Breaking the Low-Resolution Limit: High-Temporal-Resolution Satellite Bias Estimation for Enhancing Multi-Frequency PPP Ambiguity Resolution

Imagine trying to navigate the world using only the faint, shifting whispers of satellites orbiting high above. For decades, scientists have relied on these signals to pinpoint locations on Earth with incredible precision, a technology known as Global Navigation Satellite Systems. However, as these systems have evolved to broadcast more complex signals on multiple frequencies, a hidden problem has emerged. The satellites and the ground receivers they talk to have tiny, internal electronic delays—like slight lags in a conversation—that change over time. For a long time, scientists treated these delays as if they were static, unchanging constants, calculating them once a day or even once a month. But in the real world, these delays are not so still; they ripple and shift with the environment, the satellite's power usage, and even the software running inside the satellite. When these subtle, time-varying changes are ignored, the resulting position calculations lose their sharpness, drifting away from the true location.

A team of researchers has now tackled this issue by breaking the "low-resolution" limit that has held back satellite navigation for years. They developed a new way to measure and correct these satellite biases, or internal delays, with a level of detail that captures their movement in real-time. Instead of assuming the delays are fixed, their new system tracks them as they happen, updating the corrections every few seconds. By doing this, they have created a unified method that cleans up the noise in the signals from multiple satellite networks, including GPS, China's BeiDou, Europe's Galileo, and Russia's GLONASS. The result is a significant leap forward in how accurately we can determine our position on the ground, turning what was once a blurry image into a crystal-clear view.

The core of this work lies in rethinking how scientists define and measure these signal errors. Traditionally, the field has been cluttered with different terms for similar problems, such as differential code biases and uncalibrated phase delays, which were often calculated separately. This fragmentation meant that the corrections applied to the data were not always consistent with one another, leading to gaps in accuracy. The researchers proposed a cleaner approach, grouping these errors under a single concept called "Observable-specific Signal Bias." This term simply refers to the specific delay affecting a particular type of signal, whether it is the radio wave used for timing or the one used for measuring distance. By treating these biases as a unified whole rather than isolated fragments, the team could build a more robust system for correcting the data.

To achieve this, the researchers built a sophisticated software system that processes data from a global network of ground stations. Instead of waiting for a day to pass to calculate a single average value for a satellite's delay, their system analyzes the raw data continuously. They use a method that looks at the signals without combining them into simplified averages, allowing them to see the fine details of how the delays change from moment to moment. This high-temporal-resolution approach revealed something surprising: the delays are not stable. For instance, the pseudorange bias, which affects distance measurements, can shift as a satellite moves in and out of the Earth's shadow or as its internal power systems fluctuate. These changes happen on timescales of minutes or even seconds, far too fast for the old daily or monthly calculation methods to catch.

The study also uncovered that these time-varying delays are not just a nuisance for distance measurements; they significantly impact the ability to resolve the "integer ambiguity," a technical hurdle that determines whether a satellite signal can be used for centimeter-level precision. When the researchers applied their new, high-speed corrections to the data, the satellites' signals became much cleaner. They tested this system across different satellite constellations and found that it worked effectively for GPS, BeiDou, Galileo, and GLONASS. The improvements were not marginal; for GPS signals, the time it took for a receiver to lock onto a precise position dropped by nearly 30 percent in some directions. For the BeiDou system, which had previously shown slightly less stability in its signal corrections, the new method brought a noticeable boost in performance, narrowing the gap with other global systems.

Perhaps the most striking finding was the magnitude of the changes they were able to detect. In the case of GPS satellites, the researchers observed that the phase time-variant bias—the delay affecting the carrier wave used for high-precision positioning—could fluctuate by as much as a decimeter, or ten centimeters, within a single day. For other systems like BeiDou and Galileo, these fluctuations were smaller, measured in centimeters, but still significant enough to degrade accuracy if left uncorrected. By capturing these shifts in real-time, the new system allows users to correct for them immediately, rather than relying on outdated data that no longer reflects the satellite's current state.

The researchers validated their findings by comparing their new high-resolution products against traditional, lower-resolution methods. The results were clear: the new approach consistently delivered better positioning accuracy and faster convergence times. In tests, the horizontal accuracy of the position fix improved by over 50 percent for GPS when using the new high-speed corrections compared to the old float solutions. Even more impressively, the vertical accuracy, which is notoriously difficult to pin down, saw improvements of up to 35 percent. The system also proved its worth in multi-frequency scenarios, where it successfully combined data from three different frequencies to further refine the position, demonstrating that the unified bias estimation method works seamlessly across complex signal environments.

This work represents a fundamental shift in how satellite navigation data is processed. By moving away from the assumption that satellite biases are static and embracing their dynamic nature, the researchers have unlocked a new level of precision. Their unified system does not just fix one type of error; it provides a self-consistent framework that corrects multiple types of biases simultaneously, ensuring that the final position calculation is as accurate as the raw data allows. For the scientists and engineers who rely on these systems for everything from surveying land to guiding autonomous vehicles, this breakthrough means that the gap between the satellite signal and the true location on the ground is shrinking. The era of waiting for daily updates is over; the future of navigation is now measured in seconds, offering a clearer, more reliable path forward for anyone looking to find their way.