the neutral atmosphere for microwave signals is one of the main sources of measurement error in PPP. As one component of this delay, the hydrostatic delay is usually compensated by using an empirical correction model. However, how to eliminate the effects of the wet delay during a weather event is a challenge because current troposphere models are not capable of considering the complex atmosphere around the receiver during situations such as typhoons, storms, heavy rainfall, et cetera. Thus, how positioning results can be improved if the residual wet delays are taken into account needs to be investigated . In this contribution, a real-time procedure of recursive detection, identification and adaptation (DIA) is applied to detect the model errors which have the same effects on both phase and code observables; e.g., the model error caused by the tropospheric delay. Once the model errors are identified, additional parameters are added to the functional model to account for the measurement residuals. This approach is evaluated with Global Positioning System (GPS) data during two rainfall events in Darwin, Australia, proving the usefulness of compensated residual slant wet delay for positioning results. Comparisons with the standard approach show that the precision of the up component is improved significantly during the periods of the weather events; for the two case studies, 72.46 % and 64.41 % improvements of root mean squared error (RMS) resulted, and the precision of the horizontal component obtained by the proposed approach is also improved more than 30 % compared to the standard approach. The results also show that the identified model errors are concentrated at the beginning of both heavy rainfall processes when the front causes significant spatial and temporal gradients of the integrated water vapor above the receiver.
精密单点定位（PPP）是全球导航卫星系统（GNSS）的著名应用之一，它利用精确的卫星轨道和时钟产品提供精密定位解决方案。微波信号在中性大气中产生的对流层延迟是PPP测量误差的主要来源之一。作为这种延迟的一个组成部分，静力延迟通常通过使用经验校正模型来补偿。然而，在天气事件期间如何消除湿延迟的影响是一个挑战，因为当前的对流层模型无法考虑台风、风暴、暴雨等情况下接收机周围复杂的大气环境。因此，需要研究如果考虑残余湿延迟，定位结果如何能够得到改善。在本论文中，应用一种递归检测、识别和自适应（DIA）的实时程序来检测对相位和码观测值具有相同影响的模型误差；例如，由对流层延迟引起的模型误差。一旦识别出模型误差，就会在功能模型中添加额外的参数来考虑测量残差。利用澳大利亚达尔文市两次降雨事件期间的全球定位系统（GPS）数据对该方法进行了评估，证明了补偿后的残余斜湿延迟对定位结果的有用性。与标准方法的比较表明，在天气事件期间，垂直分量的精度有显著提高；对于两个案例研究，均方根误差（RMS）分别提高了72.46%和64.41%，并且所提方法获得的水平分量精度相比标准方法也提高了30%以上。结果还表明，识别出的模型误差集中在两次强降雨过程的开始阶段，此时锋面在接收机上方引起综合水汽的显著时空梯度。
