Tan Xiaorong, Xu Jiangning, Li Fangneng, Wu Miao, Liang Yifeng, Chen Ding, Zhu Bing
School of Electronic Engineering, Jiujiang University, Jiujiang, 332005, China.
Department of Navigation Engineering, Naval University of Engineering, Wuhan, 430000, China.
Sci Rep. 2024 Aug 2;14(1):17897. doi: 10.1038/s41598-024-69005-2.
Precise forecasting of satellite clock bias is crucial for ensuring service quality and enhancing the efficiency of real-time precise point positioning (PPP).The grey model with many benefits is an excellent choice for predicting real-time clock bias. However, the absolute prediction error of a small number of satellites is very high in actual forecasting process. In order to address this issue, a non-homogeneous white exponential law grey model has been constructed specifically for predicting clock bias sequences with non-homogeneous class ratio approximating constants. Considering that any model is difficult to apply to different kinds of satellite clocks and clock bias signals, an adaptive selection strategy for forecast model is proposed to ensure more excellent prediction accuracy. Afterwards, a prediction scenario based on the observed products of the BeiDou satellite navigation system (BDS) is created to demonstrate the effectiveness of the approach described in this article. The outcomes of the method are then compared with those of a single grey model and the ultra-rapid predicted products. The outcomes demonstrate that this strategy's prediction accuracy is less than 1 ns/day and that its prediction uncertainty is much decreased, which may improve data selection for real-time applications like real-time kinematics (RTK) and PPP.
精确预测卫星钟差对于确保服务质量和提高实时精密单点定位（PPP）的效率至关重要。具有诸多优点的灰色模型是预测实时钟差的理想选择。然而，在实际预测过程中，少数卫星的绝对预测误差非常大。为了解决这个问题，专门构建了一种非齐次白指数律灰色模型，用于预测类比近似常数的非齐次钟差序列。考虑到任何模型都难以适用于不同类型的卫星钟和钟差信号，提出了一种预测模型的自适应选择策略，以确保更高的预测精度。之后，创建了一个基于北斗卫星导航系统（BDS）观测产品的预测场景，以证明本文所述方法的有效性。然后将该方法的结果与单一灰色模型和超快速预测产品的结果进行比较。结果表明，该策略的预测精度小于1纳秒/天，且其预测不确定性大大降低，这可能会改善实时动态定位（RTK）和PPP等实时应用的数据选择。
