基于 SSA-ELM 模型的改进的 BDS-2/3 卫星超快速时钟偏差预测。

Improved BDS-2/3 Satellite Ultra-Fast Clock Bias Prediction Based with the SSA-ELM Model.

机构信息

School of Geomatics, Anhui University of Science and Technology, Huainan 232001, China.

Key Laboratory of Aviation-Aerospace-Ground Cooperative Monitoring and Early Warning of Coal Mining-Induced Disasters of Anhui Higher Education Institutes, Anhui University of Science and Technology, KLAHEI (KLAHEI18015), Huainan 232001, China.

出版信息

Sensors (Basel). 2023 Feb 22;23(5):2453. doi: 10.3390/s23052453.

DOI:10.3390/s23052453

PMID:36904657

原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC10007449/

Abstract

Ultra-fast satellite clock bias (SCB) products play an important role in real-time precise point positioning. Considering the low accuracy of ultra-fast SCB, which is unable to meet the requirements of precise point position, in this paper, we propose a sparrow search algorithm to optimize the extreme learning machine (SSA-ELM) algorithm in order to improve the performance of SCB prediction in the Beidou satellite navigation system (BDS). By using the sparrow search algorithm's strong global search and fast convergence ability, we further improve the prediction accuracy of SCB of the extreme learning machine. This study uses ultra-fast SCB data from the international GNSS monitoring assessment system (iGMAS) to perform experiments. First, the second difference method is used to evaluate the accuracy and stability of the used data, demonstrating that the accuracy between observed data (ISUO) and predicted data (ISUP) of the ultra-fast clock (ISU) products is optimal. Moreover, the accuracy and stability of the new rubidium (Rb-II) clock and hydrogen (PHM) clock onboard BDS-3 are superior to those of BDS-2, and the choice of different reference clocks affects the accuracy of SCB. Then, SSA-ELM, quadratic polynomial (QP), and a grey model (GM) are used for SCB prediction, and the results are compared with ISUP data. The results show that when predicting 3 and 6 h based on 12 h of SCB data, the SSA-ELM model improves the prediction model by ~60.42%, 5.46%, and 57.59% and 72.27%, 44.65%, and 62.96% as compared with the ISUP, QP, and GM models, respectively. When predicting 6 h based on 12 h of SCB data, the SSA-ELM model improves the prediction model by ~53.16% and 52.09% and by 40.66% and 46.38% compared to the QP and GM models, respectively. Finally, multiday data are used for 6 h SCB prediction. The results show that the SSA-ELM model improves the prediction model by more than 25% compared to the ISUP, QP, and GM models. In addition, the prediction accuracy of the BDS-3 satellite is better than that of the BDS-2 satellite.

摘要

超快速卫星钟偏（SCB）产品在实时精密单点定位中起着重要作用。考虑到超快速 SCB 的精度较低，无法满足精密单点定位的要求，本文提出了一种麻雀搜索算法来优化极端学习机（SSA-ELM）算法，以提高北斗卫星导航系统（BDS）中 SCB 预测的性能。利用麻雀搜索算法的强大全局搜索和快速收敛能力，进一步提高了极端学习机的 SCB 预测精度。本研究使用国际全球导航卫星系统监测评估系统（iGMAS）中的超快速 SCB 数据进行实验。首先，采用二阶差分法对所使用数据的精度和稳定性进行评估，证明了超快速钟（ISU）产品观测数据（ISUO）与预测数据（ISUP）之间的精度最优。此外，BDS-3 上的新型铷（Rb-II）钟和氢（PHM）钟的精度和稳定性优于 BDS-2，不同参考钟的选择会影响 SCB 的精度。然后，采用麻雀搜索算法-极端学习机（SSA-ELM）、二次多项式（QP）和灰色模型（GM）进行 SCB 预测，并与 ISUP 数据进行比较。结果表明，当基于 12 h 的 SCB 数据预测 3 h 和 6 h 时，SSA-ELM 模型分别提高了 ISUP、QP 和 GM 模型的预测模型的~~60.42%、5.46%和 57.59%和 72.27%、44.65%和 62.96%。当基于 12 h 的 SCB 数据预测 6 h 时，SSA-ELM 模型与 QP 和 GM 模型相比，分别提高了~~53.16%和 52.09%和 40.66%和 46.38%。最后，使用多日数据进行 6 h 的 SCB 预测。结果表明，SSA-ELM 模型与 ISUP、QP 和 GM 模型相比，提高了超过 25%的预测模型。此外，BDS-3 卫星的预测精度优于 BDS-2 卫星。