多频GNSS非差非组合PPP多径误差修正及定位性能评估

Multi-frequency GNSS Undifferenced and Uncombined PPP with Multipath Error Correction and Positioning Performance Evaluation

  • 摘要: 多径误差是影响精密单点定位(precise point positioning, PPP)收敛时间和定位精度的主要因素之一,处理多频数据需对各频率建立多径误差修正模型,增加了数据处理的复杂度。选用全球分布的12个国际全球导航卫星系统(global navigation satellite system, GNSS)服务观测站的GNSS多频观测数据,以评估非差非组合PPP多径误差修正效果。首先通过恒星日滤波法对部分频率的相位多径建模,然后利用多元线性回归构建频间相位残差相关性模型并获取全频点多径改正数,最后进行定位验证。实验结果表明,任意一频点的相位多径均可由其他频点的相位多径进行拟合,且拟合精度可达91%及以上。多径修正对单系统不同频率组合定位性能的提升效果明显:静态条件下,GPS系统平均收敛时间降低了19.2%;Galileo系统双频、三频、五频收敛时间分别缩短31.2%、42.8%和32.9%;北斗系统双频、三频收敛时间分别缩短16.4%、17.0%至28 min左右,动态同样具有明显提升。三系统组合静态PPP平均收敛时间由15 min缩短至13 min左右,收敛后东、北、天 3方向定位精度分别达0.27、 0.24、 0.80 cm;动态平均收敛时间可由17 min缩短至16 min左右,3方向定位精度分别达0.82、 0.68、 2.40 cm。因此,研究频间相位多径相关关系、获取全频多径改正数对于提升多频GNSS多径修正效率及定位精度具有一定意义。

     

    Abstract:
    Objectives Multipath effect is one of the main factors that affect the convergence time and positioning accuracy of precise point positioning (PPP). For multi-frequency GNSS undifferenced and uncombined PPP, it is necessary to establish a multipath correction model for each frequency, which increases the complexity of data processing. We evaluate the effectiveness of multipath correction in undifferenced and uncombined PPP using multi-frequency GNSS observation data from 12 International GNSS(global navigation satellite system) Service stations.
    Methods We extract the raw phase residuals based on GPS/Galileo/BDS(BeiDou satellite navigation system) multi-frequency undifferenced and uncombined PPP, and use multiple linear regression to construct a correlation model of inter-frequency phase residuals. The sidereal filtering is used to establish a multipath model for the phase residuals of some frequencies, combined with the residual correlation model to recover the full-frequency multipath correction.
    Results The results show that the phase residuals of any frequency can be fitted by the phase residuals of other frequency points, with a fitting accuracy of 91% or more. In static conditions, it leads to an average decrease of 19.2% in the GPS convergence time. In contrast, for Galileo systems employing dual-frequency, triple-frequency, and five-frequency PPP positioning, the convergence time is reduced by 31.2%, 42.8%, and 32.9%, respectively. Similarly, for BDS dual-frequency and triple-frequency positioning, the convergence time shortened by 16.4% and 17.0%, stabilizing around 28 min. The kinematic results also showed significant improvement. Additionally, for GPS/Galileo/BDS combination, the average convergence time was shortened from 15 min to about 13 min with multipath correction in static mode, the resulting positioning accuracies in the east, north, and up directions reached 0.27 cm, 0.24 cm, and 0.80 cm, respectively. Under kinematic conditions, the average convergence time of GPS/Galileo/BDS was decreased from 17 min to about 16 min. Furthermore, the positioning accuracies of GPS/Galileo/BDS in the east, north, and up directions reached 0.82 cm, 0.68 cm, and 2.40 cm, respectively.
    Conclusions Therefore, studying the inter-frequency phase multipath correlation and obtaining a comprehensive multipath correction model is of significant importance for improving the efficiency of multi-frequency GNSS multipath mitigation and positioning accuracy.

     

/

返回文章
返回