![[Translate to Englisch:] Blick auf die Erde aus dem Weltraum mit Planeten und Milchstraße im Vordergrund.](/fileadmin/_processed_/6/4/csm_2021-08-26_Adobe_Stock_elen31HG-Y009_5d479f4069.jpg "Daily Space News zur Weltraumwettervorhersage")
The German Luftwaffe Space Command has been put into service and space is therefore also an area of activity for the Bundeswehr.
A meaningful space situation picture for an effective situation center also includes weather forecasts and collision warnings. The aim of such space weather forecasts is to protect satellites from "weather swings" in space that might endanger the electronics of an observation satellite, or collisions that, for example, penetrate a solar panel at high speed.
To optimize this, we have developed an AI analysis tool for space weather forecasting and an AI application for collision warning, based on the CIHBw development platform. Together with the BWI GmbH data analytics team, we worked on this in a "team-of-teams" approach with System Center 25 for the Air Force Space Situational Awareness Center.
The AI continuously evaluates all the data collected for the forecasts and compares it with the expected values of the models. If a data source fails, the AI is able to simulate data over a short period of time - in order to keep the overall picture accurate. This means that the Space Situational Awareness Center remains capable of acting and can, for example, prepare satellites in orbit for the calculated weather.
This involves not only the consolidation of momentary individual data, but also the collection of complete data series over longer periods of time. The analysis tools are based on machine learning systems.
For the Bundeswehr, such AI processes are examples of the important new capabilities that a modern army needs. AI-supported and -protected space weather increases the safety of satellites in orbit. And these in turn contribute to critical tasks, first and foremost national and alliance defense.
The second AI application is designed to warn of collisions in space. To do this, it will analyze the trajectories of unknown objects such as satellites and space debris. It will use machine learning methods to find patterns in their orbital movements. The aim is to determine the trajectories of unknown Low Earth Orbit (LEO) objects as accurately as possible in order to carry out precise collision analyses. At the same time, these objects are to be collected in a catalog and made identifiable again.
