Tag: precise orbit determination

New satellite orbit determination method could boost navigation precision for future mega-constellations

The rotation-corrected integrated POD method holds significant promise for global navigation augmentation, autonomous LEO-based navigation systems, and real-time positioning services.

Modern satellite constellations such as OneWeb, Starlink and CENTISPACE promise global communications and navigation capabilities using low-Earth orbit (LEO) constellations. However, their precise orbit determination (POD) requires dense ground station networks — costly and often limited by geopolitical or geographical constraints.

Inter-satellite links (ISLs) help reduce ground dependence but suffer from “rotational unobservability,” where the entire constellation drifts in orientation due to the lack of an absolute spatial reference. Existing fixes often require additional infrastructure or high-quality GNSS products, which increase latency and operational complexity.

Because of these challenges, a more autonomous, low-latency approach that leverages existing onboard capabilities is needed to ensure reliable, high-accuracy orbits for mega-constellations.

Wuhan University researchers have developed and validated a rotation-corrected integrated POD method that fuses ISL measurements with onboard BeiDou-3 (BDS-3) GNSS observations. Published (DOI: 10.1186/s43020-025-00175-8) in Satellite Navigation on Aug. 4, the study demonstrates how the technique simultaneously estimates the orbits of LEO and BDS-3 medium-Earth-orbit (MEO) satellites, corrects systematic rotation using BDS-3 broadcast ephemerides, and achieves centimeter-level precision.

The approach significantly reduces reliance on ground stations, making it well-suited for real-time applications in large-scale LEO constellations, the researchers said.

The team simulated a 66-satellite LEO constellation equipped with ISLs and onboard BDS-3 receivers, alongside 24 real BDS-3 MEO satellites. Two processing strategies were tested: using BDS-3 data from all LEOs, and from only a subset. In both cases, ISL and GNSS data were jointly processed to form a unified high–low constellation.

Due to internal-only measurements, the initial solutions exhibited significant systematic rotation — up to 40 cm cross-track error for LEOs and over 1 meter for MEOs.

This innovation could become a cornerstone technology for integrating LEO constellations with existing GNSS systems to enhance global navigation and timing performance.

The researchers derived rotation angles between the integrated POD coordinate frame and the BeiDou Coordinate System implied in broadcast ephemerides, then applied a Helmert transformation to correct the orbits. After correction, LEO along-track and cross-track errors dropped from 22.7 cm and 39.3 cm to 1.3 cm and 4.2 cm, respectively. MEO errors fell from over 1.2 m to about 13 cm.

Even when only 36 of 66 LEOs carried GNSS receivers, ISL connectivity propagated the correction across the constellation with minimal accuracy loss. Tests also examined the influence of predicted Earth rotation parameters and residual errors in broadcast ephemerides.

“This method tackles one of the most stubborn issues in autonomous constellation orbit determination — systematic rotation caused by the lack of absolute spatial reference,” said Kecai Jiang, corresponding author of the study. “By harnessing readily available BDS-3 broadcast ephemerides and inter-satellite measurements, we can deliver centimeter-level precision without waiting for post-processed GNSS products or building extensive ground networks. This approach is not only efficient but also scalable, paving the way for real-time, high-accuracy navigation services in future mega-constellations.”

The rotation-corrected integrated POD method holds significant promise for global navigation augmentation, autonomous LEO-based navigation systems, and real-time positioning services. By dramatically reducing reliance on ground infrastructure, it enables resilient operations in remote or geopolitically constrained regions. Its scalability makes it suitable for next-generation satellite constellations supporting broadband internet, disaster response, and precision agriculture, the researchers said.

Moreover, the ability to achieve near-uniform accuracy across all satellites — even when only part of the constellation carries GNSS receivers — lowers hardware requirements and operational costs. This innovation could become a cornerstone technology for integrating LEO constellations with existing GNSS systems to enhance global navigation and timing performance.

August 20, 2025
Research Report: Advancing precision in navigation

Photo: Government of Japan

In early 2015, the Navigation Support Office of the European Space Agency (ESA) and the Japan Aerospace Exploration Agency (JAXA) began a collaboration. At its core, the ESA-JAXA collaboration is designed to cross-validate Japan’s Quasi-Zenith Satellite System (QZSS) Precise Orbit Determination (POD) results and share expertise to improve the POD accuracy of QZSS.

The cross-validation of the QZSS POD performance was implemented by jointly analyzing QZSS observations and validating the POD results of the QZSS satellites. As a result of this joint activity, ESA and JAXA have significantly improved the robustness and accuracy of their respective POD products. This collaborative approach not only ensures the continuous improvement of QZSS force modeling and precise orbit determination performance but also demonstrates the effectiveness of international cooperation in advancing the field of space navigation, especially as the benefits of GNSS interoperability become very evident.

An important milestone in this collaboration was ESA’s role in supporting the In-Orbit Testing (IOT) activities for QZS-1R towards the end of 2021. The successful execution of these tests demonstrated the practical results of the ESA-JAXA partnership and further solidified the commitment of both agencies to enhance their capabilities for QZSS POD and associated products.

FIGURE 1 ESA’s Solar Radiation Pressure (SRP) model output in satellite-Sun frame.

The benefits of this collaboration extend beyond the agencies to the entire scientific community. Notable achievements include the revision of metadata for the QZSS constellation, such as the optical properties of the QZS-1 solar arrays, which have been refined and improved through shared expertise, while simultaneously releasing the satellite mass and attitude mode history in a machine-readable file format for easy access and adoption by the users.

To evaluate the spacecraft models and metadata for QZS-1R prior to their public release, ESA and JAXA conducted several comparative tests. Since both organizations use different software packages for satellite POD — ESA uses NAPEOS (Dow, Springer 2009, Enderle et al., 2019 and 2022) and JAXA uses MADOCA (Kawate et al., 2023) — their results can be considered as largely independent. One comparison involved the Solar Radiation Pressure (SRP) model results produced by both organizations. FIGURE 1 shows the accelerations in satellite-Sun frame computed by ESA’s SRP model. The comparison of the computed SRP accelerations in different reference frames, spacecraft-fixed and inertial, showed excellent agreement with differences of less than 0.1 nm/s².

FIGURE 2 One-way Satellite Laser Ranging (SLR) range residuals calculated with respect to QZS- 1R orbits generated with (green) and without (blue) a-priori radiation force models and displayed as function of the Earth-Probe-Sun angle.

In addition, pseudo-range and carrier phase dual-frequency measurement data from 200 tracking stations of the International GNSS Service (IGS) network were used to generate precise QZS-1R satellite orbits and clock offsets on a day-to-day basis over a 12-month period spanning from January to December 2022. Comparison between ESA and JAXA solutions yielded a root-mean-square (RMS) agreement of 8.6 centimeters (orbit) and 0.21 nanoseconds (clock), respectively. Analysis of Satellite Laser Ranging (SLR) data from seven stations of the International Laser Ranging Service (ILRS) suggests a radial RMS accuracy of the generated orbital trajectories of about 4 cm. Without applying the analytical models for SRP and other non-gravitational perturbation forces, such as antenna thrust (AT), the RMS accuracy decreases by a factor of five (FIGURE 2).

In conclusion, the ESA-JAXA collaboration on Japanese Quasi-Zenith Satellite System POD has been a resounding success. Through this continuous and mutual support, performance cross-validation and knowledge sharing, significant improvements related to modeling and subsequently to POD accuracy could be achieved for ESA as well as for JAXA. Additionally, the global scientific community benefitted from this ESA/JAXA collaboration via improved QZSS POD products and validated metadata.

Figure 1 and 2 courtesy of the authors

May 20, 2024
Copernicus Sentinel-3 and Sentinel-6 GNSS orbital products available
Artist’s depiction of the Copernicus Sentinel-6 satellite, launched in November 2020. (Image: ESA)

The Copernicus Precise Orbit Determination (CPOD) Service, in charge of computing precise orbits for the Copernicus Sentinel-1, -2, -3 and -6 missions, routinely publishes GNSS and quaternions data and precise orbital products of these missions on the POD Data Hub of the Copernicus Open Access Hub.

The following products are published:
1. Sentinel-1, 2, 3 A&B GNSS RINEX observation files (AUX_GNSSRD)
2. Sentinel-1, 2, 3 A&B Quaternions files (AUX_PROQUA)
3. Sentinel-1 A&B CPOD Predicted Orbits (AUX_PREORB)
4. Sentinel-1 A&B CPOD Restituted Orbits (AUX_RESORB)
5. Sentinel-1 A&B CPOD Precise Orbits (AUX_POEORB)
6. Sentinel-3 A&B CPOD Restituted Orbits (SR___ROE_AX)
7. Sentinel-3 A&B CPOD Medium Orbit (AUX_RESORB)
8. Sentinel-3 A&B CPOD Precise Orbits (AUX_POEORB)
9. Sentinel-3 A&B CPOD Precise Platform data (AUX_PRCPTF)
The following new products from Sentinel-3 and Sentinel-6 are now available as well. The Sentinel-6A GNSS RINEX observations include GPS and Galileo data — the first publicly available Galileo data obtained from an orbiting receiver.
1. Sentinel-3A&B CNES Medium Orbit Ephemeris (SR___MDO_AX)
2. Sentinel-3A&B CNES Precise Orbit Ephemeris (SR___POE_AX)
3. Sentinel-6A CNES Medium Orbit Ephemeris (AX____MOED_AX)
4. Sentinel-6A CNES Precise Orbit Ephemeris (AX____POE__AX)
5. Sentinel-6A CPOD Restituted Orbit Ephemeris (AX____ROE__AX)
6. Sentinel-6A GNSS RINEX observation files (AUX_GNSSRD)
7. Sentinel-6A Quaternions files (AUX_PROQUA)
The GNSS RINEX (AUX_GNSSRD) and Quaternions files (AUX_PROQUA), together with the final orbital products (AUX_POEORB, AUX_PRCPTF, SR___POE_AX, and AX____POE__AX) are available at the beginning of each mission.

The other orbital products (AUX_RESORB, SR___ROE_AX, SR___MDO_AX, AX____MOED_AX, and AX____ROE__AX) are available for at least one month, until the final products are available.

The typical accuracy of the orbital products can be found in the Regular Service Reviews carried out by the CPOD Service quarterly.

Details about these products can be found in the POD Product Handbook.

Auxiliary data needed for precise orbit determination, such as maneuvering information, can be found in the Sentinel online:
Please send questions to mailto:[email protected].
December 20, 2022