A Tracklet-Score Metric for a passive-optical Telescope Network

The sky above us is no longer empty. It is filled with a silent, swirling traffic of human-made objects, from active satellites that power our global communications to the countless fragments of debris left behind by decades of space activity. To keep this traffic safe and predictable, astronomers and engineers must constantly track these objects, knowing exactly where they are and where they are going. This task relies on networks of telescopes that watch the heavens, capturing brief glimpses of these objects as they cross the sky. Each glimpse is not just a single photo, but a short sequence of observations called a tracklet. These tracklets are the raw building blocks used to calculate orbits and predict if two objects might collide. However, not every glimpse is equally useful. A telescope with a small lens might see a bright satellite clearly, while a larger one might catch a faint, tumbling piece of debris. Some observations are sent to the central database instantly, while others arrive hours later. The challenge for the scientists managing these networks is to figure out which of these thousands of daily glimpses are the most valuable for keeping the catalog of space objects accurate and up to date.

Johannes Herzog and Hauke Fiedler, researchers at the German Aerospace Center, have developed a new way to answer this question. They created a scoring system that acts like a quality rating for every single tracklet collected by the SMARTnet telescope network. Instead of trying to analyze the complex math inside every single observation to judge its worth, their method looks at the circumstances under which the observation was made. They ask simple, practical questions: How fast was the data delivered? How many photos were taken in the sequence? How long did the telescope watch the object? Was there a large gap where the object disappeared from view? And finally, how big is the telescope's main mirror? By assigning a score to each of these factors and combining them, they produce a single number that tells operators how much value a specific tracklet brings to the overall effort of monitoring space.

The researchers tested this new system against an older method that was already in use. The old approach tried to judge a tracklet by looking at the precision of the measurements themselves, comparing the observed position and speed of the object against a perfect reference. While this sounds thorough, the authors found it had a major flaw. Because it relied on calculating speeds and positions from the data, a single lucky measurement with very low error could make a tracklet appear thousands of times more valuable than it really was. This created a chaotic scoring system where some tracklets received massive, unrealistic scores while others with similar practical value were rated almost worthless. It was like judging a book solely by the thickness of one page, ignoring the story entirely.

In contrast, the new scoring system is built on the idea that the value of a tracklet comes from the reliability of the sensor and the strategy used to capture it. The researchers defined five key ingredients for their score. First, they value speed. A tracklet that arrives at the database within one hour of being taken gets the highest possible rating for timeliness, because immediate data is crucial for predicting close encounters between satellites. As the delay grows, the score drops, but it never disappears completely, acknowledging that even old data is useful for refining orbits later. Second, they look at the number of photos. A sequence with more measurements is generally more robust, helping to filter out errors, so the score rises as the count increases, though it grows more slowly to avoid overvaluing extremely long sequences.

The third factor considers the length of the observation, both in time and in the angle the object moved across the sky. This is designed to be fair to all types of orbits. Fast-moving objects in low orbits might only be visible for a few minutes, while slow-moving objects in high orbits might be watched for hours. The system awards points if the observation covers a significant amount of time or a significant amount of sky, taking the better of the two measures so that neither fast nor slow objects are penalized. The fourth factor checks for gaps. If the telescope loses sight of the object for a long time, perhaps because it passed behind a star, the score dips slightly, but only if the gap becomes very large. This ensures that operators are not punished for the natural difficulties of observing faint objects or for testing new strategies. The final factor is the size of the telescope's primary mirror. A larger mirror collects more light, allowing the network to see fainter, smaller objects that smaller telescopes simply cannot detect. This capability is highly valued, so larger mirrors receive higher scores.

To turn these five ingredients into a final number, the researchers assigned different weights to each one. They decided that how quickly the data arrives and how many measurements it contains are the most critical factors, giving them the highest importance. The size of the telescope and the gaps in the data are considered less critical, receiving lower weights. When they applied this system to nearly forty-three thousand tracklets collected in 2023, the results were remarkably consistent. The scores fell into a tight, manageable range, mostly between 1.5 and 9.4. This stands in stark contrast to the old method, which produced scores that jumped wildly across millions of different values, making it difficult to compare the contributions of different telescopes fairly.

The authors note that this new metric is not a tool to judge the performance of the people operating the telescopes, as different partners have different missions and constraints. Instead, it serves as a transparent way to understand the overall health and composition of the network. It helps them see which types of observations are most common and how well the network is covering the sky. By focusing on the characteristics of the sensors and the delivery process rather than trying to mathematically perfect every single data point, the researchers have created a tool that is both simpler and more reliable. It allows the network to treat every tracklet with a clear sense of its true worth, ensuring that the catalog of space objects remains accurate and that the safety of our orbital environment is maintained.