Space is becoming increasingly crowded with resident space objects (RSOs), including satellites and other debris. As these objects operate in close proximity, the risk of collisions and operational impacts escalates. Understanding and evaluating their maneuvers is crucial for maintaining a safe and functional space environment.

​

The Importance of Evaluating Maneuvers

​

RSOs that operate near each other must be closely monitored for potential collision risks and operational impacts. Evaluating these maneuvers ensures that these objects can coexist without endangering each other or their missions. Our research investigates an analytical solution for estimating the maneuvers of RSOs in rendezvous and proximity operations (RPO) scenarios. By focusing on space-based observers, we aim to develop an efficient method to estimate these maneuvers accurately.

​

Leveraging Deep Learning Techniques

​

To enhance the maneuver estimation process, we incorporate deep learning techniques. These techniques classify and estimate maneuvers for different space-based observer orbits and RPO scenarios. By utilizing pixel-spatial-temporal data and observer metadata, we can quickly characterize these maneuvers.

​

Moving Beyond Classical Image Processing

​

Traditional methods for maneuver estimation rely on classical image processing and orbit determination, which can be time-consuming and inefficient. Our approach eliminates the need for these classical methods, providing a faster and more accurate estimation process.[1] 

​

Conclusion

​

In summary, our research offers a novel analytical solution for evaluating RSOs' maneuvers using deep learning techniques. This approach ensures quick and accurate characterization of maneuvers, contributing to the safe and efficient operation of RSOs in close proximity. As space continues to fill with more objects, our method will be crucial for minimizing collision risks and maintaining a secure space environment.

​