Comprehensive VLBI observations of Galileo satellites with the AuScope array

This paper presents the first comprehensive VLBI observations of Galileo navigation satellites using the Australian AuScope array, demonstrating the feasibility of deriving station coordinates and establishing inter-technique ties between VLBI and GNSS reference frames to support future geodetic co-location missions like ESA's Genesis.

The Gist

The Cosmic Game of "Pin the Tail on the Donkey" (But with Satellites)

Imagine you are trying to measure the exact distance between three friends standing in different cities (Hobart, Katherine, and Yarragadee in Australia). Usually, to do this with extreme precision, scientists use radio telescopes to listen to the faint, static-like whispers of distant black holes and galaxies billions of light-years away. It's like trying to hear a pin drop in a stadium from three miles away.

But this paper is about a new, much louder game: listening to the satellites orbiting right above our heads.

Specifically, the team used the AuScope VLBI array (a team of three 12-meter radio dishes) to listen to Galileo satellites (the European version of GPS). The goal? To create a "space tie"—a super-accurate ruler connecting the ground-based measurement system (VLBI) with the satellite navigation system (GNSS).

Here is the breakdown of their adventure, explained simply:

1. The Challenge: Listening to a Shout in a Library

Normally, these radio telescopes are built to hear the faintest whispers of the universe. Galileo satellites, however, are screaming. Their signals are millions of times stronger than the cosmic whispers the telescopes usually listen for.

• The Analogy: It's like trying to listen to a moth fluttering in a library, but suddenly someone starts playing a heavy metal concert next door. The equipment wasn't built for that volume; it could easily get "blown out" or damaged.
• The Fix: The team had to build a special "volume knob" and a new signal path (an L-band signal chain) to handle the loud satellite signals without frying their sensitive equipment.

2. The Problem: The "Stop-and-Go" Dance

To get a clear picture, a radio telescope needs to keep its "eye" locked on the target. But these telescopes are big and heavy; they can't smoothly glide to follow a fast-moving satellite like a modern camera on a drone.

• The Analogy: Imagine trying to keep a camera focused on a race car zooming by. Instead of a smooth pan, the camera operator has to snap the camera to a new position every few seconds, wait, snap again, wait, and snap again. This is called stepwise tracking.
• The Result: This "jittery" movement creates a "striped" pattern in the data (like a bad video connection). The team tested different speeds: snapping every 30 seconds, 20, 10, or 5. They found that snapping every 10 seconds was the sweet spot—fast enough to keep up, but slow enough not to break the system.

3. The Digital Upgrade: 2-Bit vs. 8-Bit

When recording these signals, the team had to decide how much detail to save.

• The Analogy: Think of 2-bit recording as a low-resolution sketch (just black and white dots). It's the standard for listening to faint stars. 8-bit recording is like a high-definition photo with millions of colors.
• The Discovery: Because satellite signals are so complex and "digital" (they are modulated data streams), the high-definition 8-bit recording was much better. It captured the signal with far less "noise" (static), especially for the E1 frequency band. It was like switching from a grainy security camera to a 4K movie camera.

4. The "Space Tie" Achievement

The biggest breakthrough in this paper is that they successfully used these ground telescopes to measure the position of the satellites, and then used that data to figure out exactly where the telescopes themselves were standing on Earth.

• The Result: They connected the "Ground Frame" (where the telescopes are) with the "Satellite Frame" (where the GPS is).
• The Accuracy: They got the telescope positions right to within a few meters.
  • Wait, isn't that bad? For standard GPS, meters are great. But for high-precision science (like measuring tectonic plates moving millimeters a year), meters is a huge error.
  • Why the error? The paper admits there are still "ghost signals" and unexplained systematic errors in the data (like the "striped" tracking noise or signals bouncing off the satellite's own antenna in weird ways). The "noise" is currently drowning out the "signal" of the precise measurement.

5. Why This Matters: The "Genesis" Mission

You might wonder, "If the accuracy is only at the meter level, why is this a big deal?"

• The Future: The European Space Agency is launching a mission called Genesis in 2028. This satellite will carry all the tools needed to measure the Earth's shape and gravity.
• The Goal: To make the Earth's reference map perfect (accurate to 1 millimeter), we need to know exactly how the satellite's tools relate to the ground tools.
• The Paper's Role: This study is the "test flight." It proves that we can do this. It shows us where the problems are (the tracking jitter, the signal noise) so engineers can fix them before the big Genesis mission launches.

The Bottom Line

This paper is a "proof of concept." It's like building a prototype car to see if it can drive on a new, bumpy road.

• Did it work? Yes, the car drove!
• Is it smooth? No, it's still bumpy (meter-level accuracy).
• Did we learn anything? Yes! We learned exactly where the bumps are (tracking intervals, signal processing, bit-depth) so we can smooth the road out for the future.

The team successfully turned a set of telescopes designed for the deep universe into a tool that can "see" our own neighborhood in space, paving the way for a future where our maps of Earth are as perfect as humanly possible.

Technical Summary

1. Problem Statement

The International Terrestrial Reference Frame (ITRF) relies on combining four space-geodetic techniques: VLBI, GNSS, SLR, and DORIS. A major limitation in current ITRF realizations is the "inter-technique tie," which links these techniques via terrestrial local surveys at co-location sites. Discrepancies of several centimeters between these local ties and global estimates degrade the overall frame consistency.

To overcome this, the concept of "space ties" (co-location in space) has been proposed, where a single satellite carries instruments for multiple techniques. The European Space Agency's (ESA) Genesis mission (scheduled for 2028) will be the first to combine all four techniques on one platform. However, for VLBI to support this, it must successfully observe Earth-orbiting satellites. This is currently non-standard because:

• Signal Mismatch: VLBI traditionally observes weak, broadband signals from distant extragalactic sources. GNSS satellites transmit extremely strong, narrowband L-band signals.
• Processing Gaps: Standard VLBI processing chains are not designed for near-field targets or modulated navigation signals.
• Tracking Limitations: Many VLBI antennas cannot continuously track moving satellites, requiring stepwise tracking which introduces geometric errors.

2. Methodology

The study utilized the AuScope VLBI array in Australia, consisting of three 12-m VGOS antennas in Hobart, Katherine, and Yarragadee.

• Hardware & Signal Chain: The antennas were equipped with a new L-band signal chain (500–1650 MHz) to receive Galileo E1 (1575.42 MHz) and E6 (1278.75 MHz) signals. The system used VGOS cryogenic LNAs and a DBBC3 sampler.
• Observation Strategy:
  • Tracking: Due to hardware limitations, a stepwise tracking approach was used. Antennas updated their pointing at fixed intervals (tested at 30, 20, 10, and 5 seconds) rather than continuous tracking.
  • Recording Modes: Two modes were tested:
    • 2-bit: Standard geodetic VLBI depth (Rec 4x32 2).
    • 8-bit: Higher depth (Rec 2x64 8) to better capture the deterministic, modulated nature of GNSS signals.
  • Sessions: The study included two short test experiments (Sessions A1/A2) and four full-scale 24-hour observing sessions (Sessions B1–B4) targeting Galileo satellites.
• Data Processing:
  • Correlation: Performed using the DiFX software correlator. Three methods for generating near-field delay models were compared:
    1. M1: TLEs + SPICE kernel.
    2. M2: SP3 orbit files + difxcalc.
    3. M3: SP3 orbit files + VieVS (Vienna VLBI and Satellite software) using the Klioner model.
  • Fringe Fitting: Performed using fourfit (HOPS). Various fringe modes were tested, including optimized bandpass filtering to maximize Signal-to-Noise Ratio (SNR).
  • Analysis: Geodetic parameters (station coordinates, troposphere, clocks) were estimated using VieVS with a least-squares adjustment.

3. Key Contributions

• First Inter-Technique Tie: This study presents the first-ever estimation of VLBI station coordinates derived from observations of Galileo navigation satellites, establishing a direct observational link between the VLBI and GNSS reference frames.
• 8-bit vs. 2-bit Quantification: It demonstrates that 8-bit digitization significantly improves precision for GNSS signals compared to the standard 2-bit, particularly in the E1 band, by better handling the modulated signal structure.
• Optimized Processing Workflow: The paper establishes a robust workflow for correlating near-field targets, identifying that SP3-based delay models (M3) significantly outperform TLE-based models in terms of coherence and SNR.
• Stepwise Tracking Analysis: It quantifies the impact of discrete antenna repositioning, identifying that 10-second intervals are sufficient to minimize tracking-induced artifacts for current antenna capabilities.

4. Key Results

• Delay Precision:
  • E1 Band: Achieved precisions of 3.9 to 11.1 picoseconds (ps) for 1-second integration times.
  • E6 Band: Precisions ranged from 19.4 to 135.7 ps. The lower precision in E6 is attributed to lower transmitted power and reduced antenna gain at lower L-band frequencies.
  • 8-bit Advantage: 8-bit recordings showed up to 2x better precision than 2-bit in the E1 band.
• Tracking Impact: Stepwise tracking with 30s and 20s intervals introduced a distinct "striped" systematic error of ~20 ps in the residuals. Reducing the interval to 10 seconds eliminated this visible pattern, though frequent repositioning may still contribute to noise.
• Delay Models: The M3 (VieVS + SP3) implementation yielded the highest SNR and lowest residual phase rates, outperforming TLE-based models by ~80% in SNR.
• Station Coordinates:
  • Estimated station positions agreed with a priori values at the metre level (residuals up to ±6 m).
  • Baseline lengths agreed at the sub-metre level (systematically shorter by ~40–85 cm).
  • The large coordinate offsets are attributed to unmodelled systematic effects (e.g., phase center offsets, signal chain artifacts) rather than random noise, as the network showed a consistent translational shift.
• Unmodelled Signals: Significant systematic residuals (up to ±1.5 meters or ±5000 ps) were observed, varying by satellite and frequency band. These are likely caused by unmodelled phase center offsets or signal chain distortions, not by the delay models themselves.

5. Significance

This work provides the critical groundwork for the upcoming ESA Genesis mission. It proves that:

1. Feasibility: Large-scale VLBI observations of GNSS satellites are technically feasible using existing VGOS infrastructure with minor signal chain modifications.
2. Precision Potential: While current results are at the metre level (limited by systematic errors), the observables themselves have picosecond-level precision. This indicates that the technique is capable of millimetre-level accuracy once systematic effects (like phase center calibrations) are resolved.
3. Future Roadmap: The study identifies specific areas for development required before routine operations, including continuous satellite tracking, phase calibration, and the coherent combination of polarisation products.

In summary, this paper successfully transitions VLBI satellite tracking from theoretical simulation to practical, large-scale observation, validating the concept of "space ties" and paving the way for the next generation of millimetre-accurate global reference frames.