Abstract
Two-way time transfer (TWTT) is one of the most accurate time synchronization methods applied to spacecraft and ground stations. Because this method does not require prior knowledge of satellite locations and cancels out transmission path delays, it significantly boosts synchronization accuracy. However, traditional TWTT assumes that the participating platforms are relatively static, a condition typically met only by GEO satellites. This study presents a dynamic inter-satellite TWTT algorithm tailored for mobile satellites, specifically Inclined Geo-Synchronous Orbit (IGSO) and Medium Earth Orbit (MEO) satellites. By combining inter-satellite pseudo-range polynomial fitting with clock-offset polynomial fitting, the algorithm identifies the moment of minimum range to solve for high-accuracy clock offsets. Simulation results indicate that this approach can control time transfer errors to within 1ns.
Introduction
As space-based information networks become more integrated, precise time synchronization between spacecraft is essential for cooperative tasks, distributed sensing, and navigation. Two-Way Time Transfer (TWTT) has long been the gold standard for this purpose. In a static scenario, Satellite A sends a signal to Satellite B, and B simultaneously sends one to A. If the signal path is identical, the propagation delays ($\tau_{AB}$ and $\tau_{BA}$) cancel out, leaving only the clock offset.
However, for "mobile" satellites in MEO or IGSO orbits, high relative velocities mean the distance between the satellites changes during the signal's flight time. This motion creates an asymmetry where $\tau_{BA} \neq \tau_{AB}$, introducing significant errors that traditional static algorithms cannot resolve. This paper proposes a dynamic algorithm to mitigate these motion-induced biases.
Methodology
Dynamic Two-Way Time Transfer Modeling
The core challenge in mobile TWTT is the time-varying nature of the inter-satellite range. The proposed method begins by establishing a geometric model to determine mutual visibility windows between IGSO and MEO satellites. Once a link is established, the timing sequence must account for the relative motion between the platforms.
As shown in the timing diagram, the signal exchange involves specific departure and arrival times that are influenced by the distance at the moment of transmission. The algorithm collects a sequence of these pseudo-range measurements and calculated clock offsets over a set observation window (typically 20 minutes).
Dual Polynomial Fitting and the Minimum Range Principle
To extract the true clock offset from the noisy, motion-biased data, the authors employ a dual polynomial fitting technique:
- Pseudo-range Fitting: A polynomial $p = f_1(t)$ is fitted to the sequence of measured ranges.
- Clock-offset Fitting: A second polynomial $\Delta t = f_2(t)$ is fitted to the calculated offsets.
The innovation lies in the "Minimum Range Principle." The error in TWTT is minimized when the satellites are at their closest point, as this is the moment of maximum path symmetry. By finding the time $t_3$ where the derivative of the pseudo-range polynomial is zero (the vertex of the parabola), and plugging that time into the clock-offset polynomial, the algorithm achieves its highest accuracy.
Simulation and Results
The study utilized Satellite Tool Kit (STK) to simulate an IGSO satellite (35,786 km altitude, 55° inclination) and an MEO satellite (22,116 km altitude, 55° inclination). The simulation, conducted over a period in June 2008, showed that these satellites maintain mutual visibility for approximately 98.55% of their orbits.
Key results include:
- Accuracy: The algorithm consistently controlled inter-satellite time transfer errors to within 1ns.
- Precision: In optimized test cases using a 20-minute data window, the minimum error reached as low as 0.0836 ns.
- Motion Compensation: The polynomial fitting effectively smoothed out the biases caused by satellites approaching or receding from one another.
Conclusion
Traditional TWTT is insufficient for the high-velocity environments of MEO and IGSO constellations. By leveraging the geometric symmetry of the minimum range point and applying dual polynomial fitting, this dynamic algorithm provides a robust solution for sub-nanosecond synchronization. Future implementations will need to account for hardware-specific internal latencies and potential atmospheric grazing for low-altitude links, but the current simulation proves the algorithm's viability for high-accuracy space-based timekeeping.
Summary updated