Reduction of space debris collision prediction uncertainty based on Q-Sat precise orbit

Human space activities are becoming increasingly frequent, generating a significant amount of space debris that poses a serious threat to the safety of orbiting spacecraft. Tracking and predicting space debris to provide early warnings of potential catastrophic collisions is a commonly used method. However, due to the orbital determination errors of spacecraft and space debris, the collision prediction results are highly uncertain. Consequently, evasive maneuvers taken by spacecraft in response to collision warnings may be unnecessary, wasting valuable spacecraft fuel and significantly increasing the burden of space asset management and control. With the rapid increase in orbiting spacecraft and space debris, this burden will become unsustainable. Therefore, improving the accuracy of collision warnings is an urgent issue that needs to be addressed.

In a recent article published in Space: Science & Technology by the Distributed and Intelligent Space System Lab (DSSL) of Tsinghua University, a method is proposed to reduce the uncertainty in space debris collision predictions using precise orbital data from spacecraft. The method was applied to analyze and evaluate a hazardous conjunction event involving space debris and the Tsinghua University Gravity and Atmospheric Science Satellite (Q-Sat), effectively ruling out the possibility of a collision. The widespread application of this method will significantly enhance spacecraft safety management.

Firstly, the article briefly introduces the main methods and assumptions for calculating collision probability in collision warnings. From the calculation formulas, the main factors influencing the collision probability are the sizes of the composite objects, the conjunction distance, and the position error.

The conjunction distance is the closest distance between two objects, obtained through dynamical model predictions. The article proposes a low-Earth orbit spacecraft orbital prediction method based on precise orbital data. By using high-precision orbital data to perform short- and medium-term corrections to the atmospheric density model, the corrected model is incorporated into the "correction-prediction" process. This allows for the correction of the atmospheric density model and identification of the atmospheric drag coefficient, thereby improving the accuracy of low-Earth orbit spacecraft orbital predictions. Empirical data indicate that this method can achieve an orbital prediction accuracy better than 150 meters for 24 hours.

Position error is the prediction error after the initial orbital determination error of the two objects. The article proposes an artificial intelligence-based prediction error calculation method, using historical and predicted data of the targets to generate a prediction error dataset. A back propagation (BP) neural network is established, and through extensive training, it simulates the distribution characteristics of the position error. Empirical data show that this method can solve the problem of prediction error propagation.

The article uses the example of a hazardous conjunction event involving Q-Sat and space debris reported by online media on January 18, 2022, to analyze space debris collision warnings. Q-Sat, developed by Tsinghua University, is China's first satellite dedicated to gravity and atmospheric science measurements, with a mission to jointly detect low-Earth orbit atmospheric density and long-wavelength gravity fields. The satellite weighs 21.2 kg, with a diameter of 510 mm, and its initial mission orbit is a sun-synchronous orbit at an altitude of 499.3 km. It is equipped with a self-developed dual-frequency global navigation receiver, achieving orbit determination accuracy at the centimeter level. Using the high-precision orbital data obtained from the satellite and the collision prediction method proposed in the article, it was determined that the closest distance between the two targets was no less than 2.7 km, with a collision probability of 1.16×10⁻¹¹, well below the international evasion threshold of 10⁻⁴. This was a false alarm, and subsequent observational data confirmed that no collision occurred.

The article concludes that with the rapid growth of orbiting spacecraft, the risks and costs of space asset safety management are increasing. In the future, carrying low-cost precise orbit determination payloads on orbiting spacecraft, combined with the collision prediction calculation method proposed in the article, can significantly reduce the ambiguity in collision predictions, identify and eliminate low-risk hazardous conjunction warnings early, and improve space safety management levels.