TL;DR
This paper mathematically analyzes Horizontal Dilution of Precision (HDOP) and Vertical Dilution of Precision (VDOP) for dual Global Navigation Satellite Systems (GNSS) under an ideal satellite placement scenario. The authors propose a method to optimize satellite selection for improved positioning accuracy, demonstrating its effectiveness with simulations and real-world data.
What Problem Does This Paper Address?
As the number of satellites in Global Navigation Satellite Systems (GNSS) like GPS, GLONASS, Galileo, and BeiDou increases, so does the potential for improved positioning accuracy and availability. However, simply using more satellites can lead to increased computational load and power consumption, especially for mobile devices. Furthermore, the accuracy improvement rate diminishes with more satellites, and the overall positioning accuracy is heavily influenced by the geometric arrangement of the visible satellites. This geometric arrangement is quantified by Dilution of Precision (DOP) values, with lower DOP indicating better accuracy. While previous research focused on minimizing the overall Geometric DOP (GDOP), this paper specifically investigates the Horizontal DOP (HDOP) and Vertical DOP (VDOP) for dual GNSS constellations, recognizing that different applications prioritize horizontal or vertical accuracy. The challenge is to find an optimal satellite selection strategy that balances accuracy, computational load, and availability, particularly when using multiple GNSS constellations.
Methods
The core of this research lies in the mathematical derivation and analysis of HDOP and VDOP for dual GNSS under an ideal satellite placement known as the Zenith + Horizon (ZH) method. This method, previously proposed for minimizing GDOP, places satellites at the zenith and uniformly around the horizon. The authors extend this by:
- Mathematically deriving HDOP and VDOP formulas: They derive these metrics for both single and dual GNSS scenarios, separating the GDOP into its horizontal and vertical components. This allows for application-specific optimization.
- Developing a dual GNSS framework: The mathematical model is extended to account for two GNSS constellations simultaneously, addressing the complexities of differing reference times and receiver hardware delays.
- Simulations: The derived formulas are validated through simulations using various satellite configurations to assess HDOP and VDOP.
- Satellite selection algorithm: A novel algorithm is proposed to adjust satellite selection based on the derived HDOP and VDOP, aiming to improve these metrics by changing satellite combinations, particularly in dual GNSS scenarios. This algorithm prioritizes reducing SST (Squared Sum Trigonometric function), which is found to correlate with lower HDOP.
- Real-world data verification: The effectiveness of the derived formulas and the satellite selection algorithm is tested using actual GNSS data from GPS and Galileo.
Datasets
The study utilizes actual satellite data acquired over approximately 29 hours from 01:00 on December 22, 2021, to 06:00 on December 23, 2021. The data was collected on the roof of the New Engineering Building at Konkuk University, Seoul, South Korea, using a NOVATEL OEM6 receiver. GPS and Galileo were selected as the dual GNSS constellations for the analysis.
Key Insights and Ideas
- HDOP and VDOP are more application-specific than GDOP: While GDOP provides an overall measure of accuracy, separating it into HDOP and VDOP allows for tailoring satellite selection to applications that prioritize horizontal (e.g., car navigation) or vertical (e.g., aerial navigation) accuracy.
- Uniform satellite placement is key for low HDOP: The study highlights that increasing the number of uniformly placed satellites (especially in dual GNSS) and achieving uniformity within each GNSS constellation (measured by SST) significantly reduces HDOP.
- VDOP is less directly correlated with simple satellite counts: Unlike HDOP, VDOP's behavior is more complex and less directly predictable from simple satellite counts. Simulations are crucial for understanding VDOP tendencies.
- Dual GNSS satellite selection requires careful consideration of constellation types: The study demonstrates that simply selecting satellites closest to an ideal placement isn't always optimal for dual GNSS. A satellite changing algorithm that considers the GNSS type is necessary to achieve lower HDOP and VDOP.
- Ideal ZH placement is a strong baseline, but real-world data requires adaptation: While the ZH placement provides a good starting point for minimizing DOP, real-world satellite visibility and availability necessitate dynamic satellite selection algorithms to maintain low DOP values.
Key Results
- Mathematical derivations: The paper successfully derives HDOP and VDOP formulas for both single and dual GNSS under the ZH placement conditions.
- Simulation validation: Simulations confirm that HDOP decreases with an increasing number of uniform satellites. For VDOP, it decreases when the number of zenith and uniform satellites becomes more balanced.
- Real-world data analysis:
- The satellite selection algorithm, which prioritizes decreasing SST, led to a significant reduction in HDOP (e.g., an 18.5552 reduction in one instance).
- Similarly, for VDOP, a satellite changing algorithm that considered GNSS types resulted in a substantial reduction (e.g., 19.2897).
- The study found that while single GNSS generally exhibited lower HDOP and VDOP than dual GNSS in ideal placements, the dual GNSS approach becomes more critical when actual satellite visibility is limited, as it offers more flexibility.
Limitations and Future Work
- Mathematical complexity of dual GNSS: The study was limited to investigating dual GNSS due to the significant mathematical complexity involved. The authors suggest that the observed tendencies are likely to extend to multiple GNSS.
- Optimality of satellite selection: While the study demonstrates the necessity and effectiveness of satellite selection, it did not delve into analyzing the absolute optimality of the selection method in detail.
- Limited number of satellites in simulations: The simulations were performed with a limited number of satellites (up to 6) due to computational constraints. The authors believe the derived tendencies will hold for a larger number of satellites.
- Future work: Further research is needed on satellite placement strategies for dual GNSS, especially when actual satellites closest to the ideal ZH placement are selected, as this is more challenging with fewer satellites. Investigating the optimality of satellite selection methods is also a promising avenue.
Summary updated